train_gcn_nofuture.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "3f47fa9d",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-07-22 00:29:52.020080: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.\n",
      "To enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2023-07-22 00:29:53.674412: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from tensorflow.keras import Model, Input, layers\n",
    "from math import sqrt\n",
    "from tensorflow import keras\n",
    "from sklearn.metrics import mean_squared_error as mse\n",
    "from sklearn.metrics import mean_absolute_error as mae\n",
    "from tensorflow.keras.models import load_model\n",
    "from tensorflow.keras.layers import *\n",
    "from tensorflow.keras.models import *\n",
    "from tensorflow.keras.optimizers import *\n",
    "from tensorflow.keras.callbacks import *\n",
    "\n",
    "from spektral.layers import GCNConv\n",
    "\n",
    "import keras_tuner as kt\n",
    "from BaselinePrerocess import gcn_process\n",
    "from graph import graph_topology\n",
    "from baselines.gcn import gcn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "0092be29",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"4\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f9b08072",
   "metadata": {},
   "source": [
    "### Parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "0cef91ff",
   "metadata": {},
   "outputs": [],
   "source": [
    "# ====== preprocessing parameters ======\n",
    "n_hours = 72\n",
    "K = 24 \n",
    "masked_value = 1e-10\n",
    "split_1 = 0.8\n",
    "split_2 = 0.9\n",
    "sigma2 = 0.1\n",
    "epsilon = 0.5\n",
    "\n",
    "\n",
    "# ===== model parameters ======\n",
    "gcn_unit1 = 32\n",
    "gcn_unit2 = 16\n",
    "lstm_units = 32\n",
    "dropout = 0\n",
    "learning_rate = 5e-4\n",
    "decay_steps = 10000\n",
    "decay_rate = 0.95\n",
    "PATIENCE = 500\n",
    "EPOCHS = 3000\n",
    "BATCH = 512"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3788f5e1",
   "metadata": {},
   "source": [
    "### Preprocess"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "39e5c08d",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['WS_S1', 'WS_S4', 'FLOW_S25A', 'GATE_S25A', 'HWS_S25A', 'TWS_S25A',\n",
      "       'FLOW_S25B', 'GATE_S25B', 'GATE_S25B2', 'HWS_S25B', 'TWS_S25B',\n",
      "       'PUMP_S25B', 'FLOW_S26', 'GATE_S26_1', 'GATE_S26_2', 'HWS_S26',\n",
      "       'TWS_S26', 'PUMP_S26', 'MEAN_RAIN'],\n",
      "      dtype='object')\n"
     ]
    }
   ],
   "source": [
    "train_X_mask, val_X_mask, test_X_mask, \\\n",
    "train_ws_y, val_ws_y, test_ws_y, \\\n",
    "scaler, ws_scaler = gcn_process(n_hours, K, masked_value, split_1, split_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "61f87690",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(77069, 14, 96) (9634, 14, 96) (19268, 14, 96) (77069, 96) (9634, 96) (19268, 96)\n"
     ]
    }
   ],
   "source": [
    "print(train_X_mask.shape, val_X_mask.shape, test_X_mask.shape, train_ws_y.shape, val_ws_y.shape, test_ws_y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "d19c698c",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_X_mask = train_X_mask[:, :, :n_hours]\n",
    "val_X_mask = val_X_mask[:, :, :n_hours]\n",
    "test_X_mask = test_X_mask[:, :, :n_hours]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "239a2039",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(77069, 14, 72) (9634, 14, 72) (19268, 14, 72) (77069, 96) (9634, 96) (19268, 96)\n"
     ]
    }
   ],
   "source": [
    "print(train_X_mask.shape, val_X_mask.shape, test_X_mask.shape, train_ws_y.shape, val_ws_y.shape, test_ws_y.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "995e4cec",
   "metadata": {},
   "source": [
    "### Graph topology"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "e6f1452c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "node_indices: [ 0  0  0  0  0  1  1  1  2  2  2  3  3  4  4  4  4  5  5  5  6  6  6  7\n",
      "  7  7  8  8  8  8  8  9  9  9 10 10 10 11 11 11 12 12 12 12 13 13 13 13\n",
      " 13 13 13 13 13 13 13 13 13 13] \n",
      "neighbor_indices: [ 1  4  8 12 13  1  4 13  3  4 13  2 13  0  1  2 13  7  8 13  7  8 13  5\n",
      "  6 13  0  5  6 12 13 11 12 13 11 12 13  9 10 13  0  9 10 13  0  1  2  3\n",
      "  4  5  6  7  8  9 10 11 12 13]\n",
      "number of nodes: 14, number of edges: 58\n"
     ]
    }
   ],
   "source": [
    "train_adj_mat, val_adj_mat, test_adj_mat = graph_topology(n_hours, K, sigma2, epsilon, len(train_ws_y), len(val_ws_y), len(test_ws_y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "f7b6cba2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(77069, 14, 14) (9634, 14, 14) (19268, 14, 14)\n"
     ]
    }
   ],
   "source": [
    "print(train_adj_mat.shape, val_adj_mat.shape, test_adj_mat.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8003a2dc",
   "metadata": {},
   "source": [
    "### Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "6bf53ab9",
   "metadata": {},
   "outputs": [],
   "source": [
    "n_nodes = train_adj_mat.shape[-1]\n",
    "n_timesteps = train_X_mask.shape[-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "a7d6a860",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Layer lstm_1 will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU.\n",
      "Model: \"model_1\"\n",
      "__________________________________________________________________________________________________\n",
      " Layer (type)                Output Shape                 Param #   Connected to                  \n",
      "==================================================================================================\n",
      " input_4 (InputLayer)        [(None, 14, 72)]             0         []                            \n",
      "                                                                                                  \n",
      " input_3 (InputLayer)        [(None, 14, 14)]             0         []                            \n",
      "                                                                                                  \n",
      " gcn_conv_2 (GCNConv)        (None, 14, 32)               2336      ['input_4[0][0]',             \n",
      "                                                                     'input_3[0][0]']             \n",
      "                                                                                                  \n",
      " gcn_conv_3 (GCNConv)        (None, 14, 16)               528       ['gcn_conv_2[0][0]',          \n",
      "                                                                     'input_3[0][0]']             \n",
      "                                                                                                  \n",
      " lstm_1 (LSTM)               (None, 14, 32)               13440     ['input_4[0][0]']             \n",
      "                                                                                                  \n",
      " concatenate_1 (Concatenate  (None, 14, 48)               0         ['gcn_conv_3[0][0]',          \n",
      " )                                                                   'lstm_1[0][0]']              \n",
      "                                                                                                  \n",
      " flatten_1 (Flatten)         (None, 672)                  0         ['concatenate_1[0][0]']       \n",
      "                                                                                                  \n",
      " dropout_1 (Dropout)         (None, 672)                  0         ['flatten_1[0][0]']           \n",
      "                                                                                                  \n",
      " dense_1 (Dense)             (None, 96)                   64608     ['dropout_1[0][0]']           \n",
      "                                                                                                  \n",
      "==================================================================================================\n",
      "Total params: 80912 (316.06 KB)\n",
      "Trainable params: 80912 (316.06 KB)\n",
      "Non-trainable params: 0 (0.00 Byte)\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# n_nodes, n_timesteps, gcn1, gcn2, lstm_unit\n",
    "model, GCNConv = gcn(gcn1=gcn_unit1,\n",
    "                     gcn2=gcn_unit2,\n",
    "                     n_nodes=n_nodes,\n",
    "                     n_timesteps=n_timesteps,\n",
    "                     lstm_unit=lstm_units,\n",
    "                     dropout=dropout,\n",
    "                     masked_value=masked_value\n",
    "                    )\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "ad13d3ec",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/3000\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-07-22 00:32:16.710568: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x7f0d9401b030 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:\n",
      "2023-07-22 00:32:16.710623: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): NVIDIA GeForce GTX 1080 Ti, Compute Capability 6.1\n",
      "2023-07-22 00:32:16.720720: I tensorflow/compiler/mlir/tensorflow/utils/dump_mlir_util.cc:255] disabling MLIR crash reproducer, set env var `MLIR_CRASH_REPRODUCER_DIRECTORY` to enable.\n",
      "2023-07-22 00:32:16.951844: I tensorflow/compiler/xla/stream_executor/cuda/cuda_dnn.cc:432] Loaded cuDNN version 8600\n",
      "2023-07-22 00:32:17.368660: I ./tensorflow/compiler/jit/device_compiler.h:186] Compiled cluster using XLA!  This line is logged at most once for the lifetime of the process.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 1: val_mae improved from inf to 0.10513, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 10s - loss: 0.1354 - mae: 0.1862 - val_loss: 0.0188 - val_mae: 0.1051 - 10s/epoch - 64ms/step\n",
      "Epoch 2/3000\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/aul/homes/jshi008/miniconda3/envs/tf213/lib/python3.10/site-packages/keras/src/engine/training.py:3000: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.\n",
      "  saving_api.save_model(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 2: val_mae improved from 0.10513 to 0.07130, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 0.0076 - mae: 0.0637 - val_loss: 0.0084 - val_mae: 0.0713 - 4s/epoch - 24ms/step\n",
      "Epoch 3/3000\n",
      "\n",
      "Epoch 3: val_mae improved from 0.07130 to 0.05977, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 0.0036 - mae: 0.0453 - val_loss: 0.0058 - val_mae: 0.0598 - 4s/epoch - 25ms/step\n",
      "Epoch 4/3000\n",
      "\n",
      "Epoch 4: val_mae improved from 0.05977 to 0.05532, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 0.0025 - mae: 0.0380 - val_loss: 0.0049 - val_mae: 0.0553 - 4s/epoch - 25ms/step\n",
      "Epoch 5/3000\n",
      "\n",
      "Epoch 5: val_mae improved from 0.05532 to 0.05124, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 0.0021 - mae: 0.0345 - val_loss: 0.0042 - val_mae: 0.0512 - 4s/epoch - 25ms/step\n",
      "Epoch 6/3000\n",
      "\n",
      "Epoch 6: val_mae improved from 0.05124 to 0.04772, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 0.0018 - mae: 0.0319 - val_loss: 0.0036 - val_mae: 0.0477 - 4s/epoch - 25ms/step\n",
      "Epoch 7/3000\n",
      "\n",
      "Epoch 7: val_mae improved from 0.04772 to 0.04492, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 0.0016 - mae: 0.0299 - val_loss: 0.0032 - val_mae: 0.0449 - 4s/epoch - 25ms/step\n",
      "Epoch 8/3000\n",
      "\n",
      "Epoch 8: val_mae improved from 0.04492 to 0.04383, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 0.0014 - mae: 0.0283 - val_loss: 0.0031 - val_mae: 0.0438 - 4s/epoch - 25ms/step\n",
      "Epoch 9/3000\n",
      "\n",
      "Epoch 9: val_mae improved from 0.04383 to 0.04137, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 0.0013 - mae: 0.0269 - val_loss: 0.0027 - val_mae: 0.0414 - 4s/epoch - 24ms/step\n",
      "Epoch 10/3000\n",
      "\n",
      "Epoch 10: val_mae improved from 0.04137 to 0.04013, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 0.0012 - mae: 0.0257 - val_loss: 0.0026 - val_mae: 0.0401 - 4s/epoch - 24ms/step\n",
      "Epoch 11/3000\n",
      "\n",
      "Epoch 11: val_mae improved from 0.04013 to 0.03849, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 0.0011 - mae: 0.0246 - val_loss: 0.0024 - val_mae: 0.0385 - 4s/epoch - 24ms/step\n",
      "Epoch 12/3000\n",
      "\n",
      "Epoch 12: val_mae improved from 0.03849 to 0.03786, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 9.9826e-04 - mae: 0.0236 - val_loss: 0.0023 - val_mae: 0.0379 - 4s/epoch - 24ms/step\n",
      "Epoch 13/3000\n",
      "\n",
      "Epoch 13: val_mae improved from 0.03786 to 0.03598, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 9.3120e-04 - mae: 0.0227 - val_loss: 0.0021 - val_mae: 0.0360 - 4s/epoch - 24ms/step\n",
      "Epoch 14/3000\n",
      "\n",
      "Epoch 14: val_mae improved from 0.03598 to 0.03542, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 8.7223e-04 - mae: 0.0219 - val_loss: 0.0020 - val_mae: 0.0354 - 4s/epoch - 24ms/step\n",
      "Epoch 15/3000\n",
      "\n",
      "Epoch 15: val_mae improved from 0.03542 to 0.03470, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 8.1580e-04 - mae: 0.0211 - val_loss: 0.0019 - val_mae: 0.0347 - 4s/epoch - 24ms/step\n",
      "Epoch 16/3000\n",
      "\n",
      "Epoch 16: val_mae improved from 0.03470 to 0.03390, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 7.6803e-04 - mae: 0.0204 - val_loss: 0.0018 - val_mae: 0.0339 - 3s/epoch - 23ms/step\n",
      "Epoch 17/3000\n",
      "\n",
      "Epoch 17: val_mae improved from 0.03390 to 0.03319, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 7.2134e-04 - mae: 0.0197 - val_loss: 0.0018 - val_mae: 0.0332 - 4s/epoch - 24ms/step\n",
      "Epoch 18/3000\n",
      "\n",
      "Epoch 18: val_mae improved from 0.03319 to 0.03237, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 6.7909e-04 - mae: 0.0190 - val_loss: 0.0017 - val_mae: 0.0324 - 4s/epoch - 24ms/step\n",
      "Epoch 19/3000\n",
      "\n",
      "Epoch 19: val_mae improved from 0.03237 to 0.03210, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 6.4250e-04 - mae: 0.0184 - val_loss: 0.0017 - val_mae: 0.0321 - 4s/epoch - 24ms/step\n",
      "Epoch 20/3000\n",
      "\n",
      "Epoch 20: val_mae improved from 0.03210 to 0.03147, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 6.0989e-04 - mae: 0.0179 - val_loss: 0.0016 - val_mae: 0.0315 - 4s/epoch - 23ms/step\n",
      "Epoch 21/3000\n",
      "\n",
      "Epoch 21: val_mae did not improve from 0.03147\n",
      "151/151 - 3s - loss: 5.8026e-04 - mae: 0.0174 - val_loss: 0.0016 - val_mae: 0.0315 - 3s/epoch - 23ms/step\n",
      "Epoch 22/3000\n",
      "\n",
      "Epoch 22: val_mae improved from 0.03147 to 0.03113, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 5.5387e-04 - mae: 0.0169 - val_loss: 0.0016 - val_mae: 0.0311 - 4s/epoch - 24ms/step\n",
      "Epoch 23/3000\n",
      "\n",
      "Epoch 23: val_mae improved from 0.03113 to 0.03042, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 5.3089e-04 - mae: 0.0165 - val_loss: 0.0015 - val_mae: 0.0304 - 4s/epoch - 26ms/step\n",
      "Epoch 24/3000\n",
      "\n",
      "Epoch 24: val_mae improved from 0.03042 to 0.02970, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 5.0969e-04 - mae: 0.0162 - val_loss: 0.0014 - val_mae: 0.0297 - 4s/epoch - 25ms/step\n",
      "Epoch 25/3000\n",
      "\n",
      "Epoch 25: val_mae improved from 0.02970 to 0.02918, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 4.8963e-04 - mae: 0.0158 - val_loss: 0.0014 - val_mae: 0.0292 - 4s/epoch - 26ms/step\n",
      "Epoch 26/3000\n",
      "\n",
      "Epoch 26: val_mae did not improve from 0.02918\n",
      "151/151 - 4s - loss: 4.7440e-04 - mae: 0.0155 - val_loss: 0.0015 - val_mae: 0.0302 - 4s/epoch - 24ms/step\n",
      "Epoch 27/3000\n",
      "\n",
      "Epoch 27: val_mae did not improve from 0.02918\n",
      "151/151 - 4s - loss: 4.5768e-04 - mae: 0.0152 - val_loss: 0.0014 - val_mae: 0.0297 - 4s/epoch - 24ms/step\n",
      "Epoch 28/3000\n",
      "\n",
      "Epoch 28: val_mae did not improve from 0.02918\n",
      "151/151 - 4s - loss: 4.4427e-04 - mae: 0.0149 - val_loss: 0.0014 - val_mae: 0.0294 - 4s/epoch - 24ms/step\n",
      "Epoch 29/3000\n",
      "\n",
      "Epoch 29: val_mae did not improve from 0.02918\n",
      "151/151 - 4s - loss: 4.3174e-04 - mae: 0.0147 - val_loss: 0.0014 - val_mae: 0.0298 - 4s/epoch - 24ms/step\n",
      "Epoch 30/3000\n",
      "\n",
      "Epoch 30: val_mae did not improve from 0.02918\n",
      "151/151 - 4s - loss: 4.2009e-04 - mae: 0.0145 - val_loss: 0.0014 - val_mae: 0.0295 - 4s/epoch - 24ms/step\n",
      "Epoch 31/3000\n",
      "\n",
      "Epoch 31: val_mae improved from 0.02918 to 0.02910, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 4.0929e-04 - mae: 0.0143 - val_loss: 0.0014 - val_mae: 0.0291 - 4s/epoch - 24ms/step\n",
      "Epoch 32/3000\n",
      "\n",
      "Epoch 32: val_mae improved from 0.02910 to 0.02884, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 4.0144e-04 - mae: 0.0141 - val_loss: 0.0014 - val_mae: 0.0288 - 4s/epoch - 25ms/step\n",
      "Epoch 33/3000\n",
      "\n",
      "Epoch 33: val_mae improved from 0.02884 to 0.02858, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 3.9131e-04 - mae: 0.0139 - val_loss: 0.0013 - val_mae: 0.0286 - 4s/epoch - 24ms/step\n",
      "Epoch 34/3000\n",
      "\n",
      "Epoch 34: val_mae did not improve from 0.02858\n",
      "151/151 - 4s - loss: 3.8434e-04 - mae: 0.0138 - val_loss: 0.0014 - val_mae: 0.0293 - 4s/epoch - 25ms/step\n",
      "Epoch 35/3000\n",
      "\n",
      "Epoch 35: val_mae improved from 0.02858 to 0.02828, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 3.7627e-04 - mae: 0.0136 - val_loss: 0.0013 - val_mae: 0.0283 - 4s/epoch - 25ms/step\n",
      "Epoch 36/3000\n",
      "\n",
      "Epoch 36: val_mae did not improve from 0.02828\n",
      "151/151 - 4s - loss: 3.6846e-04 - mae: 0.0134 - val_loss: 0.0013 - val_mae: 0.0284 - 4s/epoch - 24ms/step\n",
      "Epoch 37/3000\n",
      "\n",
      "Epoch 37: val_mae improved from 0.02828 to 0.02793, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 3.6268e-04 - mae: 0.0133 - val_loss: 0.0013 - val_mae: 0.0279 - 4s/epoch - 25ms/step\n",
      "Epoch 38/3000\n",
      "\n",
      "Epoch 38: val_mae improved from 0.02793 to 0.02771, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 3.5636e-04 - mae: 0.0132 - val_loss: 0.0013 - val_mae: 0.0277 - 4s/epoch - 24ms/step\n",
      "Epoch 39/3000\n",
      "\n",
      "Epoch 39: val_mae did not improve from 0.02771\n",
      "151/151 - 4s - loss: 3.5051e-04 - mae: 0.0130 - val_loss: 0.0013 - val_mae: 0.0279 - 4s/epoch - 25ms/step\n",
      "Epoch 40/3000\n",
      "\n",
      "Epoch 40: val_mae did not improve from 0.02771\n",
      "151/151 - 4s - loss: 3.4530e-04 - mae: 0.0129 - val_loss: 0.0013 - val_mae: 0.0281 - 4s/epoch - 25ms/step\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 41/3000\n",
      "\n",
      "Epoch 41: val_mae improved from 0.02771 to 0.02758, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 3.4112e-04 - mae: 0.0128 - val_loss: 0.0013 - val_mae: 0.0276 - 4s/epoch - 25ms/step\n",
      "Epoch 42/3000\n",
      "\n",
      "Epoch 42: val_mae did not improve from 0.02758\n",
      "151/151 - 4s - loss: 3.3465e-04 - mae: 0.0127 - val_loss: 0.0013 - val_mae: 0.0276 - 4s/epoch - 24ms/step\n",
      "Epoch 43/3000\n",
      "\n",
      "Epoch 43: val_mae did not improve from 0.02758\n",
      "151/151 - 4s - loss: 3.2938e-04 - mae: 0.0126 - val_loss: 0.0013 - val_mae: 0.0277 - 4s/epoch - 24ms/step\n",
      "Epoch 44/3000\n",
      "\n",
      "Epoch 44: val_mae did not improve from 0.02758\n",
      "151/151 - 4s - loss: 3.2541e-04 - mae: 0.0125 - val_loss: 0.0013 - val_mae: 0.0281 - 4s/epoch - 23ms/step\n",
      "Epoch 45/3000\n",
      "\n",
      "Epoch 45: val_mae improved from 0.02758 to 0.02755, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 3.2048e-04 - mae: 0.0124 - val_loss: 0.0012 - val_mae: 0.0276 - 4s/epoch - 24ms/step\n",
      "Epoch 46/3000\n",
      "\n",
      "Epoch 46: val_mae improved from 0.02755 to 0.02700, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 3.1633e-04 - mae: 0.0123 - val_loss: 0.0012 - val_mae: 0.0270 - 4s/epoch - 24ms/step\n",
      "Epoch 47/3000\n",
      "\n",
      "Epoch 47: val_mae did not improve from 0.02700\n",
      "151/151 - 3s - loss: 3.1339e-04 - mae: 0.0122 - val_loss: 0.0013 - val_mae: 0.0278 - 3s/epoch - 23ms/step\n",
      "Epoch 48/3000\n",
      "\n",
      "Epoch 48: val_mae did not improve from 0.02700\n",
      "151/151 - 3s - loss: 3.0859e-04 - mae: 0.0121 - val_loss: 0.0013 - val_mae: 0.0281 - 3s/epoch - 23ms/step\n",
      "Epoch 49/3000\n",
      "\n",
      "Epoch 49: val_mae did not improve from 0.02700\n",
      "151/151 - 4s - loss: 3.0546e-04 - mae: 0.0120 - val_loss: 0.0012 - val_mae: 0.0273 - 4s/epoch - 23ms/step\n",
      "Epoch 50/3000\n",
      "\n",
      "Epoch 50: val_mae improved from 0.02700 to 0.02665, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 3.0109e-04 - mae: 0.0119 - val_loss: 0.0012 - val_mae: 0.0266 - 4s/epoch - 24ms/step\n",
      "Epoch 51/3000\n",
      "\n",
      "Epoch 51: val_mae did not improve from 0.02665\n",
      "151/151 - 3s - loss: 2.9816e-04 - mae: 0.0118 - val_loss: 0.0012 - val_mae: 0.0268 - 3s/epoch - 23ms/step\n",
      "Epoch 52/3000\n",
      "\n",
      "Epoch 52: val_mae improved from 0.02665 to 0.02607, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.9419e-04 - mae: 0.0117 - val_loss: 0.0011 - val_mae: 0.0261 - 4s/epoch - 24ms/step\n",
      "Epoch 53/3000\n",
      "\n",
      "Epoch 53: val_mae did not improve from 0.02607\n",
      "151/151 - 4s - loss: 2.9194e-04 - mae: 0.0117 - val_loss: 0.0012 - val_mae: 0.0262 - 4s/epoch - 24ms/step\n",
      "Epoch 54/3000\n",
      "\n",
      "Epoch 54: val_mae did not improve from 0.02607\n",
      "151/151 - 4s - loss: 2.8879e-04 - mae: 0.0116 - val_loss: 0.0012 - val_mae: 0.0270 - 4s/epoch - 24ms/step\n",
      "Epoch 55/3000\n",
      "\n",
      "Epoch 55: val_mae did not improve from 0.02607\n",
      "151/151 - 4s - loss: 2.8586e-04 - mae: 0.0115 - val_loss: 0.0012 - val_mae: 0.0264 - 4s/epoch - 23ms/step\n",
      "Epoch 56/3000\n",
      "\n",
      "Epoch 56: val_mae did not improve from 0.02607\n",
      "151/151 - 3s - loss: 2.8235e-04 - mae: 0.0114 - val_loss: 0.0012 - val_mae: 0.0263 - 3s/epoch - 23ms/step\n",
      "Epoch 57/3000\n",
      "\n",
      "Epoch 57: val_mae improved from 0.02607 to 0.02514, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 2.7957e-04 - mae: 0.0114 - val_loss: 0.0011 - val_mae: 0.0251 - 3s/epoch - 23ms/step\n",
      "Epoch 58/3000\n",
      "\n",
      "Epoch 58: val_mae did not improve from 0.02514\n",
      "151/151 - 3s - loss: 2.7637e-04 - mae: 0.0113 - val_loss: 0.0011 - val_mae: 0.0258 - 3s/epoch - 23ms/step\n",
      "Epoch 59/3000\n",
      "\n",
      "Epoch 59: val_mae did not improve from 0.02514\n",
      "151/151 - 3s - loss: 2.7487e-04 - mae: 0.0113 - val_loss: 0.0012 - val_mae: 0.0264 - 3s/epoch - 23ms/step\n",
      "Epoch 60/3000\n",
      "\n",
      "Epoch 60: val_mae improved from 0.02514 to 0.02487, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 2.7203e-04 - mae: 0.0112 - val_loss: 0.0010 - val_mae: 0.0249 - 3s/epoch - 23ms/step\n",
      "Epoch 61/3000\n",
      "\n",
      "Epoch 61: val_mae did not improve from 0.02487\n",
      "151/151 - 4s - loss: 2.6967e-04 - mae: 0.0111 - val_loss: 0.0011 - val_mae: 0.0255 - 4s/epoch - 24ms/step\n",
      "Epoch 62/3000\n",
      "\n",
      "Epoch 62: val_mae did not improve from 0.02487\n",
      "151/151 - 4s - loss: 2.6823e-04 - mae: 0.0111 - val_loss: 0.0010 - val_mae: 0.0250 - 4s/epoch - 24ms/step\n",
      "Epoch 63/3000\n",
      "\n",
      "Epoch 63: val_mae did not improve from 0.02487\n",
      "151/151 - 3s - loss: 2.6487e-04 - mae: 0.0110 - val_loss: 0.0011 - val_mae: 0.0254 - 3s/epoch - 23ms/step\n",
      "Epoch 64/3000\n",
      "\n",
      "Epoch 64: val_mae did not improve from 0.02487\n",
      "151/151 - 3s - loss: 2.6225e-04 - mae: 0.0109 - val_loss: 0.0011 - val_mae: 0.0254 - 3s/epoch - 23ms/step\n",
      "Epoch 65/3000\n",
      "\n",
      "Epoch 65: val_mae did not improve from 0.02487\n",
      "151/151 - 4s - loss: 2.6024e-04 - mae: 0.0109 - val_loss: 0.0011 - val_mae: 0.0259 - 4s/epoch - 23ms/step\n",
      "Epoch 66/3000\n",
      "\n",
      "Epoch 66: val_mae did not improve from 0.02487\n",
      "151/151 - 4s - loss: 2.5798e-04 - mae: 0.0108 - val_loss: 0.0011 - val_mae: 0.0256 - 4s/epoch - 25ms/step\n",
      "Epoch 67/3000\n",
      "\n",
      "Epoch 67: val_mae improved from 0.02487 to 0.02471, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.5557e-04 - mae: 0.0108 - val_loss: 0.0010 - val_mae: 0.0247 - 4s/epoch - 24ms/step\n",
      "Epoch 68/3000\n",
      "\n",
      "Epoch 68: val_mae improved from 0.02471 to 0.02409, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.5461e-04 - mae: 0.0108 - val_loss: 9.7821e-04 - val_mae: 0.0241 - 4s/epoch - 23ms/step\n",
      "Epoch 69/3000\n",
      "\n",
      "Epoch 69: val_mae improved from 0.02409 to 0.02402, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.5166e-04 - mae: 0.0107 - val_loss: 9.4926e-04 - val_mae: 0.0240 - 4s/epoch - 24ms/step\n",
      "Epoch 70/3000\n",
      "\n",
      "Epoch 70: val_mae improved from 0.02402 to 0.02347, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.5123e-04 - mae: 0.0107 - val_loss: 9.3555e-04 - val_mae: 0.0235 - 4s/epoch - 24ms/step\n",
      "Epoch 71/3000\n",
      "\n",
      "Epoch 71: val_mae improved from 0.02347 to 0.02337, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.4792e-04 - mae: 0.0106 - val_loss: 9.2853e-04 - val_mae: 0.0234 - 4s/epoch - 24ms/step\n",
      "Epoch 72/3000\n",
      "\n",
      "Epoch 72: val_mae did not improve from 0.02337\n",
      "151/151 - 4s - loss: 2.4653e-04 - mae: 0.0105 - val_loss: 0.0010 - val_mae: 0.0246 - 4s/epoch - 24ms/step\n",
      "Epoch 73/3000\n",
      "\n",
      "Epoch 73: val_mae did not improve from 0.02337\n",
      "151/151 - 4s - loss: 2.4438e-04 - mae: 0.0105 - val_loss: 9.2934e-04 - val_mae: 0.0236 - 4s/epoch - 24ms/step\n",
      "Epoch 74/3000\n",
      "\n",
      "Epoch 74: val_mae did not improve from 0.02337\n",
      "151/151 - 3s - loss: 2.4333e-04 - mae: 0.0105 - val_loss: 9.6667e-04 - val_mae: 0.0239 - 3s/epoch - 23ms/step\n",
      "Epoch 75/3000\n",
      "\n",
      "Epoch 75: val_mae did not improve from 0.02337\n",
      "151/151 - 3s - loss: 2.4117e-04 - mae: 0.0104 - val_loss: 9.7420e-04 - val_mae: 0.0240 - 3s/epoch - 23ms/step\n",
      "Epoch 76/3000\n",
      "\n",
      "Epoch 76: val_mae did not improve from 0.02337\n",
      "151/151 - 3s - loss: 2.3954e-04 - mae: 0.0104 - val_loss: 9.7787e-04 - val_mae: 0.0241 - 3s/epoch - 22ms/step\n",
      "Epoch 77/3000\n",
      "\n",
      "Epoch 77: val_mae did not improve from 0.02337\n",
      "151/151 - 3s - loss: 2.3752e-04 - mae: 0.0103 - val_loss: 9.4916e-04 - val_mae: 0.0234 - 3s/epoch - 23ms/step\n",
      "Epoch 78/3000\n",
      "\n",
      "Epoch 78: val_mae did not improve from 0.02337\n",
      "151/151 - 3s - loss: 2.3606e-04 - mae: 0.0103 - val_loss: 9.2852e-04 - val_mae: 0.0234 - 3s/epoch - 23ms/step\n",
      "Epoch 79/3000\n",
      "\n",
      "Epoch 79: val_mae did not improve from 0.02337\n",
      "151/151 - 3s - loss: 2.3478e-04 - mae: 0.0102 - val_loss: 9.5662e-04 - val_mae: 0.0235 - 3s/epoch - 23ms/step\n",
      "Epoch 80/3000\n",
      "\n",
      "Epoch 80: val_mae did not improve from 0.02337\n",
      "151/151 - 3s - loss: 2.3314e-04 - mae: 0.0102 - val_loss: 9.3557e-04 - val_mae: 0.0234 - 3s/epoch - 23ms/step\n",
      "Epoch 81/3000\n",
      "\n",
      "Epoch 81: val_mae improved from 0.02337 to 0.02301, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.3111e-04 - mae: 0.0101 - val_loss: 8.9946e-04 - val_mae: 0.0230 - 4s/epoch - 24ms/step\n",
      "Epoch 82/3000\n",
      "\n",
      "Epoch 82: val_mae improved from 0.02301 to 0.02252, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.2992e-04 - mae: 0.0101 - val_loss: 8.6123e-04 - val_mae: 0.0225 - 4s/epoch - 24ms/step\n",
      "Epoch 83/3000\n",
      "\n",
      "Epoch 83: val_mae improved from 0.02252 to 0.02250, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.2832e-04 - mae: 0.0100 - val_loss: 8.5221e-04 - val_mae: 0.0225 - 4s/epoch - 24ms/step\n",
      "Epoch 84/3000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 84: val_mae did not improve from 0.02250\n",
      "151/151 - 3s - loss: 2.2768e-04 - mae: 0.0100 - val_loss: 9.2875e-04 - val_mae: 0.0232 - 3s/epoch - 23ms/step\n",
      "Epoch 85/3000\n",
      "\n",
      "Epoch 85: val_mae did not improve from 0.02250\n",
      "151/151 - 3s - loss: 2.2546e-04 - mae: 0.0100 - val_loss: 8.9412e-04 - val_mae: 0.0227 - 3s/epoch - 23ms/step\n",
      "Epoch 86/3000\n",
      "\n",
      "Epoch 86: val_mae improved from 0.02250 to 0.02232, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.2315e-04 - mae: 0.0099 - val_loss: 8.7233e-04 - val_mae: 0.0223 - 4s/epoch - 24ms/step\n",
      "Epoch 87/3000\n",
      "\n",
      "Epoch 87: val_mae did not improve from 0.02232\n",
      "151/151 - 4s - loss: 2.2392e-04 - mae: 0.0099 - val_loss: 8.4606e-04 - val_mae: 0.0223 - 4s/epoch - 24ms/step\n",
      "Epoch 88/3000\n",
      "\n",
      "Epoch 88: val_mae improved from 0.02232 to 0.02203, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.2221e-04 - mae: 0.0099 - val_loss: 8.4928e-04 - val_mae: 0.0220 - 4s/epoch - 24ms/step\n",
      "Epoch 89/3000\n",
      "\n",
      "Epoch 89: val_mae did not improve from 0.02203\n",
      "151/151 - 4s - loss: 2.2051e-04 - mae: 0.0098 - val_loss: 8.9428e-04 - val_mae: 0.0227 - 4s/epoch - 24ms/step\n",
      "Epoch 90/3000\n",
      "\n",
      "Epoch 90: val_mae improved from 0.02203 to 0.02158, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.1855e-04 - mae: 0.0098 - val_loss: 8.0135e-04 - val_mae: 0.0216 - 4s/epoch - 24ms/step\n",
      "Epoch 91/3000\n",
      "\n",
      "Epoch 91: val_mae did not improve from 0.02158\n",
      "151/151 - 4s - loss: 2.1849e-04 - mae: 0.0098 - val_loss: 8.2447e-04 - val_mae: 0.0218 - 4s/epoch - 24ms/step\n",
      "Epoch 92/3000\n",
      "\n",
      "Epoch 92: val_mae did not improve from 0.02158\n",
      "151/151 - 4s - loss: 2.1650e-04 - mae: 0.0097 - val_loss: 8.4785e-04 - val_mae: 0.0221 - 4s/epoch - 24ms/step\n",
      "Epoch 93/3000\n",
      "\n",
      "Epoch 93: val_mae did not improve from 0.02158\n",
      "151/151 - 4s - loss: 2.1513e-04 - mae: 0.0097 - val_loss: 8.1155e-04 - val_mae: 0.0217 - 4s/epoch - 24ms/step\n",
      "Epoch 94/3000\n",
      "\n",
      "Epoch 94: val_mae did not improve from 0.02158\n",
      "151/151 - 4s - loss: 2.1537e-04 - mae: 0.0097 - val_loss: 8.3868e-04 - val_mae: 0.0218 - 4s/epoch - 24ms/step\n",
      "Epoch 95/3000\n",
      "\n",
      "Epoch 95: val_mae did not improve from 0.02158\n",
      "151/151 - 3s - loss: 2.1295e-04 - mae: 0.0096 - val_loss: 8.3383e-04 - val_mae: 0.0223 - 3s/epoch - 23ms/step\n",
      "Epoch 96/3000\n",
      "\n",
      "Epoch 96: val_mae improved from 0.02158 to 0.02133, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.1311e-04 - mae: 0.0096 - val_loss: 7.8039e-04 - val_mae: 0.0213 - 4s/epoch - 23ms/step\n",
      "Epoch 97/3000\n",
      "\n",
      "Epoch 97: val_mae did not improve from 0.02133\n",
      "151/151 - 4s - loss: 2.1134e-04 - mae: 0.0096 - val_loss: 8.3148e-04 - val_mae: 0.0219 - 4s/epoch - 24ms/step\n",
      "Epoch 98/3000\n",
      "\n",
      "Epoch 98: val_mae did not improve from 0.02133\n",
      "151/151 - 4s - loss: 2.1073e-04 - mae: 0.0096 - val_loss: 8.0121e-04 - val_mae: 0.0216 - 4s/epoch - 25ms/step\n",
      "Epoch 99/3000\n",
      "\n",
      "Epoch 99: val_mae improved from 0.02133 to 0.02094, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.0846e-04 - mae: 0.0095 - val_loss: 7.5307e-04 - val_mae: 0.0209 - 4s/epoch - 24ms/step\n",
      "Epoch 100/3000\n",
      "\n",
      "Epoch 100: val_mae did not improve from 0.02094\n",
      "151/151 - 4s - loss: 2.0813e-04 - mae: 0.0095 - val_loss: 8.4580e-04 - val_mae: 0.0218 - 4s/epoch - 25ms/step\n",
      "Epoch 101/3000\n",
      "\n",
      "Epoch 101: val_mae did not improve from 0.02094\n",
      "151/151 - 4s - loss: 2.0818e-04 - mae: 0.0095 - val_loss: 8.0012e-04 - val_mae: 0.0212 - 4s/epoch - 24ms/step\n",
      "Epoch 102/3000\n",
      "\n",
      "Epoch 102: val_mae did not improve from 0.02094\n",
      "151/151 - 4s - loss: 2.0553e-04 - mae: 0.0094 - val_loss: 7.9648e-04 - val_mae: 0.0215 - 4s/epoch - 23ms/step\n",
      "Epoch 103/3000\n",
      "\n",
      "Epoch 103: val_mae did not improve from 0.02094\n",
      "151/151 - 3s - loss: 2.0465e-04 - mae: 0.0094 - val_loss: 8.0450e-04 - val_mae: 0.0213 - 3s/epoch - 23ms/step\n",
      "Epoch 104/3000\n",
      "\n",
      "Epoch 104: val_mae did not improve from 0.02094\n",
      "151/151 - 4s - loss: 2.0545e-04 - mae: 0.0094 - val_loss: 7.8346e-04 - val_mae: 0.0213 - 4s/epoch - 24ms/step\n",
      "Epoch 105/3000\n",
      "\n",
      "Epoch 105: val_mae improved from 0.02094 to 0.02042, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 2.0374e-04 - mae: 0.0094 - val_loss: 7.1760e-04 - val_mae: 0.0204 - 4s/epoch - 24ms/step\n",
      "Epoch 106/3000\n",
      "\n",
      "Epoch 106: val_mae did not improve from 0.02042\n",
      "151/151 - 4s - loss: 2.0237e-04 - mae: 0.0093 - val_loss: 7.3734e-04 - val_mae: 0.0205 - 4s/epoch - 24ms/step\n",
      "Epoch 107/3000\n",
      "\n",
      "Epoch 107: val_mae did not improve from 0.02042\n",
      "151/151 - 3s - loss: 2.0258e-04 - mae: 0.0093 - val_loss: 7.3052e-04 - val_mae: 0.0206 - 3s/epoch - 23ms/step\n",
      "Epoch 108/3000\n",
      "\n",
      "Epoch 108: val_mae did not improve from 0.02042\n",
      "151/151 - 3s - loss: 2.0044e-04 - mae: 0.0092 - val_loss: 7.8736e-04 - val_mae: 0.0215 - 3s/epoch - 23ms/step\n",
      "Epoch 109/3000\n",
      "\n",
      "Epoch 109: val_mae did not improve from 0.02042\n",
      "151/151 - 3s - loss: 1.9978e-04 - mae: 0.0092 - val_loss: 7.9563e-04 - val_mae: 0.0214 - 3s/epoch - 22ms/step\n",
      "Epoch 110/3000\n",
      "\n",
      "Epoch 110: val_mae improved from 0.02042 to 0.02038, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 1.9889e-04 - mae: 0.0092 - val_loss: 7.2348e-04 - val_mae: 0.0204 - 3s/epoch - 22ms/step\n",
      "Epoch 111/3000\n",
      "\n",
      "Epoch 111: val_mae did not improve from 0.02038\n",
      "151/151 - 3s - loss: 1.9837e-04 - mae: 0.0092 - val_loss: 7.4846e-04 - val_mae: 0.0206 - 3s/epoch - 23ms/step\n",
      "Epoch 112/3000\n",
      "\n",
      "Epoch 112: val_mae did not improve from 0.02038\n",
      "151/151 - 3s - loss: 1.9786e-04 - mae: 0.0092 - val_loss: 7.3642e-04 - val_mae: 0.0204 - 3s/epoch - 23ms/step\n",
      "Epoch 113/3000\n",
      "\n",
      "Epoch 113: val_mae did not improve from 0.02038\n",
      "151/151 - 3s - loss: 1.9669e-04 - mae: 0.0091 - val_loss: 7.3828e-04 - val_mae: 0.0208 - 3s/epoch - 23ms/step\n",
      "Epoch 114/3000\n",
      "\n",
      "Epoch 114: val_mae improved from 0.02038 to 0.02038, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.9616e-04 - mae: 0.0091 - val_loss: 7.1368e-04 - val_mae: 0.0204 - 4s/epoch - 24ms/step\n",
      "Epoch 115/3000\n",
      "\n",
      "Epoch 115: val_mae improved from 0.02038 to 0.02028, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.9475e-04 - mae: 0.0091 - val_loss: 7.2813e-04 - val_mae: 0.0203 - 4s/epoch - 23ms/step\n",
      "Epoch 116/3000\n",
      "\n",
      "Epoch 116: val_mae improved from 0.02028 to 0.02000, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 1.9425e-04 - mae: 0.0091 - val_loss: 6.9912e-04 - val_mae: 0.0200 - 3s/epoch - 23ms/step\n",
      "Epoch 117/3000\n",
      "\n",
      "Epoch 117: val_mae did not improve from 0.02000\n",
      "151/151 - 3s - loss: 1.9528e-04 - mae: 0.0091 - val_loss: 7.2900e-04 - val_mae: 0.0203 - 3s/epoch - 23ms/step\n",
      "Epoch 118/3000\n",
      "\n",
      "Epoch 118: val_mae did not improve from 0.02000\n",
      "151/151 - 3s - loss: 1.9412e-04 - mae: 0.0091 - val_loss: 7.1980e-04 - val_mae: 0.0203 - 3s/epoch - 22ms/step\n",
      "Epoch 119/3000\n",
      "\n",
      "Epoch 119: val_mae did not improve from 0.02000\n",
      "151/151 - 4s - loss: 1.9181e-04 - mae: 0.0090 - val_loss: 7.0406e-04 - val_mae: 0.0200 - 4s/epoch - 24ms/step\n",
      "Epoch 120/3000\n",
      "\n",
      "Epoch 120: val_mae did not improve from 0.02000\n",
      "151/151 - 4s - loss: 1.9130e-04 - mae: 0.0090 - val_loss: 7.2073e-04 - val_mae: 0.0201 - 4s/epoch - 24ms/step\n",
      "Epoch 121/3000\n",
      "\n",
      "Epoch 121: val_mae improved from 0.02000 to 0.01966, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.9065e-04 - mae: 0.0090 - val_loss: 6.6055e-04 - val_mae: 0.0197 - 4s/epoch - 24ms/step\n",
      "Epoch 122/3000\n",
      "\n",
      "Epoch 122: val_mae did not improve from 0.01966\n",
      "151/151 - 4s - loss: 1.9028e-04 - mae: 0.0090 - val_loss: 7.0561e-04 - val_mae: 0.0201 - 4s/epoch - 24ms/step\n",
      "Epoch 123/3000\n",
      "\n",
      "Epoch 123: val_mae did not improve from 0.01966\n",
      "151/151 - 4s - loss: 1.8975e-04 - mae: 0.0089 - val_loss: 7.1967e-04 - val_mae: 0.0201 - 4s/epoch - 24ms/step\n",
      "Epoch 124/3000\n",
      "\n",
      "Epoch 124: val_mae did not improve from 0.01966\n",
      "151/151 - 4s - loss: 1.8984e-04 - mae: 0.0090 - val_loss: 7.0767e-04 - val_mae: 0.0199 - 4s/epoch - 23ms/step\n",
      "Epoch 125/3000\n",
      "\n",
      "Epoch 125: val_mae did not improve from 0.01966\n",
      "151/151 - 3s - loss: 1.8887e-04 - mae: 0.0089 - val_loss: 6.8250e-04 - val_mae: 0.0197 - 3s/epoch - 23ms/step\n",
      "Epoch 126/3000\n",
      "\n",
      "Epoch 126: val_mae did not improve from 0.01966\n",
      "151/151 - 4s - loss: 1.8817e-04 - mae: 0.0089 - val_loss: 6.8432e-04 - val_mae: 0.0198 - 4s/epoch - 24ms/step\n",
      "Epoch 127/3000\n",
      "\n",
      "Epoch 127: val_mae improved from 0.01966 to 0.01955, saving model to saved_models/gcn_gate_pump_nofuture.h5\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "151/151 - 4s - loss: 1.8727e-04 - mae: 0.0089 - val_loss: 6.6341e-04 - val_mae: 0.0195 - 4s/epoch - 24ms/step\n",
      "Epoch 128/3000\n",
      "\n",
      "Epoch 128: val_mae improved from 0.01955 to 0.01953, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.8617e-04 - mae: 0.0088 - val_loss: 6.7082e-04 - val_mae: 0.0195 - 4s/epoch - 24ms/step\n",
      "Epoch 129/3000\n",
      "\n",
      "Epoch 129: val_mae did not improve from 0.01953\n",
      "151/151 - 4s - loss: 1.8604e-04 - mae: 0.0088 - val_loss: 6.8928e-04 - val_mae: 0.0197 - 4s/epoch - 24ms/step\n",
      "Epoch 130/3000\n",
      "\n",
      "Epoch 130: val_mae did not improve from 0.01953\n",
      "151/151 - 4s - loss: 1.8541e-04 - mae: 0.0088 - val_loss: 6.7529e-04 - val_mae: 0.0197 - 4s/epoch - 24ms/step\n",
      "Epoch 131/3000\n",
      "\n",
      "Epoch 131: val_mae did not improve from 0.01953\n",
      "151/151 - 4s - loss: 1.8561e-04 - mae: 0.0088 - val_loss: 6.8224e-04 - val_mae: 0.0197 - 4s/epoch - 24ms/step\n",
      "Epoch 132/3000\n",
      "\n",
      "Epoch 132: val_mae improved from 0.01953 to 0.01906, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.8503e-04 - mae: 0.0088 - val_loss: 6.3548e-04 - val_mae: 0.0191 - 4s/epoch - 25ms/step\n",
      "Epoch 133/3000\n",
      "\n",
      "Epoch 133: val_mae improved from 0.01906 to 0.01893, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.8289e-04 - mae: 0.0087 - val_loss: 6.2791e-04 - val_mae: 0.0189 - 4s/epoch - 24ms/step\n",
      "Epoch 134/3000\n",
      "\n",
      "Epoch 134: val_mae did not improve from 0.01893\n",
      "151/151 - 4s - loss: 1.8378e-04 - mae: 0.0088 - val_loss: 6.7386e-04 - val_mae: 0.0196 - 4s/epoch - 24ms/step\n",
      "Epoch 135/3000\n",
      "\n",
      "Epoch 135: val_mae did not improve from 0.01893\n",
      "151/151 - 4s - loss: 1.8287e-04 - mae: 0.0087 - val_loss: 6.7221e-04 - val_mae: 0.0195 - 4s/epoch - 24ms/step\n",
      "Epoch 136/3000\n",
      "\n",
      "Epoch 136: val_mae did not improve from 0.01893\n",
      "151/151 - 4s - loss: 1.8205e-04 - mae: 0.0087 - val_loss: 7.0704e-04 - val_mae: 0.0201 - 4s/epoch - 23ms/step\n",
      "Epoch 137/3000\n",
      "\n",
      "Epoch 137: val_mae did not improve from 0.01893\n",
      "151/151 - 4s - loss: 1.8228e-04 - mae: 0.0087 - val_loss: 6.7666e-04 - val_mae: 0.0196 - 4s/epoch - 24ms/step\n",
      "Epoch 138/3000\n",
      "\n",
      "Epoch 138: val_mae improved from 0.01893 to 0.01858, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.8055e-04 - mae: 0.0087 - val_loss: 6.0932e-04 - val_mae: 0.0186 - 4s/epoch - 24ms/step\n",
      "Epoch 139/3000\n",
      "\n",
      "Epoch 139: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.8104e-04 - mae: 0.0087 - val_loss: 6.7292e-04 - val_mae: 0.0195 - 4s/epoch - 24ms/step\n",
      "Epoch 140/3000\n",
      "\n",
      "Epoch 140: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.8126e-04 - mae: 0.0087 - val_loss: 6.8419e-04 - val_mae: 0.0195 - 4s/epoch - 25ms/step\n",
      "Epoch 141/3000\n",
      "\n",
      "Epoch 141: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.8007e-04 - mae: 0.0087 - val_loss: 6.9493e-04 - val_mae: 0.0198 - 4s/epoch - 24ms/step\n",
      "Epoch 142/3000\n",
      "\n",
      "Epoch 142: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.7922e-04 - mae: 0.0086 - val_loss: 6.4658e-04 - val_mae: 0.0191 - 4s/epoch - 24ms/step\n",
      "Epoch 143/3000\n",
      "\n",
      "Epoch 143: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.7881e-04 - mae: 0.0086 - val_loss: 6.1674e-04 - val_mae: 0.0188 - 4s/epoch - 25ms/step\n",
      "Epoch 144/3000\n",
      "\n",
      "Epoch 144: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.7828e-04 - mae: 0.0086 - val_loss: 6.4701e-04 - val_mae: 0.0190 - 4s/epoch - 25ms/step\n",
      "Epoch 145/3000\n",
      "\n",
      "Epoch 145: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.7831e-04 - mae: 0.0086 - val_loss: 6.6351e-04 - val_mae: 0.0193 - 4s/epoch - 24ms/step\n",
      "Epoch 146/3000\n",
      "\n",
      "Epoch 146: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.7721e-04 - mae: 0.0086 - val_loss: 6.3621e-04 - val_mae: 0.0189 - 4s/epoch - 24ms/step\n",
      "Epoch 147/3000\n",
      "\n",
      "Epoch 147: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.7768e-04 - mae: 0.0086 - val_loss: 6.9188e-04 - val_mae: 0.0196 - 4s/epoch - 24ms/step\n",
      "Epoch 148/3000\n",
      "\n",
      "Epoch 148: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.7634e-04 - mae: 0.0086 - val_loss: 6.7650e-04 - val_mae: 0.0196 - 4s/epoch - 23ms/step\n",
      "Epoch 149/3000\n",
      "\n",
      "Epoch 149: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.7553e-04 - mae: 0.0085 - val_loss: 6.1482e-04 - val_mae: 0.0186 - 4s/epoch - 24ms/step\n",
      "Epoch 150/3000\n",
      "\n",
      "Epoch 150: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.7569e-04 - mae: 0.0085 - val_loss: 6.5039e-04 - val_mae: 0.0190 - 4s/epoch - 24ms/step\n",
      "Epoch 151/3000\n",
      "\n",
      "Epoch 151: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.7539e-04 - mae: 0.0085 - val_loss: 6.5012e-04 - val_mae: 0.0189 - 4s/epoch - 24ms/step\n",
      "Epoch 152/3000\n",
      "\n",
      "Epoch 152: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.7446e-04 - mae: 0.0085 - val_loss: 6.2048e-04 - val_mae: 0.0187 - 4s/epoch - 25ms/step\n",
      "Epoch 153/3000\n",
      "\n",
      "Epoch 153: val_mae did not improve from 0.01858\n",
      "151/151 - 4s - loss: 1.7417e-04 - mae: 0.0085 - val_loss: 6.4136e-04 - val_mae: 0.0189 - 4s/epoch - 24ms/step\n",
      "Epoch 154/3000\n",
      "\n",
      "Epoch 154: val_mae improved from 0.01858 to 0.01846, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.7384e-04 - mae: 0.0085 - val_loss: 6.0780e-04 - val_mae: 0.0185 - 4s/epoch - 24ms/step\n",
      "Epoch 155/3000\n",
      "\n",
      "Epoch 155: val_mae improved from 0.01846 to 0.01829, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.7404e-04 - mae: 0.0085 - val_loss: 5.9526e-04 - val_mae: 0.0183 - 4s/epoch - 25ms/step\n",
      "Epoch 156/3000\n",
      "\n",
      "Epoch 156: val_mae improved from 0.01829 to 0.01825, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.7381e-04 - mae: 0.0085 - val_loss: 5.9738e-04 - val_mae: 0.0183 - 4s/epoch - 24ms/step\n",
      "Epoch 157/3000\n",
      "\n",
      "Epoch 157: val_mae did not improve from 0.01825\n",
      "151/151 - 4s - loss: 1.7272e-04 - mae: 0.0084 - val_loss: 6.1843e-04 - val_mae: 0.0186 - 4s/epoch - 24ms/step\n",
      "Epoch 158/3000\n",
      "\n",
      "Epoch 158: val_mae did not improve from 0.01825\n",
      "151/151 - 4s - loss: 1.7312e-04 - mae: 0.0085 - val_loss: 6.1243e-04 - val_mae: 0.0185 - 4s/epoch - 23ms/step\n",
      "Epoch 159/3000\n",
      "\n",
      "Epoch 159: val_mae did not improve from 0.01825\n",
      "151/151 - 4s - loss: 1.7240e-04 - mae: 0.0084 - val_loss: 6.1330e-04 - val_mae: 0.0186 - 4s/epoch - 24ms/step\n",
      "Epoch 160/3000\n",
      "\n",
      "Epoch 160: val_mae did not improve from 0.01825\n",
      "151/151 - 4s - loss: 1.7152e-04 - mae: 0.0084 - val_loss: 6.2644e-04 - val_mae: 0.0187 - 4s/epoch - 23ms/step\n",
      "Epoch 161/3000\n",
      "\n",
      "Epoch 161: val_mae improved from 0.01825 to 0.01794, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.7157e-04 - mae: 0.0084 - val_loss: 5.7172e-04 - val_mae: 0.0179 - 4s/epoch - 24ms/step\n",
      "Epoch 162/3000\n",
      "\n",
      "Epoch 162: val_mae did not improve from 0.01794\n",
      "151/151 - 4s - loss: 1.7062e-04 - mae: 0.0084 - val_loss: 6.4368e-04 - val_mae: 0.0190 - 4s/epoch - 24ms/step\n",
      "Epoch 163/3000\n",
      "\n",
      "Epoch 163: val_mae did not improve from 0.01794\n",
      "151/151 - 4s - loss: 1.7047e-04 - mae: 0.0084 - val_loss: 5.9519e-04 - val_mae: 0.0183 - 4s/epoch - 24ms/step\n",
      "Epoch 164/3000\n",
      "\n",
      "Epoch 164: val_mae did not improve from 0.01794\n",
      "151/151 - 4s - loss: 1.7101e-04 - mae: 0.0084 - val_loss: 5.9758e-04 - val_mae: 0.0183 - 4s/epoch - 24ms/step\n",
      "Epoch 165/3000\n",
      "\n",
      "Epoch 165: val_mae did not improve from 0.01794\n",
      "151/151 - 4s - loss: 1.6988e-04 - mae: 0.0084 - val_loss: 5.8901e-04 - val_mae: 0.0180 - 4s/epoch - 24ms/step\n",
      "Epoch 166/3000\n",
      "\n",
      "Epoch 166: val_mae did not improve from 0.01794\n",
      "151/151 - 4s - loss: 1.6987e-04 - mae: 0.0084 - val_loss: 6.0458e-04 - val_mae: 0.0183 - 4s/epoch - 24ms/step\n",
      "Epoch 167/3000\n",
      "\n",
      "Epoch 167: val_mae did not improve from 0.01794\n",
      "151/151 - 4s - loss: 1.6886e-04 - mae: 0.0083 - val_loss: 6.1897e-04 - val_mae: 0.0188 - 4s/epoch - 24ms/step\n",
      "Epoch 168/3000\n",
      "\n",
      "Epoch 168: val_mae did not improve from 0.01794\n",
      "151/151 - 3s - loss: 1.6961e-04 - mae: 0.0084 - val_loss: 6.0422e-04 - val_mae: 0.0183 - 3s/epoch - 23ms/step\n",
      "Epoch 169/3000\n",
      "\n",
      "Epoch 169: val_mae improved from 0.01794 to 0.01783, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.6743e-04 - mae: 0.0083 - val_loss: 5.6098e-04 - val_mae: 0.0178 - 4s/epoch - 24ms/step\n",
      "Epoch 170/3000\n",
      "\n",
      "Epoch 170: val_mae did not improve from 0.01783\n",
      "151/151 - 4s - loss: 1.6877e-04 - mae: 0.0083 - val_loss: 5.8539e-04 - val_mae: 0.0181 - 4s/epoch - 23ms/step\n",
      "Epoch 171/3000\n",
      "\n",
      "Epoch 171: val_mae did not improve from 0.01783\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "151/151 - 4s - loss: 1.6826e-04 - mae: 0.0083 - val_loss: 6.0216e-04 - val_mae: 0.0182 - 4s/epoch - 24ms/step\n",
      "Epoch 172/3000\n",
      "\n",
      "Epoch 172: val_mae did not improve from 0.01783\n",
      "151/151 - 4s - loss: 1.6709e-04 - mae: 0.0083 - val_loss: 5.8862e-04 - val_mae: 0.0180 - 4s/epoch - 24ms/step\n",
      "Epoch 173/3000\n",
      "\n",
      "Epoch 173: val_mae did not improve from 0.01783\n",
      "151/151 - 4s - loss: 1.6703e-04 - mae: 0.0083 - val_loss: 5.8670e-04 - val_mae: 0.0181 - 4s/epoch - 24ms/step\n",
      "Epoch 174/3000\n",
      "\n",
      "Epoch 174: val_mae improved from 0.01783 to 0.01768, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.6735e-04 - mae: 0.0083 - val_loss: 5.6127e-04 - val_mae: 0.0177 - 4s/epoch - 24ms/step\n",
      "Epoch 175/3000\n",
      "\n",
      "Epoch 175: val_mae did not improve from 0.01768\n",
      "151/151 - 4s - loss: 1.6679e-04 - mae: 0.0083 - val_loss: 5.9680e-04 - val_mae: 0.0183 - 4s/epoch - 24ms/step\n",
      "Epoch 176/3000\n",
      "\n",
      "Epoch 176: val_mae did not improve from 0.01768\n",
      "151/151 - 4s - loss: 1.6653e-04 - mae: 0.0083 - val_loss: 5.8627e-04 - val_mae: 0.0181 - 4s/epoch - 24ms/step\n",
      "Epoch 177/3000\n",
      "\n",
      "Epoch 177: val_mae did not improve from 0.01768\n",
      "151/151 - 4s - loss: 1.6622e-04 - mae: 0.0082 - val_loss: 5.9976e-04 - val_mae: 0.0183 - 4s/epoch - 24ms/step\n",
      "Epoch 178/3000\n",
      "\n",
      "Epoch 178: val_mae did not improve from 0.01768\n",
      "151/151 - 4s - loss: 1.6551e-04 - mae: 0.0082 - val_loss: 6.1070e-04 - val_mae: 0.0185 - 4s/epoch - 24ms/step\n",
      "Epoch 179/3000\n",
      "\n",
      "Epoch 179: val_mae did not improve from 0.01768\n",
      "151/151 - 4s - loss: 1.6576e-04 - mae: 0.0082 - val_loss: 5.9321e-04 - val_mae: 0.0181 - 4s/epoch - 24ms/step\n",
      "Epoch 180/3000\n",
      "\n",
      "Epoch 180: val_mae did not improve from 0.01768\n",
      "151/151 - 4s - loss: 1.6515e-04 - mae: 0.0082 - val_loss: 5.7690e-04 - val_mae: 0.0179 - 4s/epoch - 23ms/step\n",
      "Epoch 181/3000\n",
      "\n",
      "Epoch 181: val_mae improved from 0.01768 to 0.01759, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.6529e-04 - mae: 0.0082 - val_loss: 5.6075e-04 - val_mae: 0.0176 - 4s/epoch - 23ms/step\n",
      "Epoch 182/3000\n",
      "\n",
      "Epoch 182: val_mae did not improve from 0.01759\n",
      "151/151 - 3s - loss: 1.6448e-04 - mae: 0.0082 - val_loss: 5.7956e-04 - val_mae: 0.0179 - 3s/epoch - 23ms/step\n",
      "Epoch 183/3000\n",
      "\n",
      "Epoch 183: val_mae did not improve from 0.01759\n",
      "151/151 - 4s - loss: 1.6438e-04 - mae: 0.0082 - val_loss: 5.7735e-04 - val_mae: 0.0179 - 4s/epoch - 25ms/step\n",
      "Epoch 184/3000\n",
      "\n",
      "Epoch 184: val_mae improved from 0.01759 to 0.01759, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.6438e-04 - mae: 0.0082 - val_loss: 5.5518e-04 - val_mae: 0.0176 - 4s/epoch - 25ms/step\n",
      "Epoch 185/3000\n",
      "\n",
      "Epoch 185: val_mae did not improve from 0.01759\n",
      "151/151 - 4s - loss: 1.6358e-04 - mae: 0.0082 - val_loss: 5.8087e-04 - val_mae: 0.0179 - 4s/epoch - 24ms/step\n",
      "Epoch 186/3000\n",
      "\n",
      "Epoch 186: val_mae did not improve from 0.01759\n",
      "151/151 - 4s - loss: 1.6401e-04 - mae: 0.0082 - val_loss: 5.5980e-04 - val_mae: 0.0177 - 4s/epoch - 24ms/step\n",
      "Epoch 187/3000\n",
      "\n",
      "Epoch 187: val_mae did not improve from 0.01759\n",
      "151/151 - 4s - loss: 1.6285e-04 - mae: 0.0081 - val_loss: 5.6112e-04 - val_mae: 0.0177 - 4s/epoch - 24ms/step\n",
      "Epoch 188/3000\n",
      "\n",
      "Epoch 188: val_mae did not improve from 0.01759\n",
      "151/151 - 4s - loss: 1.6301e-04 - mae: 0.0081 - val_loss: 5.9049e-04 - val_mae: 0.0181 - 4s/epoch - 24ms/step\n",
      "Epoch 189/3000\n",
      "\n",
      "Epoch 189: val_mae improved from 0.01759 to 0.01747, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.6255e-04 - mae: 0.0081 - val_loss: 5.5413e-04 - val_mae: 0.0175 - 4s/epoch - 24ms/step\n",
      "Epoch 190/3000\n",
      "\n",
      "Epoch 190: val_mae did not improve from 0.01747\n",
      "151/151 - 4s - loss: 1.6216e-04 - mae: 0.0081 - val_loss: 5.6272e-04 - val_mae: 0.0177 - 4s/epoch - 24ms/step\n",
      "Epoch 191/3000\n",
      "\n",
      "Epoch 191: val_mae did not improve from 0.01747\n",
      "151/151 - 4s - loss: 1.6164e-04 - mae: 0.0081 - val_loss: 5.5525e-04 - val_mae: 0.0177 - 4s/epoch - 24ms/step\n",
      "Epoch 192/3000\n",
      "\n",
      "Epoch 192: val_mae improved from 0.01747 to 0.01743, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.6246e-04 - mae: 0.0081 - val_loss: 5.4498e-04 - val_mae: 0.0174 - 4s/epoch - 24ms/step\n",
      "Epoch 193/3000\n",
      "\n",
      "Epoch 193: val_mae did not improve from 0.01743\n",
      "151/151 - 4s - loss: 1.6150e-04 - mae: 0.0081 - val_loss: 5.6240e-04 - val_mae: 0.0178 - 4s/epoch - 23ms/step\n",
      "Epoch 194/3000\n",
      "\n",
      "Epoch 194: val_mae improved from 0.01743 to 0.01741, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.6135e-04 - mae: 0.0081 - val_loss: 5.4817e-04 - val_mae: 0.0174 - 4s/epoch - 24ms/step\n",
      "Epoch 195/3000\n",
      "\n",
      "Epoch 195: val_mae did not improve from 0.01741\n",
      "151/151 - 4s - loss: 1.6090e-04 - mae: 0.0081 - val_loss: 5.7991e-04 - val_mae: 0.0179 - 4s/epoch - 23ms/step\n",
      "Epoch 196/3000\n",
      "\n",
      "Epoch 196: val_mae did not improve from 0.01741\n",
      "151/151 - 4s - loss: 1.6087e-04 - mae: 0.0081 - val_loss: 5.7683e-04 - val_mae: 0.0179 - 4s/epoch - 24ms/step\n",
      "Epoch 197/3000\n",
      "\n",
      "Epoch 197: val_mae improved from 0.01741 to 0.01739, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.6055e-04 - mae: 0.0081 - val_loss: 5.4662e-04 - val_mae: 0.0174 - 4s/epoch - 24ms/step\n",
      "Epoch 198/3000\n",
      "\n",
      "Epoch 198: val_mae did not improve from 0.01739\n",
      "151/151 - 4s - loss: 1.6011e-04 - mae: 0.0080 - val_loss: 5.8842e-04 - val_mae: 0.0178 - 4s/epoch - 24ms/step\n",
      "Epoch 199/3000\n",
      "\n",
      "Epoch 199: val_mae did not improve from 0.01739\n",
      "151/151 - 4s - loss: 1.6009e-04 - mae: 0.0081 - val_loss: 5.7956e-04 - val_mae: 0.0177 - 4s/epoch - 23ms/step\n",
      "Epoch 200/3000\n",
      "\n",
      "Epoch 200: val_mae improved from 0.01739 to 0.01724, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.5986e-04 - mae: 0.0080 - val_loss: 5.4213e-04 - val_mae: 0.0172 - 4s/epoch - 24ms/step\n",
      "Epoch 201/3000\n",
      "\n",
      "Epoch 201: val_mae did not improve from 0.01724\n",
      "151/151 - 4s - loss: 1.5973e-04 - mae: 0.0080 - val_loss: 5.5469e-04 - val_mae: 0.0174 - 4s/epoch - 24ms/step\n",
      "Epoch 202/3000\n",
      "\n",
      "Epoch 202: val_mae improved from 0.01724 to 0.01707, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.5957e-04 - mae: 0.0080 - val_loss: 5.1307e-04 - val_mae: 0.0171 - 4s/epoch - 23ms/step\n",
      "Epoch 203/3000\n",
      "\n",
      "Epoch 203: val_mae improved from 0.01707 to 0.01693, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.5946e-04 - mae: 0.0080 - val_loss: 5.1780e-04 - val_mae: 0.0169 - 4s/epoch - 24ms/step\n",
      "Epoch 204/3000\n",
      "\n",
      "Epoch 204: val_mae did not improve from 0.01693\n",
      "151/151 - 4s - loss: 1.5823e-04 - mae: 0.0080 - val_loss: 5.4521e-04 - val_mae: 0.0172 - 4s/epoch - 25ms/step\n",
      "Epoch 205/3000\n",
      "\n",
      "Epoch 205: val_mae did not improve from 0.01693\n",
      "151/151 - 4s - loss: 1.5916e-04 - mae: 0.0080 - val_loss: 5.8835e-04 - val_mae: 0.0181 - 4s/epoch - 24ms/step\n",
      "Epoch 206/3000\n",
      "\n",
      "Epoch 206: val_mae did not improve from 0.01693\n",
      "151/151 - 4s - loss: 1.5865e-04 - mae: 0.0080 - val_loss: 5.5855e-04 - val_mae: 0.0176 - 4s/epoch - 24ms/step\n",
      "Epoch 207/3000\n",
      "\n",
      "Epoch 207: val_mae did not improve from 0.01693\n",
      "151/151 - 4s - loss: 1.5870e-04 - mae: 0.0080 - val_loss: 5.6463e-04 - val_mae: 0.0177 - 4s/epoch - 24ms/step\n",
      "Epoch 208/3000\n",
      "\n",
      "Epoch 208: val_mae did not improve from 0.01693\n",
      "151/151 - 4s - loss: 1.5813e-04 - mae: 0.0080 - val_loss: 5.6607e-04 - val_mae: 0.0176 - 4s/epoch - 25ms/step\n",
      "Epoch 209/3000\n",
      "\n",
      "Epoch 209: val_mae did not improve from 0.01693\n",
      "151/151 - 4s - loss: 1.5825e-04 - mae: 0.0080 - val_loss: 5.2863e-04 - val_mae: 0.0170 - 4s/epoch - 23ms/step\n",
      "Epoch 210/3000\n",
      "\n",
      "Epoch 210: val_mae did not improve from 0.01693\n",
      "151/151 - 4s - loss: 1.5786e-04 - mae: 0.0080 - val_loss: 5.3108e-04 - val_mae: 0.0171 - 4s/epoch - 24ms/step\n",
      "Epoch 211/3000\n",
      "\n",
      "Epoch 211: val_mae did not improve from 0.01693\n",
      "151/151 - 4s - loss: 1.5817e-04 - mae: 0.0080 - val_loss: 5.3700e-04 - val_mae: 0.0173 - 4s/epoch - 23ms/step\n",
      "Epoch 212/3000\n",
      "\n",
      "Epoch 212: val_mae improved from 0.01693 to 0.01692, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 1.5741e-04 - mae: 0.0080 - val_loss: 5.1303e-04 - val_mae: 0.0169 - 3s/epoch - 23ms/step\n",
      "Epoch 213/3000\n",
      "\n",
      "Epoch 213: val_mae improved from 0.01692 to 0.01678, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.5709e-04 - mae: 0.0080 - val_loss: 5.0224e-04 - val_mae: 0.0168 - 4s/epoch - 24ms/step\n",
      "Epoch 214/3000\n",
      "\n",
      "Epoch 214: val_mae did not improve from 0.01678\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "151/151 - 4s - loss: 1.5706e-04 - mae: 0.0080 - val_loss: 5.3501e-04 - val_mae: 0.0173 - 4s/epoch - 24ms/step\n",
      "Epoch 215/3000\n",
      "\n",
      "Epoch 215: val_mae did not improve from 0.01678\n",
      "151/151 - 4s - loss: 1.5648e-04 - mae: 0.0079 - val_loss: 5.2659e-04 - val_mae: 0.0171 - 4s/epoch - 24ms/step\n",
      "Epoch 216/3000\n",
      "\n",
      "Epoch 216: val_mae improved from 0.01678 to 0.01666, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.5649e-04 - mae: 0.0079 - val_loss: 4.9844e-04 - val_mae: 0.0167 - 4s/epoch - 25ms/step\n",
      "Epoch 217/3000\n",
      "\n",
      "Epoch 217: val_mae did not improve from 0.01666\n",
      "151/151 - 4s - loss: 1.5738e-04 - mae: 0.0080 - val_loss: 5.2985e-04 - val_mae: 0.0171 - 4s/epoch - 24ms/step\n",
      "Epoch 218/3000\n",
      "\n",
      "Epoch 218: val_mae did not improve from 0.01666\n",
      "151/151 - 4s - loss: 1.5609e-04 - mae: 0.0079 - val_loss: 5.3651e-04 - val_mae: 0.0173 - 4s/epoch - 24ms/step\n",
      "Epoch 219/3000\n",
      "\n",
      "Epoch 219: val_mae did not improve from 0.01666\n",
      "151/151 - 4s - loss: 1.5552e-04 - mae: 0.0079 - val_loss: 5.1572e-04 - val_mae: 0.0169 - 4s/epoch - 25ms/step\n",
      "Epoch 220/3000\n",
      "\n",
      "Epoch 220: val_mae did not improve from 0.01666\n",
      "151/151 - 4s - loss: 1.5527e-04 - mae: 0.0079 - val_loss: 5.2755e-04 - val_mae: 0.0171 - 4s/epoch - 23ms/step\n",
      "Epoch 221/3000\n",
      "\n",
      "Epoch 221: val_mae did not improve from 0.01666\n",
      "151/151 - 4s - loss: 1.5581e-04 - mae: 0.0079 - val_loss: 5.3106e-04 - val_mae: 0.0171 - 4s/epoch - 24ms/step\n",
      "Epoch 222/3000\n",
      "\n",
      "Epoch 222: val_mae did not improve from 0.01666\n",
      "151/151 - 4s - loss: 1.5576e-04 - mae: 0.0079 - val_loss: 5.1952e-04 - val_mae: 0.0169 - 4s/epoch - 24ms/step\n",
      "Epoch 223/3000\n",
      "\n",
      "Epoch 223: val_mae did not improve from 0.01666\n",
      "151/151 - 4s - loss: 1.5538e-04 - mae: 0.0079 - val_loss: 5.4357e-04 - val_mae: 0.0172 - 4s/epoch - 24ms/step\n",
      "Epoch 224/3000\n",
      "\n",
      "Epoch 224: val_mae improved from 0.01666 to 0.01650, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.5495e-04 - mae: 0.0079 - val_loss: 4.9570e-04 - val_mae: 0.0165 - 4s/epoch - 25ms/step\n",
      "Epoch 225/3000\n",
      "\n",
      "Epoch 225: val_mae did not improve from 0.01650\n",
      "151/151 - 3s - loss: 1.5434e-04 - mae: 0.0079 - val_loss: 5.2467e-04 - val_mae: 0.0170 - 3s/epoch - 23ms/step\n",
      "Epoch 226/3000\n",
      "\n",
      "Epoch 226: val_mae did not improve from 0.01650\n",
      "151/151 - 4s - loss: 1.5494e-04 - mae: 0.0079 - val_loss: 5.1191e-04 - val_mae: 0.0167 - 4s/epoch - 24ms/step\n",
      "Epoch 227/3000\n",
      "\n",
      "Epoch 227: val_mae did not improve from 0.01650\n",
      "151/151 - 4s - loss: 1.5428e-04 - mae: 0.0079 - val_loss: 5.3088e-04 - val_mae: 0.0170 - 4s/epoch - 23ms/step\n",
      "Epoch 228/3000\n",
      "\n",
      "Epoch 228: val_mae did not improve from 0.01650\n",
      "151/151 - 4s - loss: 1.5540e-04 - mae: 0.0079 - val_loss: 5.6162e-04 - val_mae: 0.0175 - 4s/epoch - 24ms/step\n",
      "Epoch 229/3000\n",
      "\n",
      "Epoch 229: val_mae did not improve from 0.01650\n",
      "151/151 - 4s - loss: 1.5449e-04 - mae: 0.0079 - val_loss: 5.1394e-04 - val_mae: 0.0168 - 4s/epoch - 25ms/step\n",
      "Epoch 230/3000\n",
      "\n",
      "Epoch 230: val_mae did not improve from 0.01650\n",
      "151/151 - 4s - loss: 1.5359e-04 - mae: 0.0078 - val_loss: 5.2168e-04 - val_mae: 0.0169 - 4s/epoch - 24ms/step\n",
      "Epoch 231/3000\n",
      "\n",
      "Epoch 231: val_mae did not improve from 0.01650\n",
      "151/151 - 4s - loss: 1.5290e-04 - mae: 0.0078 - val_loss: 5.1515e-04 - val_mae: 0.0169 - 4s/epoch - 24ms/step\n",
      "Epoch 232/3000\n",
      "\n",
      "Epoch 232: val_mae improved from 0.01650 to 0.01629, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.5346e-04 - mae: 0.0078 - val_loss: 4.8154e-04 - val_mae: 0.0163 - 4s/epoch - 24ms/step\n",
      "Epoch 233/3000\n",
      "\n",
      "Epoch 233: val_mae did not improve from 0.01629\n",
      "151/151 - 4s - loss: 1.5326e-04 - mae: 0.0078 - val_loss: 4.9934e-04 - val_mae: 0.0166 - 4s/epoch - 24ms/step\n",
      "Epoch 234/3000\n",
      "\n",
      "Epoch 234: val_mae did not improve from 0.01629\n",
      "151/151 - 4s - loss: 1.5306e-04 - mae: 0.0078 - val_loss: 5.3284e-04 - val_mae: 0.0171 - 4s/epoch - 24ms/step\n",
      "Epoch 235/3000\n",
      "\n",
      "Epoch 235: val_mae did not improve from 0.01629\n",
      "151/151 - 4s - loss: 1.5305e-04 - mae: 0.0078 - val_loss: 5.0148e-04 - val_mae: 0.0166 - 4s/epoch - 24ms/step\n",
      "Epoch 236/3000\n",
      "\n",
      "Epoch 236: val_mae did not improve from 0.01629\n",
      "151/151 - 4s - loss: 1.5306e-04 - mae: 0.0078 - val_loss: 5.2756e-04 - val_mae: 0.0172 - 4s/epoch - 23ms/step\n",
      "Epoch 237/3000\n",
      "\n",
      "Epoch 237: val_mae did not improve from 0.01629\n",
      "151/151 - 4s - loss: 1.5341e-04 - mae: 0.0078 - val_loss: 5.1106e-04 - val_mae: 0.0168 - 4s/epoch - 24ms/step\n",
      "Epoch 238/3000\n",
      "\n",
      "Epoch 238: val_mae did not improve from 0.01629\n",
      "151/151 - 4s - loss: 1.5237e-04 - mae: 0.0078 - val_loss: 5.0300e-04 - val_mae: 0.0166 - 4s/epoch - 24ms/step\n",
      "Epoch 239/3000\n",
      "\n",
      "Epoch 239: val_mae did not improve from 0.01629\n",
      "151/151 - 4s - loss: 1.5222e-04 - mae: 0.0078 - val_loss: 5.2189e-04 - val_mae: 0.0170 - 4s/epoch - 24ms/step\n",
      "Epoch 240/3000\n",
      "\n",
      "Epoch 240: val_mae improved from 0.01629 to 0.01626, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.5263e-04 - mae: 0.0078 - val_loss: 4.8125e-04 - val_mae: 0.0163 - 4s/epoch - 24ms/step\n",
      "Epoch 241/3000\n",
      "\n",
      "Epoch 241: val_mae did not improve from 0.01626\n",
      "151/151 - 4s - loss: 1.5225e-04 - mae: 0.0078 - val_loss: 4.8688e-04 - val_mae: 0.0164 - 4s/epoch - 24ms/step\n",
      "Epoch 242/3000\n",
      "\n",
      "Epoch 242: val_mae improved from 0.01626 to 0.01598, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.5168e-04 - mae: 0.0078 - val_loss: 4.6118e-04 - val_mae: 0.0160 - 4s/epoch - 24ms/step\n",
      "Epoch 243/3000\n",
      "\n",
      "Epoch 243: val_mae did not improve from 0.01598\n",
      "151/151 - 4s - loss: 1.5275e-04 - mae: 0.0078 - val_loss: 5.1953e-04 - val_mae: 0.0170 - 4s/epoch - 24ms/step\n",
      "Epoch 244/3000\n",
      "\n",
      "Epoch 244: val_mae did not improve from 0.01598\n",
      "151/151 - 4s - loss: 1.5118e-04 - mae: 0.0078 - val_loss: 5.3707e-04 - val_mae: 0.0172 - 4s/epoch - 27ms/step\n",
      "Epoch 245/3000\n",
      "\n",
      "Epoch 245: val_mae did not improve from 0.01598\n",
      "151/151 - 4s - loss: 1.5185e-04 - mae: 0.0078 - val_loss: 4.7055e-04 - val_mae: 0.0160 - 4s/epoch - 24ms/step\n",
      "Epoch 246/3000\n",
      "\n",
      "Epoch 246: val_mae did not improve from 0.01598\n",
      "151/151 - 4s - loss: 1.5114e-04 - mae: 0.0078 - val_loss: 4.8333e-04 - val_mae: 0.0164 - 4s/epoch - 23ms/step\n",
      "Epoch 247/3000\n",
      "\n",
      "Epoch 247: val_mae did not improve from 0.01598\n",
      "151/151 - 4s - loss: 1.5182e-04 - mae: 0.0078 - val_loss: 4.8553e-04 - val_mae: 0.0163 - 4s/epoch - 23ms/step\n",
      "Epoch 248/3000\n",
      "\n",
      "Epoch 248: val_mae did not improve from 0.01598\n",
      "151/151 - 4s - loss: 1.5105e-04 - mae: 0.0078 - val_loss: 5.0202e-04 - val_mae: 0.0166 - 4s/epoch - 24ms/step\n",
      "Epoch 249/3000\n",
      "\n",
      "Epoch 249: val_mae did not improve from 0.01598\n",
      "151/151 - 4s - loss: 1.5092e-04 - mae: 0.0078 - val_loss: 5.0546e-04 - val_mae: 0.0168 - 4s/epoch - 24ms/step\n",
      "Epoch 250/3000\n",
      "\n",
      "Epoch 250: val_mae did not improve from 0.01598\n",
      "151/151 - 4s - loss: 1.5066e-04 - mae: 0.0077 - val_loss: 4.9243e-04 - val_mae: 0.0165 - 4s/epoch - 25ms/step\n",
      "Epoch 251/3000\n",
      "\n",
      "Epoch 251: val_mae did not improve from 0.01598\n",
      "151/151 - 4s - loss: 1.5011e-04 - mae: 0.0077 - val_loss: 5.0461e-04 - val_mae: 0.0165 - 4s/epoch - 24ms/step\n",
      "Epoch 252/3000\n",
      "\n",
      "Epoch 252: val_mae did not improve from 0.01598\n",
      "151/151 - 4s - loss: 1.5006e-04 - mae: 0.0077 - val_loss: 4.7848e-04 - val_mae: 0.0162 - 4s/epoch - 25ms/step\n",
      "Epoch 253/3000\n",
      "\n",
      "Epoch 253: val_mae improved from 0.01598 to 0.01586, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.5030e-04 - mae: 0.0077 - val_loss: 4.5700e-04 - val_mae: 0.0159 - 4s/epoch - 25ms/step\n",
      "Epoch 254/3000\n",
      "\n",
      "Epoch 254: val_mae did not improve from 0.01586\n",
      "151/151 - 4s - loss: 1.5053e-04 - mae: 0.0078 - val_loss: 5.2903e-04 - val_mae: 0.0171 - 4s/epoch - 24ms/step\n",
      "Epoch 255/3000\n",
      "\n",
      "Epoch 255: val_mae did not improve from 0.01586\n",
      "151/151 - 4s - loss: 1.5073e-04 - mae: 0.0078 - val_loss: 5.2335e-04 - val_mae: 0.0167 - 4s/epoch - 24ms/step\n",
      "Epoch 256/3000\n",
      "\n",
      "Epoch 256: val_mae did not improve from 0.01586\n",
      "151/151 - 4s - loss: 1.5022e-04 - mae: 0.0077 - val_loss: 5.0906e-04 - val_mae: 0.0167 - 4s/epoch - 24ms/step\n",
      "Epoch 257/3000\n",
      "\n",
      "Epoch 257: val_mae did not improve from 0.01586\n",
      "151/151 - 3s - loss: 1.5053e-04 - mae: 0.0078 - val_loss: 5.0001e-04 - val_mae: 0.0166 - 3s/epoch - 23ms/step\n",
      "Epoch 258/3000\n",
      "\n",
      "Epoch 258: val_mae did not improve from 0.01586\n",
      "151/151 - 4s - loss: 1.4948e-04 - mae: 0.0077 - val_loss: 5.4330e-04 - val_mae: 0.0171 - 4s/epoch - 24ms/step\n",
      "Epoch 259/3000\n",
      "\n",
      "Epoch 259: val_mae did not improve from 0.01586\n",
      "151/151 - 3s - loss: 1.4928e-04 - mae: 0.0077 - val_loss: 4.9371e-04 - val_mae: 0.0166 - 3s/epoch - 23ms/step\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 260/3000\n",
      "\n",
      "Epoch 260: val_mae did not improve from 0.01586\n",
      "151/151 - 4s - loss: 1.4968e-04 - mae: 0.0077 - val_loss: 5.2026e-04 - val_mae: 0.0168 - 4s/epoch - 24ms/step\n",
      "Epoch 261/3000\n",
      "\n",
      "Epoch 261: val_mae did not improve from 0.01586\n",
      "151/151 - 3s - loss: 1.4872e-04 - mae: 0.0077 - val_loss: 4.7232e-04 - val_mae: 0.0161 - 3s/epoch - 23ms/step\n",
      "Epoch 262/3000\n",
      "\n",
      "Epoch 262: val_mae did not improve from 0.01586\n",
      "151/151 - 4s - loss: 1.4847e-04 - mae: 0.0077 - val_loss: 4.7769e-04 - val_mae: 0.0162 - 4s/epoch - 24ms/step\n",
      "Epoch 263/3000\n",
      "\n",
      "Epoch 263: val_mae did not improve from 0.01586\n",
      "151/151 - 4s - loss: 1.4994e-04 - mae: 0.0077 - val_loss: 5.2334e-04 - val_mae: 0.0169 - 4s/epoch - 24ms/step\n",
      "Epoch 264/3000\n",
      "\n",
      "Epoch 264: val_mae improved from 0.01586 to 0.01578, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.4911e-04 - mae: 0.0077 - val_loss: 4.5604e-04 - val_mae: 0.0158 - 4s/epoch - 24ms/step\n",
      "Epoch 265/3000\n",
      "\n",
      "Epoch 265: val_mae did not improve from 0.01578\n",
      "151/151 - 4s - loss: 1.4861e-04 - mae: 0.0077 - val_loss: 4.9877e-04 - val_mae: 0.0166 - 4s/epoch - 25ms/step\n",
      "Epoch 266/3000\n",
      "\n",
      "Epoch 266: val_mae improved from 0.01578 to 0.01571, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.4836e-04 - mae: 0.0077 - val_loss: 4.5292e-04 - val_mae: 0.0157 - 4s/epoch - 24ms/step\n",
      "Epoch 267/3000\n",
      "\n",
      "Epoch 267: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4762e-04 - mae: 0.0076 - val_loss: 5.2646e-04 - val_mae: 0.0169 - 4s/epoch - 24ms/step\n",
      "Epoch 268/3000\n",
      "\n",
      "Epoch 268: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4782e-04 - mae: 0.0076 - val_loss: 4.9774e-04 - val_mae: 0.0163 - 4s/epoch - 24ms/step\n",
      "Epoch 269/3000\n",
      "\n",
      "Epoch 269: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4786e-04 - mae: 0.0077 - val_loss: 4.5389e-04 - val_mae: 0.0157 - 4s/epoch - 24ms/step\n",
      "Epoch 270/3000\n",
      "\n",
      "Epoch 270: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4737e-04 - mae: 0.0076 - val_loss: 5.0491e-04 - val_mae: 0.0168 - 4s/epoch - 24ms/step\n",
      "Epoch 271/3000\n",
      "\n",
      "Epoch 271: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4831e-04 - mae: 0.0077 - val_loss: 4.9707e-04 - val_mae: 0.0165 - 4s/epoch - 24ms/step\n",
      "Epoch 272/3000\n",
      "\n",
      "Epoch 272: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4847e-04 - mae: 0.0077 - val_loss: 4.9092e-04 - val_mae: 0.0164 - 4s/epoch - 24ms/step\n",
      "Epoch 273/3000\n",
      "\n",
      "Epoch 273: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4719e-04 - mae: 0.0076 - val_loss: 4.7397e-04 - val_mae: 0.0160 - 4s/epoch - 24ms/step\n",
      "Epoch 274/3000\n",
      "\n",
      "Epoch 274: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4757e-04 - mae: 0.0077 - val_loss: 4.6652e-04 - val_mae: 0.0160 - 4s/epoch - 24ms/step\n",
      "Epoch 275/3000\n",
      "\n",
      "Epoch 275: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4760e-04 - mae: 0.0077 - val_loss: 5.0783e-04 - val_mae: 0.0166 - 4s/epoch - 24ms/step\n",
      "Epoch 276/3000\n",
      "\n",
      "Epoch 276: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4685e-04 - mae: 0.0076 - val_loss: 4.8495e-04 - val_mae: 0.0163 - 4s/epoch - 24ms/step\n",
      "Epoch 277/3000\n",
      "\n",
      "Epoch 277: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4694e-04 - mae: 0.0076 - val_loss: 4.9526e-04 - val_mae: 0.0165 - 4s/epoch - 24ms/step\n",
      "Epoch 278/3000\n",
      "\n",
      "Epoch 278: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4694e-04 - mae: 0.0076 - val_loss: 4.5801e-04 - val_mae: 0.0158 - 4s/epoch - 23ms/step\n",
      "Epoch 279/3000\n",
      "\n",
      "Epoch 279: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4728e-04 - mae: 0.0076 - val_loss: 4.9927e-04 - val_mae: 0.0165 - 4s/epoch - 24ms/step\n",
      "Epoch 280/3000\n",
      "\n",
      "Epoch 280: val_mae did not improve from 0.01571\n",
      "151/151 - 3s - loss: 1.4686e-04 - mae: 0.0076 - val_loss: 5.1172e-04 - val_mae: 0.0167 - 3s/epoch - 23ms/step\n",
      "Epoch 281/3000\n",
      "\n",
      "Epoch 281: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4685e-04 - mae: 0.0076 - val_loss: 5.0965e-04 - val_mae: 0.0166 - 4s/epoch - 24ms/step\n",
      "Epoch 282/3000\n",
      "\n",
      "Epoch 282: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4632e-04 - mae: 0.0076 - val_loss: 5.3874e-04 - val_mae: 0.0171 - 4s/epoch - 24ms/step\n",
      "Epoch 283/3000\n",
      "\n",
      "Epoch 283: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4605e-04 - mae: 0.0076 - val_loss: 4.8012e-04 - val_mae: 0.0162 - 4s/epoch - 24ms/step\n",
      "Epoch 284/3000\n",
      "\n",
      "Epoch 284: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4617e-04 - mae: 0.0076 - val_loss: 4.5298e-04 - val_mae: 0.0158 - 4s/epoch - 24ms/step\n",
      "Epoch 285/3000\n",
      "\n",
      "Epoch 285: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4605e-04 - mae: 0.0076 - val_loss: 4.5538e-04 - val_mae: 0.0158 - 4s/epoch - 24ms/step\n",
      "Epoch 286/3000\n",
      "\n",
      "Epoch 286: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4588e-04 - mae: 0.0076 - val_loss: 4.9507e-04 - val_mae: 0.0164 - 4s/epoch - 24ms/step\n",
      "Epoch 287/3000\n",
      "\n",
      "Epoch 287: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4575e-04 - mae: 0.0076 - val_loss: 4.8916e-04 - val_mae: 0.0163 - 4s/epoch - 25ms/step\n",
      "Epoch 288/3000\n",
      "\n",
      "Epoch 288: val_mae did not improve from 0.01571\n",
      "151/151 - 4s - loss: 1.4603e-04 - mae: 0.0076 - val_loss: 4.9771e-04 - val_mae: 0.0164 - 4s/epoch - 24ms/step\n",
      "Epoch 289/3000\n",
      "\n",
      "Epoch 289: val_mae improved from 0.01571 to 0.01561, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.4534e-04 - mae: 0.0076 - val_loss: 4.3975e-04 - val_mae: 0.0156 - 4s/epoch - 24ms/step\n",
      "Epoch 290/3000\n",
      "\n",
      "Epoch 290: val_mae did not improve from 0.01561\n",
      "151/151 - 4s - loss: 1.4632e-04 - mae: 0.0076 - val_loss: 4.5975e-04 - val_mae: 0.0160 - 4s/epoch - 24ms/step\n",
      "Epoch 291/3000\n",
      "\n",
      "Epoch 291: val_mae did not improve from 0.01561\n",
      "151/151 - 3s - loss: 1.4563e-04 - mae: 0.0076 - val_loss: 4.7617e-04 - val_mae: 0.0161 - 3s/epoch - 23ms/step\n",
      "Epoch 292/3000\n",
      "\n",
      "Epoch 292: val_mae did not improve from 0.01561\n",
      "151/151 - 4s - loss: 1.4488e-04 - mae: 0.0076 - val_loss: 4.8855e-04 - val_mae: 0.0165 - 4s/epoch - 25ms/step\n",
      "Epoch 293/3000\n",
      "\n",
      "Epoch 293: val_mae did not improve from 0.01561\n",
      "151/151 - 4s - loss: 1.4560e-04 - mae: 0.0076 - val_loss: 4.6969e-04 - val_mae: 0.0159 - 4s/epoch - 24ms/step\n",
      "Epoch 294/3000\n",
      "\n",
      "Epoch 294: val_mae did not improve from 0.01561\n",
      "151/151 - 4s - loss: 1.4485e-04 - mae: 0.0076 - val_loss: 4.4530e-04 - val_mae: 0.0156 - 4s/epoch - 25ms/step\n",
      "Epoch 295/3000\n",
      "\n",
      "Epoch 295: val_mae did not improve from 0.01561\n",
      "151/151 - 4s - loss: 1.4474e-04 - mae: 0.0075 - val_loss: 4.7056e-04 - val_mae: 0.0160 - 4s/epoch - 25ms/step\n",
      "Epoch 296/3000\n",
      "\n",
      "Epoch 296: val_mae did not improve from 0.01561\n",
      "151/151 - 4s - loss: 1.4505e-04 - mae: 0.0076 - val_loss: 4.5134e-04 - val_mae: 0.0158 - 4s/epoch - 24ms/step\n",
      "Epoch 297/3000\n",
      "\n",
      "Epoch 297: val_mae improved from 0.01561 to 0.01560, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.4499e-04 - mae: 0.0076 - val_loss: 4.4480e-04 - val_mae: 0.0156 - 4s/epoch - 24ms/step\n",
      "Epoch 298/3000\n",
      "\n",
      "Epoch 298: val_mae did not improve from 0.01560\n",
      "151/151 - 4s - loss: 1.4422e-04 - mae: 0.0075 - val_loss: 4.9107e-04 - val_mae: 0.0163 - 4s/epoch - 23ms/step\n",
      "Epoch 299/3000\n",
      "\n",
      "Epoch 299: val_mae did not improve from 0.01560\n",
      "151/151 - 4s - loss: 1.4384e-04 - mae: 0.0075 - val_loss: 4.5527e-04 - val_mae: 0.0158 - 4s/epoch - 24ms/step\n",
      "Epoch 300/3000\n",
      "\n",
      "Epoch 300: val_mae improved from 0.01560 to 0.01541, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.4448e-04 - mae: 0.0075 - val_loss: 4.3607e-04 - val_mae: 0.0154 - 4s/epoch - 24ms/step\n",
      "Epoch 301/3000\n",
      "\n",
      "Epoch 301: val_mae did not improve from 0.01541\n",
      "151/151 - 4s - loss: 1.4470e-04 - mae: 0.0076 - val_loss: 4.7648e-04 - val_mae: 0.0161 - 4s/epoch - 24ms/step\n",
      "Epoch 302/3000\n",
      "\n",
      "Epoch 302: val_mae did not improve from 0.01541\n",
      "151/151 - 4s - loss: 1.4388e-04 - mae: 0.0075 - val_loss: 4.5883e-04 - val_mae: 0.0159 - 4s/epoch - 24ms/step\n",
      "Epoch 303/3000\n",
      "\n",
      "Epoch 303: val_mae did not improve from 0.01541\n",
      "151/151 - 4s - loss: 1.4442e-04 - mae: 0.0075 - val_loss: 4.6281e-04 - val_mae: 0.0160 - 4s/epoch - 24ms/step\n",
      "Epoch 304/3000\n",
      "\n",
      "Epoch 304: val_mae did not improve from 0.01541\n",
      "151/151 - 4s - loss: 1.4378e-04 - mae: 0.0075 - val_loss: 4.5808e-04 - val_mae: 0.0158 - 4s/epoch - 24ms/step\n",
      "Epoch 305/3000\n",
      "\n",
      "Epoch 305: val_mae did not improve from 0.01541\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "151/151 - 4s - loss: 1.4417e-04 - mae: 0.0075 - val_loss: 4.6927e-04 - val_mae: 0.0160 - 4s/epoch - 24ms/step\n",
      "Epoch 306/3000\n",
      "\n",
      "Epoch 306: val_mae did not improve from 0.01541\n",
      "151/151 - 3s - loss: 1.4396e-04 - mae: 0.0075 - val_loss: 4.6155e-04 - val_mae: 0.0159 - 3s/epoch - 23ms/step\n",
      "Epoch 307/3000\n",
      "\n",
      "Epoch 307: val_mae did not improve from 0.01541\n",
      "151/151 - 4s - loss: 1.4417e-04 - mae: 0.0075 - val_loss: 4.7199e-04 - val_mae: 0.0161 - 4s/epoch - 25ms/step\n",
      "Epoch 308/3000\n",
      "\n",
      "Epoch 308: val_mae did not improve from 0.01541\n",
      "151/151 - 4s - loss: 1.4321e-04 - mae: 0.0075 - val_loss: 4.9236e-04 - val_mae: 0.0163 - 4s/epoch - 24ms/step\n",
      "Epoch 309/3000\n",
      "\n",
      "Epoch 309: val_mae did not improve from 0.01541\n",
      "151/151 - 4s - loss: 1.4373e-04 - mae: 0.0075 - val_loss: 4.9515e-04 - val_mae: 0.0165 - 4s/epoch - 24ms/step\n",
      "Epoch 310/3000\n",
      "\n",
      "Epoch 310: val_mae did not improve from 0.01541\n",
      "151/151 - 4s - loss: 1.4345e-04 - mae: 0.0075 - val_loss: 4.5402e-04 - val_mae: 0.0157 - 4s/epoch - 25ms/step\n",
      "Epoch 311/3000\n",
      "\n",
      "Epoch 311: val_mae did not improve from 0.01541\n",
      "151/151 - 4s - loss: 1.4316e-04 - mae: 0.0075 - val_loss: 5.0340e-04 - val_mae: 0.0163 - 4s/epoch - 24ms/step\n",
      "Epoch 312/3000\n",
      "\n",
      "Epoch 312: val_mae did not improve from 0.01541\n",
      "151/151 - 4s - loss: 1.4285e-04 - mae: 0.0075 - val_loss: 4.8239e-04 - val_mae: 0.0163 - 4s/epoch - 23ms/step\n",
      "Epoch 313/3000\n",
      "\n",
      "Epoch 313: val_mae did not improve from 0.01541\n",
      "151/151 - 4s - loss: 1.4267e-04 - mae: 0.0075 - val_loss: 4.7986e-04 - val_mae: 0.0160 - 4s/epoch - 24ms/step\n",
      "Epoch 314/3000\n",
      "\n",
      "Epoch 314: val_mae did not improve from 0.01541\n",
      "151/151 - 4s - loss: 1.4275e-04 - mae: 0.0075 - val_loss: 4.7691e-04 - val_mae: 0.0161 - 4s/epoch - 24ms/step\n",
      "Epoch 315/3000\n",
      "\n",
      "Epoch 315: val_mae did not improve from 0.01541\n",
      "151/151 - 3s - loss: 1.4295e-04 - mae: 0.0075 - val_loss: 5.2348e-04 - val_mae: 0.0170 - 3s/epoch - 23ms/step\n",
      "Epoch 316/3000\n",
      "\n",
      "Epoch 316: val_mae did not improve from 0.01541\n",
      "151/151 - 4s - loss: 1.4278e-04 - mae: 0.0075 - val_loss: 4.8352e-04 - val_mae: 0.0163 - 4s/epoch - 24ms/step\n",
      "Epoch 317/3000\n",
      "\n",
      "Epoch 317: val_mae improved from 0.01541 to 0.01525, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.4234e-04 - mae: 0.0075 - val_loss: 4.1716e-04 - val_mae: 0.0153 - 4s/epoch - 23ms/step\n",
      "Epoch 318/3000\n",
      "\n",
      "Epoch 318: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4297e-04 - mae: 0.0075 - val_loss: 4.5977e-04 - val_mae: 0.0159 - 3s/epoch - 23ms/step\n",
      "Epoch 319/3000\n",
      "\n",
      "Epoch 319: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4251e-04 - mae: 0.0075 - val_loss: 4.5087e-04 - val_mae: 0.0156 - 3s/epoch - 23ms/step\n",
      "Epoch 320/3000\n",
      "\n",
      "Epoch 320: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4255e-04 - mae: 0.0075 - val_loss: 4.3225e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 321/3000\n",
      "\n",
      "Epoch 321: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.4202e-04 - mae: 0.0075 - val_loss: 4.7339e-04 - val_mae: 0.0161 - 4s/epoch - 24ms/step\n",
      "Epoch 322/3000\n",
      "\n",
      "Epoch 322: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.4185e-04 - mae: 0.0074 - val_loss: 4.7364e-04 - val_mae: 0.0160 - 4s/epoch - 24ms/step\n",
      "Epoch 323/3000\n",
      "\n",
      "Epoch 323: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.4201e-04 - mae: 0.0075 - val_loss: 4.9024e-04 - val_mae: 0.0163 - 4s/epoch - 23ms/step\n",
      "Epoch 324/3000\n",
      "\n",
      "Epoch 324: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.4180e-04 - mae: 0.0075 - val_loss: 4.8984e-04 - val_mae: 0.0163 - 4s/epoch - 24ms/step\n",
      "Epoch 325/3000\n",
      "\n",
      "Epoch 325: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.4241e-04 - mae: 0.0075 - val_loss: 4.8418e-04 - val_mae: 0.0163 - 4s/epoch - 24ms/step\n",
      "Epoch 326/3000\n",
      "\n",
      "Epoch 326: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.4232e-04 - mae: 0.0075 - val_loss: 4.6765e-04 - val_mae: 0.0160 - 4s/epoch - 24ms/step\n",
      "Epoch 327/3000\n",
      "\n",
      "Epoch 327: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.4185e-04 - mae: 0.0075 - val_loss: 4.6044e-04 - val_mae: 0.0160 - 4s/epoch - 24ms/step\n",
      "Epoch 328/3000\n",
      "\n",
      "Epoch 328: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.4201e-04 - mae: 0.0075 - val_loss: 4.8890e-04 - val_mae: 0.0163 - 4s/epoch - 23ms/step\n",
      "Epoch 329/3000\n",
      "\n",
      "Epoch 329: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4143e-04 - mae: 0.0074 - val_loss: 4.4920e-04 - val_mae: 0.0156 - 3s/epoch - 23ms/step\n",
      "Epoch 330/3000\n",
      "\n",
      "Epoch 330: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.4135e-04 - mae: 0.0074 - val_loss: 4.8325e-04 - val_mae: 0.0162 - 4s/epoch - 23ms/step\n",
      "Epoch 331/3000\n",
      "\n",
      "Epoch 331: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4166e-04 - mae: 0.0075 - val_loss: 4.7854e-04 - val_mae: 0.0162 - 3s/epoch - 23ms/step\n",
      "Epoch 332/3000\n",
      "\n",
      "Epoch 332: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4138e-04 - mae: 0.0074 - val_loss: 4.9151e-04 - val_mae: 0.0164 - 3s/epoch - 23ms/step\n",
      "Epoch 333/3000\n",
      "\n",
      "Epoch 333: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4080e-04 - mae: 0.0074 - val_loss: 4.5676e-04 - val_mae: 0.0158 - 3s/epoch - 23ms/step\n",
      "Epoch 334/3000\n",
      "\n",
      "Epoch 334: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4180e-04 - mae: 0.0075 - val_loss: 4.5971e-04 - val_mae: 0.0159 - 3s/epoch - 22ms/step\n",
      "Epoch 335/3000\n",
      "\n",
      "Epoch 335: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.4080e-04 - mae: 0.0074 - val_loss: 4.5748e-04 - val_mae: 0.0159 - 4s/epoch - 23ms/step\n",
      "Epoch 336/3000\n",
      "\n",
      "Epoch 336: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.4129e-04 - mae: 0.0074 - val_loss: 4.8342e-04 - val_mae: 0.0162 - 4s/epoch - 24ms/step\n",
      "Epoch 337/3000\n",
      "\n",
      "Epoch 337: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.4101e-04 - mae: 0.0074 - val_loss: 4.6009e-04 - val_mae: 0.0159 - 4s/epoch - 23ms/step\n",
      "Epoch 338/3000\n",
      "\n",
      "Epoch 338: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4120e-04 - mae: 0.0074 - val_loss: 4.8758e-04 - val_mae: 0.0163 - 3s/epoch - 23ms/step\n",
      "Epoch 339/3000\n",
      "\n",
      "Epoch 339: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4058e-04 - mae: 0.0074 - val_loss: 4.7676e-04 - val_mae: 0.0161 - 3s/epoch - 22ms/step\n",
      "Epoch 340/3000\n",
      "\n",
      "Epoch 340: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4093e-04 - mae: 0.0074 - val_loss: 4.5384e-04 - val_mae: 0.0157 - 3s/epoch - 22ms/step\n",
      "Epoch 341/3000\n",
      "\n",
      "Epoch 341: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4062e-04 - mae: 0.0074 - val_loss: 4.6368e-04 - val_mae: 0.0159 - 3s/epoch - 23ms/step\n",
      "Epoch 342/3000\n",
      "\n",
      "Epoch 342: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4027e-04 - mae: 0.0074 - val_loss: 4.3674e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 343/3000\n",
      "\n",
      "Epoch 343: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4034e-04 - mae: 0.0074 - val_loss: 4.6927e-04 - val_mae: 0.0162 - 3s/epoch - 22ms/step\n",
      "Epoch 344/3000\n",
      "\n",
      "Epoch 344: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4051e-04 - mae: 0.0074 - val_loss: 4.8829e-04 - val_mae: 0.0163 - 3s/epoch - 23ms/step\n",
      "Epoch 345/3000\n",
      "\n",
      "Epoch 345: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4029e-04 - mae: 0.0074 - val_loss: 5.0128e-04 - val_mae: 0.0165 - 3s/epoch - 23ms/step\n",
      "Epoch 346/3000\n",
      "\n",
      "Epoch 346: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.3995e-04 - mae: 0.0074 - val_loss: 4.5827e-04 - val_mae: 0.0159 - 3s/epoch - 22ms/step\n",
      "Epoch 347/3000\n",
      "\n",
      "Epoch 347: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.4010e-04 - mae: 0.0074 - val_loss: 4.4632e-04 - val_mae: 0.0157 - 4s/epoch - 23ms/step\n",
      "Epoch 348/3000\n",
      "\n",
      "Epoch 348: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.3982e-04 - mae: 0.0074 - val_loss: 4.5988e-04 - val_mae: 0.0158 - 3s/epoch - 23ms/step\n",
      "Epoch 349/3000\n",
      "\n",
      "Epoch 349: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.3972e-04 - mae: 0.0074 - val_loss: 4.3418e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 350/3000\n",
      "\n",
      "Epoch 350: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.3989e-04 - mae: 0.0074 - val_loss: 4.9284e-04 - val_mae: 0.0164 - 4s/epoch - 23ms/step\n",
      "Epoch 351/3000\n",
      "\n",
      "Epoch 351: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.4010e-04 - mae: 0.0074 - val_loss: 4.5710e-04 - val_mae: 0.0158 - 3s/epoch - 22ms/step\n",
      "Epoch 352/3000\n",
      "\n",
      "Epoch 352: val_mae did not improve from 0.01525\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "151/151 - 3s - loss: 1.3963e-04 - mae: 0.0074 - val_loss: 5.0444e-04 - val_mae: 0.0166 - 3s/epoch - 23ms/step\n",
      "Epoch 353/3000\n",
      "\n",
      "Epoch 353: val_mae did not improve from 0.01525\n",
      "151/151 - 3s - loss: 1.3949e-04 - mae: 0.0074 - val_loss: 4.5503e-04 - val_mae: 0.0157 - 3s/epoch - 23ms/step\n",
      "Epoch 354/3000\n",
      "\n",
      "Epoch 354: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.3946e-04 - mae: 0.0074 - val_loss: 4.5134e-04 - val_mae: 0.0156 - 4s/epoch - 24ms/step\n",
      "Epoch 355/3000\n",
      "\n",
      "Epoch 355: val_mae did not improve from 0.01525\n",
      "151/151 - 4s - loss: 1.3931e-04 - mae: 0.0074 - val_loss: 5.0440e-04 - val_mae: 0.0165 - 4s/epoch - 23ms/step\n",
      "Epoch 356/3000\n",
      "\n",
      "Epoch 356: val_mae improved from 0.01525 to 0.01501, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.3929e-04 - mae: 0.0074 - val_loss: 4.1436e-04 - val_mae: 0.0150 - 4s/epoch - 23ms/step\n",
      "Epoch 357/3000\n",
      "\n",
      "Epoch 357: val_mae did not improve from 0.01501\n",
      "151/151 - 3s - loss: 1.3900e-04 - mae: 0.0074 - val_loss: 4.4669e-04 - val_mae: 0.0157 - 3s/epoch - 23ms/step\n",
      "Epoch 358/3000\n",
      "\n",
      "Epoch 358: val_mae did not improve from 0.01501\n",
      "151/151 - 4s - loss: 1.3905e-04 - mae: 0.0074 - val_loss: 4.8585e-04 - val_mae: 0.0162 - 4s/epoch - 23ms/step\n",
      "Epoch 359/3000\n",
      "\n",
      "Epoch 359: val_mae did not improve from 0.01501\n",
      "151/151 - 3s - loss: 1.3888e-04 - mae: 0.0074 - val_loss: 4.7131e-04 - val_mae: 0.0161 - 3s/epoch - 23ms/step\n",
      "Epoch 360/3000\n",
      "\n",
      "Epoch 360: val_mae did not improve from 0.01501\n",
      "151/151 - 3s - loss: 1.3899e-04 - mae: 0.0074 - val_loss: 4.5974e-04 - val_mae: 0.0158 - 3s/epoch - 22ms/step\n",
      "Epoch 361/3000\n",
      "\n",
      "Epoch 361: val_mae did not improve from 0.01501\n",
      "151/151 - 3s - loss: 1.3878e-04 - mae: 0.0074 - val_loss: 4.5676e-04 - val_mae: 0.0157 - 3s/epoch - 23ms/step\n",
      "Epoch 362/3000\n",
      "\n",
      "Epoch 362: val_mae did not improve from 0.01501\n",
      "151/151 - 3s - loss: 1.3896e-04 - mae: 0.0074 - val_loss: 4.4789e-04 - val_mae: 0.0156 - 3s/epoch - 23ms/step\n",
      "Epoch 363/3000\n",
      "\n",
      "Epoch 363: val_mae did not improve from 0.01501\n",
      "151/151 - 3s - loss: 1.3859e-04 - mae: 0.0073 - val_loss: 4.2498e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 364/3000\n",
      "\n",
      "Epoch 364: val_mae did not improve from 0.01501\n",
      "151/151 - 3s - loss: 1.3917e-04 - mae: 0.0074 - val_loss: 4.4670e-04 - val_mae: 0.0156 - 3s/epoch - 23ms/step\n",
      "Epoch 365/3000\n",
      "\n",
      "Epoch 365: val_mae did not improve from 0.01501\n",
      "151/151 - 3s - loss: 1.3849e-04 - mae: 0.0073 - val_loss: 4.1822e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 366/3000\n",
      "\n",
      "Epoch 366: val_mae did not improve from 0.01501\n",
      "151/151 - 4s - loss: 1.3846e-04 - mae: 0.0073 - val_loss: 4.4085e-04 - val_mae: 0.0154 - 4s/epoch - 23ms/step\n",
      "Epoch 367/3000\n",
      "\n",
      "Epoch 367: val_mae did not improve from 0.01501\n",
      "151/151 - 3s - loss: 1.3807e-04 - mae: 0.0073 - val_loss: 4.6253e-04 - val_mae: 0.0159 - 3s/epoch - 23ms/step\n",
      "Epoch 368/3000\n",
      "\n",
      "Epoch 368: val_mae did not improve from 0.01501\n",
      "151/151 - 3s - loss: 1.3884e-04 - mae: 0.0074 - val_loss: 4.4365e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 369/3000\n",
      "\n",
      "Epoch 369: val_mae did not improve from 0.01501\n",
      "151/151 - 3s - loss: 1.3830e-04 - mae: 0.0073 - val_loss: 4.4079e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 370/3000\n",
      "\n",
      "Epoch 370: val_mae did not improve from 0.01501\n",
      "151/151 - 3s - loss: 1.3788e-04 - mae: 0.0073 - val_loss: 4.3642e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 371/3000\n",
      "\n",
      "Epoch 371: val_mae improved from 0.01501 to 0.01486, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 1.3792e-04 - mae: 0.0073 - val_loss: 4.0522e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 372/3000\n",
      "\n",
      "Epoch 372: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3820e-04 - mae: 0.0073 - val_loss: 4.5622e-04 - val_mae: 0.0157 - 3s/epoch - 23ms/step\n",
      "Epoch 373/3000\n",
      "\n",
      "Epoch 373: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3768e-04 - mae: 0.0073 - val_loss: 4.4916e-04 - val_mae: 0.0156 - 3s/epoch - 23ms/step\n",
      "Epoch 374/3000\n",
      "\n",
      "Epoch 374: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3794e-04 - mae: 0.0073 - val_loss: 4.3632e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 375/3000\n",
      "\n",
      "Epoch 375: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3817e-04 - mae: 0.0073 - val_loss: 4.6433e-04 - val_mae: 0.0159 - 3s/epoch - 23ms/step\n",
      "Epoch 376/3000\n",
      "\n",
      "Epoch 376: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3745e-04 - mae: 0.0073 - val_loss: 4.3960e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 377/3000\n",
      "\n",
      "Epoch 377: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3701e-04 - mae: 0.0073 - val_loss: 4.4285e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 378/3000\n",
      "\n",
      "Epoch 378: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3764e-04 - mae: 0.0073 - val_loss: 4.4192e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 379/3000\n",
      "\n",
      "Epoch 379: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3797e-04 - mae: 0.0073 - val_loss: 4.3393e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 380/3000\n",
      "\n",
      "Epoch 380: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3789e-04 - mae: 0.0073 - val_loss: 4.8562e-04 - val_mae: 0.0162 - 3s/epoch - 23ms/step\n",
      "Epoch 381/3000\n",
      "\n",
      "Epoch 381: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3725e-04 - mae: 0.0073 - val_loss: 4.5038e-04 - val_mae: 0.0157 - 4s/epoch - 23ms/step\n",
      "Epoch 382/3000\n",
      "\n",
      "Epoch 382: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3716e-04 - mae: 0.0073 - val_loss: 4.5481e-04 - val_mae: 0.0157 - 3s/epoch - 23ms/step\n",
      "Epoch 383/3000\n",
      "\n",
      "Epoch 383: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3711e-04 - mae: 0.0073 - val_loss: 4.2527e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 384/3000\n",
      "\n",
      "Epoch 384: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3720e-04 - mae: 0.0073 - val_loss: 4.4461e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 385/3000\n",
      "\n",
      "Epoch 385: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3739e-04 - mae: 0.0073 - val_loss: 4.4867e-04 - val_mae: 0.0155 - 4s/epoch - 24ms/step\n",
      "Epoch 386/3000\n",
      "\n",
      "Epoch 386: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3690e-04 - mae: 0.0073 - val_loss: 4.4908e-04 - val_mae: 0.0156 - 4s/epoch - 23ms/step\n",
      "Epoch 387/3000\n",
      "\n",
      "Epoch 387: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3660e-04 - mae: 0.0073 - val_loss: 4.3304e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 388/3000\n",
      "\n",
      "Epoch 388: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3687e-04 - mae: 0.0073 - val_loss: 4.3120e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 389/3000\n",
      "\n",
      "Epoch 389: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3712e-04 - mae: 0.0073 - val_loss: 4.5830e-04 - val_mae: 0.0158 - 4s/epoch - 24ms/step\n",
      "Epoch 390/3000\n",
      "\n",
      "Epoch 390: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3671e-04 - mae: 0.0073 - val_loss: 4.4719e-04 - val_mae: 0.0155 - 4s/epoch - 23ms/step\n",
      "Epoch 391/3000\n",
      "\n",
      "Epoch 391: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3662e-04 - mae: 0.0073 - val_loss: 4.7054e-04 - val_mae: 0.0160 - 3s/epoch - 23ms/step\n",
      "Epoch 392/3000\n",
      "\n",
      "Epoch 392: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3637e-04 - mae: 0.0073 - val_loss: 4.5737e-04 - val_mae: 0.0158 - 3s/epoch - 22ms/step\n",
      "Epoch 393/3000\n",
      "\n",
      "Epoch 393: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3597e-04 - mae: 0.0072 - val_loss: 4.3801e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 394/3000\n",
      "\n",
      "Epoch 394: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3635e-04 - mae: 0.0073 - val_loss: 4.3351e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 395/3000\n",
      "\n",
      "Epoch 395: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3602e-04 - mae: 0.0073 - val_loss: 4.4446e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 396/3000\n",
      "\n",
      "Epoch 396: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3606e-04 - mae: 0.0073 - val_loss: 4.9346e-04 - val_mae: 0.0164 - 3s/epoch - 22ms/step\n",
      "Epoch 397/3000\n",
      "\n",
      "Epoch 397: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3642e-04 - mae: 0.0073 - val_loss: 4.2541e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 398/3000\n",
      "\n",
      "Epoch 398: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3589e-04 - mae: 0.0072 - val_loss: 4.5721e-04 - val_mae: 0.0156 - 3s/epoch - 23ms/step\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 399/3000\n",
      "\n",
      "Epoch 399: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3575e-04 - mae: 0.0072 - val_loss: 4.5686e-04 - val_mae: 0.0157 - 3s/epoch - 23ms/step\n",
      "Epoch 400/3000\n",
      "\n",
      "Epoch 400: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3571e-04 - mae: 0.0072 - val_loss: 4.4529e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 401/3000\n",
      "\n",
      "Epoch 401: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3616e-04 - mae: 0.0073 - val_loss: 4.6919e-04 - val_mae: 0.0160 - 4s/epoch - 23ms/step\n",
      "Epoch 402/3000\n",
      "\n",
      "Epoch 402: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3632e-04 - mae: 0.0073 - val_loss: 4.9393e-04 - val_mae: 0.0164 - 4s/epoch - 24ms/step\n",
      "Epoch 403/3000\n",
      "\n",
      "Epoch 403: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3578e-04 - mae: 0.0073 - val_loss: 4.2991e-04 - val_mae: 0.0153 - 4s/epoch - 24ms/step\n",
      "Epoch 404/3000\n",
      "\n",
      "Epoch 404: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3555e-04 - mae: 0.0072 - val_loss: 4.7672e-04 - val_mae: 0.0161 - 4s/epoch - 24ms/step\n",
      "Epoch 405/3000\n",
      "\n",
      "Epoch 405: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3550e-04 - mae: 0.0072 - val_loss: 4.7748e-04 - val_mae: 0.0160 - 3s/epoch - 23ms/step\n",
      "Epoch 406/3000\n",
      "\n",
      "Epoch 406: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3563e-04 - mae: 0.0072 - val_loss: 4.7912e-04 - val_mae: 0.0161 - 3s/epoch - 23ms/step\n",
      "Epoch 407/3000\n",
      "\n",
      "Epoch 407: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3608e-04 - mae: 0.0073 - val_loss: 4.4524e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 408/3000\n",
      "\n",
      "Epoch 408: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3533e-04 - mae: 0.0072 - val_loss: 4.4184e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 409/3000\n",
      "\n",
      "Epoch 409: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3531e-04 - mae: 0.0072 - val_loss: 4.2992e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 410/3000\n",
      "\n",
      "Epoch 410: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3566e-04 - mae: 0.0072 - val_loss: 4.2557e-04 - val_mae: 0.0152 - 4s/epoch - 23ms/step\n",
      "Epoch 411/3000\n",
      "\n",
      "Epoch 411: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3493e-04 - mae: 0.0072 - val_loss: 4.3603e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 412/3000\n",
      "\n",
      "Epoch 412: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3514e-04 - mae: 0.0072 - val_loss: 4.2161e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 413/3000\n",
      "\n",
      "Epoch 413: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3529e-04 - mae: 0.0072 - val_loss: 4.8684e-04 - val_mae: 0.0161 - 4s/epoch - 23ms/step\n",
      "Epoch 414/3000\n",
      "\n",
      "Epoch 414: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3499e-04 - mae: 0.0072 - val_loss: 4.4099e-04 - val_mae: 0.0154 - 4s/epoch - 24ms/step\n",
      "Epoch 415/3000\n",
      "\n",
      "Epoch 415: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3503e-04 - mae: 0.0072 - val_loss: 4.6162e-04 - val_mae: 0.0158 - 3s/epoch - 22ms/step\n",
      "Epoch 416/3000\n",
      "\n",
      "Epoch 416: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3490e-04 - mae: 0.0072 - val_loss: 4.4141e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 417/3000\n",
      "\n",
      "Epoch 417: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3487e-04 - mae: 0.0072 - val_loss: 5.0188e-04 - val_mae: 0.0166 - 3s/epoch - 22ms/step\n",
      "Epoch 418/3000\n",
      "\n",
      "Epoch 418: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3506e-04 - mae: 0.0072 - val_loss: 4.7760e-04 - val_mae: 0.0160 - 3s/epoch - 22ms/step\n",
      "Epoch 419/3000\n",
      "\n",
      "Epoch 419: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3419e-04 - mae: 0.0072 - val_loss: 4.5256e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 420/3000\n",
      "\n",
      "Epoch 420: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3430e-04 - mae: 0.0072 - val_loss: 4.5246e-04 - val_mae: 0.0156 - 3s/epoch - 23ms/step\n",
      "Epoch 421/3000\n",
      "\n",
      "Epoch 421: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3459e-04 - mae: 0.0072 - val_loss: 4.6790e-04 - val_mae: 0.0159 - 3s/epoch - 23ms/step\n",
      "Epoch 422/3000\n",
      "\n",
      "Epoch 422: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3461e-04 - mae: 0.0072 - val_loss: 4.4888e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 423/3000\n",
      "\n",
      "Epoch 423: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3458e-04 - mae: 0.0072 - val_loss: 4.5049e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 424/3000\n",
      "\n",
      "Epoch 424: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3470e-04 - mae: 0.0072 - val_loss: 4.6636e-04 - val_mae: 0.0158 - 3s/epoch - 23ms/step\n",
      "Epoch 425/3000\n",
      "\n",
      "Epoch 425: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3451e-04 - mae: 0.0072 - val_loss: 4.7834e-04 - val_mae: 0.0162 - 3s/epoch - 23ms/step\n",
      "Epoch 426/3000\n",
      "\n",
      "Epoch 426: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3443e-04 - mae: 0.0072 - val_loss: 4.5373e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 427/3000\n",
      "\n",
      "Epoch 427: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3410e-04 - mae: 0.0072 - val_loss: 4.3720e-04 - val_mae: 0.0154 - 4s/epoch - 24ms/step\n",
      "Epoch 428/3000\n",
      "\n",
      "Epoch 428: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3442e-04 - mae: 0.0072 - val_loss: 4.3302e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 429/3000\n",
      "\n",
      "Epoch 429: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3380e-04 - mae: 0.0072 - val_loss: 4.4402e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 430/3000\n",
      "\n",
      "Epoch 430: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3411e-04 - mae: 0.0072 - val_loss: 4.3372e-04 - val_mae: 0.0154 - 3s/epoch - 21ms/step\n",
      "Epoch 431/3000\n",
      "\n",
      "Epoch 431: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3399e-04 - mae: 0.0072 - val_loss: 4.3989e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 432/3000\n",
      "\n",
      "Epoch 432: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3409e-04 - mae: 0.0072 - val_loss: 4.6953e-04 - val_mae: 0.0160 - 3s/epoch - 22ms/step\n",
      "Epoch 433/3000\n",
      "\n",
      "Epoch 433: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3416e-04 - mae: 0.0072 - val_loss: 4.4625e-04 - val_mae: 0.0155 - 3s/epoch - 21ms/step\n",
      "Epoch 434/3000\n",
      "\n",
      "Epoch 434: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3353e-04 - mae: 0.0072 - val_loss: 4.4422e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 435/3000\n",
      "\n",
      "Epoch 435: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3438e-04 - mae: 0.0072 - val_loss: 4.1657e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 436/3000\n",
      "\n",
      "Epoch 436: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3353e-04 - mae: 0.0072 - val_loss: 4.4064e-04 - val_mae: 0.0153 - 4s/epoch - 24ms/step\n",
      "Epoch 437/3000\n",
      "\n",
      "Epoch 437: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3390e-04 - mae: 0.0072 - val_loss: 4.3436e-04 - val_mae: 0.0153 - 4s/epoch - 23ms/step\n",
      "Epoch 438/3000\n",
      "\n",
      "Epoch 438: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3373e-04 - mae: 0.0072 - val_loss: 4.0790e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 439/3000\n",
      "\n",
      "Epoch 439: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3389e-04 - mae: 0.0072 - val_loss: 4.2958e-04 - val_mae: 0.0152 - 3s/epoch - 21ms/step\n",
      "Epoch 440/3000\n",
      "\n",
      "Epoch 440: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3424e-04 - mae: 0.0072 - val_loss: 4.6704e-04 - val_mae: 0.0158 - 3s/epoch - 22ms/step\n",
      "Epoch 441/3000\n",
      "\n",
      "Epoch 441: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3312e-04 - mae: 0.0072 - val_loss: 4.8720e-04 - val_mae: 0.0162 - 3s/epoch - 23ms/step\n",
      "Epoch 442/3000\n",
      "\n",
      "Epoch 442: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3357e-04 - mae: 0.0072 - val_loss: 4.2740e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 443/3000\n",
      "\n",
      "Epoch 443: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3330e-04 - mae: 0.0072 - val_loss: 4.6221e-04 - val_mae: 0.0158 - 4s/epoch - 24ms/step\n",
      "Epoch 444/3000\n",
      "\n",
      "Epoch 444: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3307e-04 - mae: 0.0072 - val_loss: 4.3250e-04 - val_mae: 0.0152 - 4s/epoch - 23ms/step\n",
      "Epoch 445/3000\n",
      "\n",
      "Epoch 445: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3327e-04 - mae: 0.0072 - val_loss: 4.2267e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 446/3000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 446: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3329e-04 - mae: 0.0072 - val_loss: 4.5024e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 447/3000\n",
      "\n",
      "Epoch 447: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3334e-04 - mae: 0.0072 - val_loss: 4.2599e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 448/3000\n",
      "\n",
      "Epoch 448: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3302e-04 - mae: 0.0072 - val_loss: 4.5211e-04 - val_mae: 0.0156 - 4s/epoch - 24ms/step\n",
      "Epoch 449/3000\n",
      "\n",
      "Epoch 449: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3318e-04 - mae: 0.0072 - val_loss: 4.5302e-04 - val_mae: 0.0156 - 3s/epoch - 23ms/step\n",
      "Epoch 450/3000\n",
      "\n",
      "Epoch 450: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3309e-04 - mae: 0.0072 - val_loss: 4.3907e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 451/3000\n",
      "\n",
      "Epoch 451: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3285e-04 - mae: 0.0071 - val_loss: 4.2665e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 452/3000\n",
      "\n",
      "Epoch 452: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3298e-04 - mae: 0.0072 - val_loss: 4.6169e-04 - val_mae: 0.0158 - 3s/epoch - 23ms/step\n",
      "Epoch 453/3000\n",
      "\n",
      "Epoch 453: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3258e-04 - mae: 0.0071 - val_loss: 4.5177e-04 - val_mae: 0.0155 - 4s/epoch - 24ms/step\n",
      "Epoch 454/3000\n",
      "\n",
      "Epoch 454: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3331e-04 - mae: 0.0072 - val_loss: 4.2460e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 455/3000\n",
      "\n",
      "Epoch 455: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3263e-04 - mae: 0.0071 - val_loss: 4.4467e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 456/3000\n",
      "\n",
      "Epoch 456: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3304e-04 - mae: 0.0072 - val_loss: 4.3788e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 457/3000\n",
      "\n",
      "Epoch 457: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3254e-04 - mae: 0.0071 - val_loss: 4.6612e-04 - val_mae: 0.0159 - 3s/epoch - 22ms/step\n",
      "Epoch 458/3000\n",
      "\n",
      "Epoch 458: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3260e-04 - mae: 0.0071 - val_loss: 4.3729e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 459/3000\n",
      "\n",
      "Epoch 459: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3233e-04 - mae: 0.0071 - val_loss: 4.6575e-04 - val_mae: 0.0159 - 4s/epoch - 23ms/step\n",
      "Epoch 460/3000\n",
      "\n",
      "Epoch 460: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3215e-04 - mae: 0.0071 - val_loss: 4.6243e-04 - val_mae: 0.0160 - 3s/epoch - 23ms/step\n",
      "Epoch 461/3000\n",
      "\n",
      "Epoch 461: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3221e-04 - mae: 0.0071 - val_loss: 4.2043e-04 - val_mae: 0.0151 - 4s/epoch - 23ms/step\n",
      "Epoch 462/3000\n",
      "\n",
      "Epoch 462: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3225e-04 - mae: 0.0071 - val_loss: 4.1368e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 463/3000\n",
      "\n",
      "Epoch 463: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3218e-04 - mae: 0.0071 - val_loss: 4.4506e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 464/3000\n",
      "\n",
      "Epoch 464: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3225e-04 - mae: 0.0071 - val_loss: 4.4573e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 465/3000\n",
      "\n",
      "Epoch 465: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3248e-04 - mae: 0.0071 - val_loss: 4.0891e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 466/3000\n",
      "\n",
      "Epoch 466: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3234e-04 - mae: 0.0071 - val_loss: 4.6476e-04 - val_mae: 0.0160 - 3s/epoch - 23ms/step\n",
      "Epoch 467/3000\n",
      "\n",
      "Epoch 467: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3267e-04 - mae: 0.0071 - val_loss: 4.4921e-04 - val_mae: 0.0156 - 3s/epoch - 23ms/step\n",
      "Epoch 468/3000\n",
      "\n",
      "Epoch 468: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3189e-04 - mae: 0.0071 - val_loss: 4.2827e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 469/3000\n",
      "\n",
      "Epoch 469: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3222e-04 - mae: 0.0071 - val_loss: 4.6948e-04 - val_mae: 0.0159 - 3s/epoch - 23ms/step\n",
      "Epoch 470/3000\n",
      "\n",
      "Epoch 470: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3170e-04 - mae: 0.0071 - val_loss: 4.3463e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 471/3000\n",
      "\n",
      "Epoch 471: val_mae did not improve from 0.01486\n",
      "151/151 - 4s - loss: 1.3162e-04 - mae: 0.0071 - val_loss: 4.6468e-04 - val_mae: 0.0158 - 4s/epoch - 23ms/step\n",
      "Epoch 472/3000\n",
      "\n",
      "Epoch 472: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3218e-04 - mae: 0.0071 - val_loss: 4.3263e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 473/3000\n",
      "\n",
      "Epoch 473: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3143e-04 - mae: 0.0071 - val_loss: 4.3642e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 474/3000\n",
      "\n",
      "Epoch 474: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3219e-04 - mae: 0.0071 - val_loss: 4.5019e-04 - val_mae: 0.0156 - 3s/epoch - 23ms/step\n",
      "Epoch 475/3000\n",
      "\n",
      "Epoch 475: val_mae did not improve from 0.01486\n",
      "151/151 - 3s - loss: 1.3157e-04 - mae: 0.0071 - val_loss: 4.6228e-04 - val_mae: 0.0158 - 3s/epoch - 23ms/step\n",
      "Epoch 476/3000\n",
      "\n",
      "Epoch 476: val_mae improved from 0.01486 to 0.01480, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 4s - loss: 1.3160e-04 - mae: 0.0071 - val_loss: 4.0701e-04 - val_mae: 0.0148 - 4s/epoch - 24ms/step\n",
      "Epoch 477/3000\n",
      "\n",
      "Epoch 477: val_mae did not improve from 0.01480\n",
      "151/151 - 4s - loss: 1.3162e-04 - mae: 0.0071 - val_loss: 4.3174e-04 - val_mae: 0.0153 - 4s/epoch - 25ms/step\n",
      "Epoch 478/3000\n",
      "\n",
      "Epoch 478: val_mae did not improve from 0.01480\n",
      "151/151 - 4s - loss: 1.3202e-04 - mae: 0.0071 - val_loss: 4.6655e-04 - val_mae: 0.0159 - 4s/epoch - 24ms/step\n",
      "Epoch 479/3000\n",
      "\n",
      "Epoch 479: val_mae did not improve from 0.01480\n",
      "151/151 - 4s - loss: 1.3142e-04 - mae: 0.0071 - val_loss: 4.1464e-04 - val_mae: 0.0150 - 4s/epoch - 24ms/step\n",
      "Epoch 480/3000\n",
      "\n",
      "Epoch 480: val_mae did not improve from 0.01480\n",
      "151/151 - 3s - loss: 1.3110e-04 - mae: 0.0071 - val_loss: 4.2580e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 481/3000\n",
      "\n",
      "Epoch 481: val_mae did not improve from 0.01480\n",
      "151/151 - 4s - loss: 1.3102e-04 - mae: 0.0071 - val_loss: 4.7158e-04 - val_mae: 0.0158 - 4s/epoch - 23ms/step\n",
      "Epoch 482/3000\n",
      "\n",
      "Epoch 482: val_mae did not improve from 0.01480\n",
      "151/151 - 3s - loss: 1.3166e-04 - mae: 0.0071 - val_loss: 4.3594e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 483/3000\n",
      "\n",
      "Epoch 483: val_mae did not improve from 0.01480\n",
      "151/151 - 3s - loss: 1.3103e-04 - mae: 0.0071 - val_loss: 4.3234e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 484/3000\n",
      "\n",
      "Epoch 484: val_mae did not improve from 0.01480\n",
      "151/151 - 3s - loss: 1.3133e-04 - mae: 0.0071 - val_loss: 4.2624e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 485/3000\n",
      "\n",
      "Epoch 485: val_mae did not improve from 0.01480\n",
      "151/151 - 3s - loss: 1.3156e-04 - mae: 0.0071 - val_loss: 4.3180e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 486/3000\n",
      "\n",
      "Epoch 486: val_mae did not improve from 0.01480\n",
      "151/151 - 3s - loss: 1.3108e-04 - mae: 0.0071 - val_loss: 4.5610e-04 - val_mae: 0.0157 - 3s/epoch - 22ms/step\n",
      "Epoch 487/3000\n",
      "\n",
      "Epoch 487: val_mae did not improve from 0.01480\n",
      "151/151 - 3s - loss: 1.3174e-04 - mae: 0.0071 - val_loss: 4.1238e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 488/3000\n",
      "\n",
      "Epoch 488: val_mae did not improve from 0.01480\n",
      "151/151 - 3s - loss: 1.3132e-04 - mae: 0.0071 - val_loss: 4.2649e-04 - val_mae: 0.0152 - 3s/epoch - 21ms/step\n",
      "Epoch 489/3000\n",
      "\n",
      "Epoch 489: val_mae improved from 0.01480 to 0.01471, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 1.3153e-04 - mae: 0.0071 - val_loss: 4.0156e-04 - val_mae: 0.0147 - 3s/epoch - 23ms/step\n",
      "Epoch 490/3000\n",
      "\n",
      "Epoch 490: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.3137e-04 - mae: 0.0071 - val_loss: 4.2745e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 491/3000\n",
      "\n",
      "Epoch 491: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.3064e-04 - mae: 0.0071 - val_loss: 4.3651e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 492/3000\n",
      "\n",
      "Epoch 492: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.3057e-04 - mae: 0.0071 - val_loss: 4.6457e-04 - val_mae: 0.0158 - 3s/epoch - 22ms/step\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 493/3000\n",
      "\n",
      "Epoch 493: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.3086e-04 - mae: 0.0071 - val_loss: 4.4168e-04 - val_mae: 0.0154 - 4s/epoch - 23ms/step\n",
      "Epoch 494/3000\n",
      "\n",
      "Epoch 494: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.3055e-04 - mae: 0.0071 - val_loss: 4.5791e-04 - val_mae: 0.0157 - 4s/epoch - 24ms/step\n",
      "Epoch 495/3000\n",
      "\n",
      "Epoch 495: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.3064e-04 - mae: 0.0071 - val_loss: 4.3324e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 496/3000\n",
      "\n",
      "Epoch 496: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.3127e-04 - mae: 0.0071 - val_loss: 4.2060e-04 - val_mae: 0.0152 - 4s/epoch - 23ms/step\n",
      "Epoch 497/3000\n",
      "\n",
      "Epoch 497: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.3040e-04 - mae: 0.0071 - val_loss: 4.5623e-04 - val_mae: 0.0157 - 3s/epoch - 23ms/step\n",
      "Epoch 498/3000\n",
      "\n",
      "Epoch 498: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.3034e-04 - mae: 0.0071 - val_loss: 4.0536e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 499/3000\n",
      "\n",
      "Epoch 499: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.3033e-04 - mae: 0.0070 - val_loss: 4.3415e-04 - val_mae: 0.0153 - 4s/epoch - 23ms/step\n",
      "Epoch 500/3000\n",
      "\n",
      "Epoch 500: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.3041e-04 - mae: 0.0071 - val_loss: 4.1751e-04 - val_mae: 0.0151 - 4s/epoch - 24ms/step\n",
      "Epoch 501/3000\n",
      "\n",
      "Epoch 501: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.3057e-04 - mae: 0.0071 - val_loss: 4.4526e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 502/3000\n",
      "\n",
      "Epoch 502: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.3064e-04 - mae: 0.0071 - val_loss: 4.3279e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 503/3000\n",
      "\n",
      "Epoch 503: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.3059e-04 - mae: 0.0071 - val_loss: 4.3889e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 504/3000\n",
      "\n",
      "Epoch 504: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.3085e-04 - mae: 0.0071 - val_loss: 4.4463e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 505/3000\n",
      "\n",
      "Epoch 505: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.3006e-04 - mae: 0.0070 - val_loss: 4.2669e-04 - val_mae: 0.0152 - 4s/epoch - 24ms/step\n",
      "Epoch 506/3000\n",
      "\n",
      "Epoch 506: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.3007e-04 - mae: 0.0070 - val_loss: 4.3189e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 507/3000\n",
      "\n",
      "Epoch 507: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.3022e-04 - mae: 0.0070 - val_loss: 4.5354e-04 - val_mae: 0.0157 - 4s/epoch - 23ms/step\n",
      "Epoch 508/3000\n",
      "\n",
      "Epoch 508: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.2982e-04 - mae: 0.0070 - val_loss: 4.3173e-04 - val_mae: 0.0153 - 4s/epoch - 23ms/step\n",
      "Epoch 509/3000\n",
      "\n",
      "Epoch 509: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.2989e-04 - mae: 0.0070 - val_loss: 4.1870e-04 - val_mae: 0.0150 - 4s/epoch - 24ms/step\n",
      "Epoch 510/3000\n",
      "\n",
      "Epoch 510: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.2993e-04 - mae: 0.0070 - val_loss: 4.2786e-04 - val_mae: 0.0152 - 4s/epoch - 23ms/step\n",
      "Epoch 511/3000\n",
      "\n",
      "Epoch 511: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.3028e-04 - mae: 0.0071 - val_loss: 4.4897e-04 - val_mae: 0.0155 - 4s/epoch - 23ms/step\n",
      "Epoch 512/3000\n",
      "\n",
      "Epoch 512: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.3048e-04 - mae: 0.0071 - val_loss: 4.3779e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 513/3000\n",
      "\n",
      "Epoch 513: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.2995e-04 - mae: 0.0070 - val_loss: 4.1729e-04 - val_mae: 0.0150 - 4s/epoch - 24ms/step\n",
      "Epoch 514/3000\n",
      "\n",
      "Epoch 514: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.3004e-04 - mae: 0.0070 - val_loss: 4.3541e-04 - val_mae: 0.0154 - 4s/epoch - 23ms/step\n",
      "Epoch 515/3000\n",
      "\n",
      "Epoch 515: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.2969e-04 - mae: 0.0070 - val_loss: 4.2471e-04 - val_mae: 0.0152 - 4s/epoch - 24ms/step\n",
      "Epoch 516/3000\n",
      "\n",
      "Epoch 516: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2986e-04 - mae: 0.0070 - val_loss: 4.4796e-04 - val_mae: 0.0156 - 3s/epoch - 23ms/step\n",
      "Epoch 517/3000\n",
      "\n",
      "Epoch 517: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.3000e-04 - mae: 0.0070 - val_loss: 4.5541e-04 - val_mae: 0.0156 - 3s/epoch - 23ms/step\n",
      "Epoch 518/3000\n",
      "\n",
      "Epoch 518: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.2985e-04 - mae: 0.0070 - val_loss: 4.2627e-04 - val_mae: 0.0152 - 4s/epoch - 23ms/step\n",
      "Epoch 519/3000\n",
      "\n",
      "Epoch 519: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.2956e-04 - mae: 0.0070 - val_loss: 4.3698e-04 - val_mae: 0.0154 - 4s/epoch - 24ms/step\n",
      "Epoch 520/3000\n",
      "\n",
      "Epoch 520: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2995e-04 - mae: 0.0070 - val_loss: 4.4102e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 521/3000\n",
      "\n",
      "Epoch 521: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2943e-04 - mae: 0.0070 - val_loss: 4.4234e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 522/3000\n",
      "\n",
      "Epoch 522: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2953e-04 - mae: 0.0070 - val_loss: 4.5723e-04 - val_mae: 0.0157 - 3s/epoch - 22ms/step\n",
      "Epoch 523/3000\n",
      "\n",
      "Epoch 523: val_mae did not improve from 0.01471\n",
      "151/151 - 4s - loss: 1.2968e-04 - mae: 0.0070 - val_loss: 4.1975e-04 - val_mae: 0.0151 - 4s/epoch - 23ms/step\n",
      "Epoch 524/3000\n",
      "\n",
      "Epoch 524: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2952e-04 - mae: 0.0070 - val_loss: 4.5604e-04 - val_mae: 0.0157 - 3s/epoch - 23ms/step\n",
      "Epoch 525/3000\n",
      "\n",
      "Epoch 525: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2938e-04 - mae: 0.0070 - val_loss: 4.4596e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 526/3000\n",
      "\n",
      "Epoch 526: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2913e-04 - mae: 0.0070 - val_loss: 4.3548e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 527/3000\n",
      "\n",
      "Epoch 527: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2942e-04 - mae: 0.0070 - val_loss: 4.3911e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 528/3000\n",
      "\n",
      "Epoch 528: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2971e-04 - mae: 0.0070 - val_loss: 4.1581e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 529/3000\n",
      "\n",
      "Epoch 529: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2996e-04 - mae: 0.0071 - val_loss: 4.4832e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 530/3000\n",
      "\n",
      "Epoch 530: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2910e-04 - mae: 0.0070 - val_loss: 4.2749e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 531/3000\n",
      "\n",
      "Epoch 531: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2921e-04 - mae: 0.0070 - val_loss: 4.3533e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 532/3000\n",
      "\n",
      "Epoch 532: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2925e-04 - mae: 0.0070 - val_loss: 4.6529e-04 - val_mae: 0.0159 - 3s/epoch - 22ms/step\n",
      "Epoch 533/3000\n",
      "\n",
      "Epoch 533: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2927e-04 - mae: 0.0070 - val_loss: 4.4588e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 534/3000\n",
      "\n",
      "Epoch 534: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2858e-04 - mae: 0.0070 - val_loss: 4.7593e-04 - val_mae: 0.0160 - 3s/epoch - 21ms/step\n",
      "Epoch 535/3000\n",
      "\n",
      "Epoch 535: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2968e-04 - mae: 0.0070 - val_loss: 4.5623e-04 - val_mae: 0.0156 - 3s/epoch - 21ms/step\n",
      "Epoch 536/3000\n",
      "\n",
      "Epoch 536: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2894e-04 - mae: 0.0070 - val_loss: 4.3503e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 537/3000\n",
      "\n",
      "Epoch 537: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2952e-04 - mae: 0.0070 - val_loss: 4.6319e-04 - val_mae: 0.0158 - 3s/epoch - 22ms/step\n",
      "Epoch 538/3000\n",
      "\n",
      "Epoch 538: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2897e-04 - mae: 0.0070 - val_loss: 4.8982e-04 - val_mae: 0.0162 - 3s/epoch - 22ms/step\n",
      "Epoch 539/3000\n",
      "\n",
      "Epoch 539: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2916e-04 - mae: 0.0070 - val_loss: 4.4537e-04 - val_mae: 0.0155 - 3s/epoch - 21ms/step\n",
      "Epoch 540/3000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 540: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2883e-04 - mae: 0.0070 - val_loss: 4.3331e-04 - val_mae: 0.0153 - 3s/epoch - 21ms/step\n",
      "Epoch 541/3000\n",
      "\n",
      "Epoch 541: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2876e-04 - mae: 0.0070 - val_loss: 4.6602e-04 - val_mae: 0.0159 - 3s/epoch - 21ms/step\n",
      "Epoch 542/3000\n",
      "\n",
      "Epoch 542: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2877e-04 - mae: 0.0070 - val_loss: 4.4041e-04 - val_mae: 0.0154 - 3s/epoch - 21ms/step\n",
      "Epoch 543/3000\n",
      "\n",
      "Epoch 543: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2840e-04 - mae: 0.0070 - val_loss: 4.3117e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 544/3000\n",
      "\n",
      "Epoch 544: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2874e-04 - mae: 0.0070 - val_loss: 4.0057e-04 - val_mae: 0.0148 - 3s/epoch - 21ms/step\n",
      "Epoch 545/3000\n",
      "\n",
      "Epoch 545: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2899e-04 - mae: 0.0070 - val_loss: 4.3220e-04 - val_mae: 0.0153 - 3s/epoch - 21ms/step\n",
      "Epoch 546/3000\n",
      "\n",
      "Epoch 546: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2855e-04 - mae: 0.0070 - val_loss: 4.0900e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 547/3000\n",
      "\n",
      "Epoch 547: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2884e-04 - mae: 0.0070 - val_loss: 4.3853e-04 - val_mae: 0.0153 - 3s/epoch - 21ms/step\n",
      "Epoch 548/3000\n",
      "\n",
      "Epoch 548: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2852e-04 - mae: 0.0070 - val_loss: 4.4463e-04 - val_mae: 0.0155 - 3s/epoch - 21ms/step\n",
      "Epoch 549/3000\n",
      "\n",
      "Epoch 549: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2850e-04 - mae: 0.0070 - val_loss: 4.5871e-04 - val_mae: 0.0158 - 3s/epoch - 21ms/step\n",
      "Epoch 550/3000\n",
      "\n",
      "Epoch 550: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2870e-04 - mae: 0.0070 - val_loss: 4.1260e-04 - val_mae: 0.0149 - 3s/epoch - 21ms/step\n",
      "Epoch 551/3000\n",
      "\n",
      "Epoch 551: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2881e-04 - mae: 0.0070 - val_loss: 4.3553e-04 - val_mae: 0.0153 - 3s/epoch - 21ms/step\n",
      "Epoch 552/3000\n",
      "\n",
      "Epoch 552: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2821e-04 - mae: 0.0070 - val_loss: 4.1977e-04 - val_mae: 0.0151 - 3s/epoch - 21ms/step\n",
      "Epoch 553/3000\n",
      "\n",
      "Epoch 553: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2856e-04 - mae: 0.0070 - val_loss: 4.1605e-04 - val_mae: 0.0150 - 3s/epoch - 21ms/step\n",
      "Epoch 554/3000\n",
      "\n",
      "Epoch 554: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2841e-04 - mae: 0.0070 - val_loss: 4.3883e-04 - val_mae: 0.0153 - 3s/epoch - 21ms/step\n",
      "Epoch 555/3000\n",
      "\n",
      "Epoch 555: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2842e-04 - mae: 0.0070 - val_loss: 4.2747e-04 - val_mae: 0.0152 - 3s/epoch - 21ms/step\n",
      "Epoch 556/3000\n",
      "\n",
      "Epoch 556: val_mae did not improve from 0.01471\n",
      "151/151 - 3s - loss: 1.2842e-04 - mae: 0.0070 - val_loss: 4.2189e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 557/3000\n",
      "\n",
      "Epoch 557: val_mae improved from 0.01471 to 0.01450, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 1.2867e-04 - mae: 0.0070 - val_loss: 3.8829e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 558/3000\n",
      "\n",
      "Epoch 558: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2843e-04 - mae: 0.0070 - val_loss: 4.4383e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 559/3000\n",
      "\n",
      "Epoch 559: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2805e-04 - mae: 0.0070 - val_loss: 4.2050e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 560/3000\n",
      "\n",
      "Epoch 560: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2864e-04 - mae: 0.0070 - val_loss: 4.8924e-04 - val_mae: 0.0164 - 3s/epoch - 22ms/step\n",
      "Epoch 561/3000\n",
      "\n",
      "Epoch 561: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2834e-04 - mae: 0.0070 - val_loss: 4.3990e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 562/3000\n",
      "\n",
      "Epoch 562: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2780e-04 - mae: 0.0070 - val_loss: 4.4456e-04 - val_mae: 0.0155 - 3s/epoch - 21ms/step\n",
      "Epoch 563/3000\n",
      "\n",
      "Epoch 563: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2822e-04 - mae: 0.0070 - val_loss: 4.7623e-04 - val_mae: 0.0161 - 3s/epoch - 22ms/step\n",
      "Epoch 564/3000\n",
      "\n",
      "Epoch 564: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2777e-04 - mae: 0.0070 - val_loss: 4.4900e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 565/3000\n",
      "\n",
      "Epoch 565: val_mae did not improve from 0.01450\n",
      "151/151 - 4s - loss: 1.2780e-04 - mae: 0.0070 - val_loss: 4.3948e-04 - val_mae: 0.0154 - 4s/epoch - 23ms/step\n",
      "Epoch 566/3000\n",
      "\n",
      "Epoch 566: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2835e-04 - mae: 0.0070 - val_loss: 4.1660e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 567/3000\n",
      "\n",
      "Epoch 567: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2795e-04 - mae: 0.0070 - val_loss: 4.1822e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 568/3000\n",
      "\n",
      "Epoch 568: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2753e-04 - mae: 0.0069 - val_loss: 4.4935e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 569/3000\n",
      "\n",
      "Epoch 569: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2784e-04 - mae: 0.0070 - val_loss: 4.4719e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 570/3000\n",
      "\n",
      "Epoch 570: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2771e-04 - mae: 0.0070 - val_loss: 4.4453e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 571/3000\n",
      "\n",
      "Epoch 571: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2790e-04 - mae: 0.0070 - val_loss: 4.3545e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 572/3000\n",
      "\n",
      "Epoch 572: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2793e-04 - mae: 0.0070 - val_loss: 4.2372e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 573/3000\n",
      "\n",
      "Epoch 573: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2828e-04 - mae: 0.0070 - val_loss: 4.4532e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 574/3000\n",
      "\n",
      "Epoch 574: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2755e-04 - mae: 0.0070 - val_loss: 4.6532e-04 - val_mae: 0.0157 - 3s/epoch - 22ms/step\n",
      "Epoch 575/3000\n",
      "\n",
      "Epoch 575: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2745e-04 - mae: 0.0069 - val_loss: 4.0913e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 576/3000\n",
      "\n",
      "Epoch 576: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2786e-04 - mae: 0.0070 - val_loss: 4.3973e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 577/3000\n",
      "\n",
      "Epoch 577: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2786e-04 - mae: 0.0070 - val_loss: 4.1680e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 578/3000\n",
      "\n",
      "Epoch 578: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2782e-04 - mae: 0.0070 - val_loss: 4.1841e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 579/3000\n",
      "\n",
      "Epoch 579: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2699e-04 - mae: 0.0069 - val_loss: 4.5378e-04 - val_mae: 0.0156 - 3s/epoch - 21ms/step\n",
      "Epoch 580/3000\n",
      "\n",
      "Epoch 580: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2741e-04 - mae: 0.0069 - val_loss: 4.4988e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 581/3000\n",
      "\n",
      "Epoch 581: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2762e-04 - mae: 0.0070 - val_loss: 4.0755e-04 - val_mae: 0.0148 - 3s/epoch - 21ms/step\n",
      "Epoch 582/3000\n",
      "\n",
      "Epoch 582: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2770e-04 - mae: 0.0070 - val_loss: 4.0502e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 583/3000\n",
      "\n",
      "Epoch 583: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2744e-04 - mae: 0.0069 - val_loss: 4.1137e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 584/3000\n",
      "\n",
      "Epoch 584: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2755e-04 - mae: 0.0070 - val_loss: 4.3400e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 585/3000\n",
      "\n",
      "Epoch 585: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2719e-04 - mae: 0.0069 - val_loss: 4.0903e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 586/3000\n",
      "\n",
      "Epoch 586: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2718e-04 - mae: 0.0069 - val_loss: 4.0940e-04 - val_mae: 0.0149 - 3s/epoch - 21ms/step\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 587/3000\n",
      "\n",
      "Epoch 587: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2775e-04 - mae: 0.0070 - val_loss: 4.3264e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 588/3000\n",
      "\n",
      "Epoch 588: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2672e-04 - mae: 0.0069 - val_loss: 4.6603e-04 - val_mae: 0.0158 - 3s/epoch - 21ms/step\n",
      "Epoch 589/3000\n",
      "\n",
      "Epoch 589: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2720e-04 - mae: 0.0069 - val_loss: 4.0458e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 590/3000\n",
      "\n",
      "Epoch 590: val_mae did not improve from 0.01450\n",
      "151/151 - 4s - loss: 1.2724e-04 - mae: 0.0069 - val_loss: 4.5458e-04 - val_mae: 0.0157 - 4s/epoch - 24ms/step\n",
      "Epoch 591/3000\n",
      "\n",
      "Epoch 591: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2698e-04 - mae: 0.0069 - val_loss: 4.3695e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 592/3000\n",
      "\n",
      "Epoch 592: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2702e-04 - mae: 0.0069 - val_loss: 4.3048e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 593/3000\n",
      "\n",
      "Epoch 593: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2683e-04 - mae: 0.0069 - val_loss: 4.2854e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 594/3000\n",
      "\n",
      "Epoch 594: val_mae did not improve from 0.01450\n",
      "151/151 - 4s - loss: 1.2704e-04 - mae: 0.0069 - val_loss: 4.3618e-04 - val_mae: 0.0153 - 4s/epoch - 24ms/step\n",
      "Epoch 595/3000\n",
      "\n",
      "Epoch 595: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2696e-04 - mae: 0.0069 - val_loss: 4.5464e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 596/3000\n",
      "\n",
      "Epoch 596: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2671e-04 - mae: 0.0069 - val_loss: 4.1315e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 597/3000\n",
      "\n",
      "Epoch 597: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2676e-04 - mae: 0.0069 - val_loss: 4.2007e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 598/3000\n",
      "\n",
      "Epoch 598: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2693e-04 - mae: 0.0069 - val_loss: 4.2850e-04 - val_mae: 0.0151 - 3s/epoch - 21ms/step\n",
      "Epoch 599/3000\n",
      "\n",
      "Epoch 599: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2691e-04 - mae: 0.0069 - val_loss: 4.0416e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 600/3000\n",
      "\n",
      "Epoch 600: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2743e-04 - mae: 0.0070 - val_loss: 4.3729e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 601/3000\n",
      "\n",
      "Epoch 601: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2654e-04 - mae: 0.0069 - val_loss: 4.4925e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 602/3000\n",
      "\n",
      "Epoch 602: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2636e-04 - mae: 0.0069 - val_loss: 4.6036e-04 - val_mae: 0.0157 - 3s/epoch - 21ms/step\n",
      "Epoch 603/3000\n",
      "\n",
      "Epoch 603: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2652e-04 - mae: 0.0069 - val_loss: 4.2989e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 604/3000\n",
      "\n",
      "Epoch 604: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2752e-04 - mae: 0.0070 - val_loss: 4.2549e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 605/3000\n",
      "\n",
      "Epoch 605: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2676e-04 - mae: 0.0069 - val_loss: 4.1087e-04 - val_mae: 0.0149 - 3s/epoch - 21ms/step\n",
      "Epoch 606/3000\n",
      "\n",
      "Epoch 606: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2668e-04 - mae: 0.0069 - val_loss: 4.4640e-04 - val_mae: 0.0154 - 3s/epoch - 21ms/step\n",
      "Epoch 607/3000\n",
      "\n",
      "Epoch 607: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2673e-04 - mae: 0.0069 - val_loss: 4.5362e-04 - val_mae: 0.0155 - 3s/epoch - 21ms/step\n",
      "Epoch 608/3000\n",
      "\n",
      "Epoch 608: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2670e-04 - mae: 0.0069 - val_loss: 3.8865e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 609/3000\n",
      "\n",
      "Epoch 609: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2710e-04 - mae: 0.0069 - val_loss: 4.4640e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 610/3000\n",
      "\n",
      "Epoch 610: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2627e-04 - mae: 0.0069 - val_loss: 4.2125e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 611/3000\n",
      "\n",
      "Epoch 611: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2667e-04 - mae: 0.0069 - val_loss: 4.3201e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 612/3000\n",
      "\n",
      "Epoch 612: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2626e-04 - mae: 0.0069 - val_loss: 4.4979e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 613/3000\n",
      "\n",
      "Epoch 613: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2660e-04 - mae: 0.0069 - val_loss: 4.3702e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 614/3000\n",
      "\n",
      "Epoch 614: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2616e-04 - mae: 0.0069 - val_loss: 4.2868e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 615/3000\n",
      "\n",
      "Epoch 615: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2622e-04 - mae: 0.0069 - val_loss: 4.4396e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 616/3000\n",
      "\n",
      "Epoch 616: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2620e-04 - mae: 0.0069 - val_loss: 4.4114e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 617/3000\n",
      "\n",
      "Epoch 617: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2610e-04 - mae: 0.0069 - val_loss: 4.1999e-04 - val_mae: 0.0150 - 3s/epoch - 21ms/step\n",
      "Epoch 618/3000\n",
      "\n",
      "Epoch 618: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2627e-04 - mae: 0.0069 - val_loss: 4.6751e-04 - val_mae: 0.0159 - 3s/epoch - 22ms/step\n",
      "Epoch 619/3000\n",
      "\n",
      "Epoch 619: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2636e-04 - mae: 0.0069 - val_loss: 4.2378e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 620/3000\n",
      "\n",
      "Epoch 620: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2640e-04 - mae: 0.0069 - val_loss: 4.2455e-04 - val_mae: 0.0151 - 3s/epoch - 21ms/step\n",
      "Epoch 621/3000\n",
      "\n",
      "Epoch 621: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2628e-04 - mae: 0.0069 - val_loss: 4.0795e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 622/3000\n",
      "\n",
      "Epoch 622: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2595e-04 - mae: 0.0069 - val_loss: 4.3954e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 623/3000\n",
      "\n",
      "Epoch 623: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2595e-04 - mae: 0.0069 - val_loss: 4.3048e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 624/3000\n",
      "\n",
      "Epoch 624: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2616e-04 - mae: 0.0069 - val_loss: 4.3995e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 625/3000\n",
      "\n",
      "Epoch 625: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2634e-04 - mae: 0.0069 - val_loss: 4.4285e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 626/3000\n",
      "\n",
      "Epoch 626: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2599e-04 - mae: 0.0069 - val_loss: 4.4124e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 627/3000\n",
      "\n",
      "Epoch 627: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2591e-04 - mae: 0.0069 - val_loss: 4.3728e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 628/3000\n",
      "\n",
      "Epoch 628: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2583e-04 - mae: 0.0069 - val_loss: 4.3869e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 629/3000\n",
      "\n",
      "Epoch 629: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2623e-04 - mae: 0.0069 - val_loss: 4.2458e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 630/3000\n",
      "\n",
      "Epoch 630: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2588e-04 - mae: 0.0069 - val_loss: 4.2727e-04 - val_mae: 0.0151 - 3s/epoch - 21ms/step\n",
      "Epoch 631/3000\n",
      "\n",
      "Epoch 631: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2594e-04 - mae: 0.0069 - val_loss: 4.3655e-04 - val_mae: 0.0154 - 3s/epoch - 21ms/step\n",
      "Epoch 632/3000\n",
      "\n",
      "Epoch 632: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2570e-04 - mae: 0.0069 - val_loss: 4.2705e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 633/3000\n",
      "\n",
      "Epoch 633: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2608e-04 - mae: 0.0069 - val_loss: 4.3045e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 634/3000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 634: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2575e-04 - mae: 0.0069 - val_loss: 4.2340e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 635/3000\n",
      "\n",
      "Epoch 635: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2562e-04 - mae: 0.0069 - val_loss: 4.6265e-04 - val_mae: 0.0157 - 3s/epoch - 23ms/step\n",
      "Epoch 636/3000\n",
      "\n",
      "Epoch 636: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2581e-04 - mae: 0.0069 - val_loss: 4.1589e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 637/3000\n",
      "\n",
      "Epoch 637: val_mae did not improve from 0.01450\n",
      "151/151 - 4s - loss: 1.2627e-04 - mae: 0.0069 - val_loss: 4.1006e-04 - val_mae: 0.0149 - 4s/epoch - 23ms/step\n",
      "Epoch 638/3000\n",
      "\n",
      "Epoch 638: val_mae did not improve from 0.01450\n",
      "151/151 - 4s - loss: 1.2578e-04 - mae: 0.0069 - val_loss: 4.0745e-04 - val_mae: 0.0149 - 4s/epoch - 23ms/step\n",
      "Epoch 639/3000\n",
      "\n",
      "Epoch 639: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2572e-04 - mae: 0.0069 - val_loss: 4.4168e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 640/3000\n",
      "\n",
      "Epoch 640: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2531e-04 - mae: 0.0069 - val_loss: 4.0081e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 641/3000\n",
      "\n",
      "Epoch 641: val_mae did not improve from 0.01450\n",
      "151/151 - 4s - loss: 1.2555e-04 - mae: 0.0069 - val_loss: 4.3477e-04 - val_mae: 0.0153 - 4s/epoch - 24ms/step\n",
      "Epoch 642/3000\n",
      "\n",
      "Epoch 642: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2568e-04 - mae: 0.0069 - val_loss: 4.2135e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 643/3000\n",
      "\n",
      "Epoch 643: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2595e-04 - mae: 0.0069 - val_loss: 4.4672e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 644/3000\n",
      "\n",
      "Epoch 644: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2534e-04 - mae: 0.0069 - val_loss: 3.9960e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 645/3000\n",
      "\n",
      "Epoch 645: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2550e-04 - mae: 0.0069 - val_loss: 4.5780e-04 - val_mae: 0.0157 - 3s/epoch - 22ms/step\n",
      "Epoch 646/3000\n",
      "\n",
      "Epoch 646: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2535e-04 - mae: 0.0069 - val_loss: 4.3425e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 647/3000\n",
      "\n",
      "Epoch 647: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2554e-04 - mae: 0.0069 - val_loss: 4.4492e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 648/3000\n",
      "\n",
      "Epoch 648: val_mae did not improve from 0.01450\n",
      "151/151 - 4s - loss: 1.2590e-04 - mae: 0.0069 - val_loss: 4.2907e-04 - val_mae: 0.0151 - 4s/epoch - 23ms/step\n",
      "Epoch 649/3000\n",
      "\n",
      "Epoch 649: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2512e-04 - mae: 0.0069 - val_loss: 4.2761e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 650/3000\n",
      "\n",
      "Epoch 650: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2529e-04 - mae: 0.0069 - val_loss: 4.8523e-04 - val_mae: 0.0162 - 3s/epoch - 23ms/step\n",
      "Epoch 651/3000\n",
      "\n",
      "Epoch 651: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2522e-04 - mae: 0.0069 - val_loss: 4.2030e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 652/3000\n",
      "\n",
      "Epoch 652: val_mae did not improve from 0.01450\n",
      "151/151 - 4s - loss: 1.2505e-04 - mae: 0.0069 - val_loss: 4.2972e-04 - val_mae: 0.0152 - 4s/epoch - 23ms/step\n",
      "Epoch 653/3000\n",
      "\n",
      "Epoch 653: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2529e-04 - mae: 0.0069 - val_loss: 4.1586e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 654/3000\n",
      "\n",
      "Epoch 654: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2572e-04 - mae: 0.0069 - val_loss: 4.2473e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 655/3000\n",
      "\n",
      "Epoch 655: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2507e-04 - mae: 0.0069 - val_loss: 4.2468e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 656/3000\n",
      "\n",
      "Epoch 656: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2472e-04 - mae: 0.0068 - val_loss: 4.3752e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 657/3000\n",
      "\n",
      "Epoch 657: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2504e-04 - mae: 0.0069 - val_loss: 4.2683e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 658/3000\n",
      "\n",
      "Epoch 658: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2495e-04 - mae: 0.0069 - val_loss: 4.2566e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 659/3000\n",
      "\n",
      "Epoch 659: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2538e-04 - mae: 0.0069 - val_loss: 4.4192e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 660/3000\n",
      "\n",
      "Epoch 660: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2477e-04 - mae: 0.0069 - val_loss: 4.2169e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 661/3000\n",
      "\n",
      "Epoch 661: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2497e-04 - mae: 0.0069 - val_loss: 4.3405e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 662/3000\n",
      "\n",
      "Epoch 662: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2520e-04 - mae: 0.0069 - val_loss: 4.4235e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 663/3000\n",
      "\n",
      "Epoch 663: val_mae did not improve from 0.01450\n",
      "151/151 - 3s - loss: 1.2497e-04 - mae: 0.0069 - val_loss: 4.4967e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 664/3000\n",
      "\n",
      "Epoch 664: val_mae improved from 0.01450 to 0.01438, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 1.2512e-04 - mae: 0.0069 - val_loss: 3.8312e-04 - val_mae: 0.0144 - 3s/epoch - 22ms/step\n",
      "Epoch 665/3000\n",
      "\n",
      "Epoch 665: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2521e-04 - mae: 0.0069 - val_loss: 4.4568e-04 - val_mae: 0.0155 - 3s/epoch - 21ms/step\n",
      "Epoch 666/3000\n",
      "\n",
      "Epoch 666: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2484e-04 - mae: 0.0069 - val_loss: 4.3067e-04 - val_mae: 0.0153 - 3s/epoch - 21ms/step\n",
      "Epoch 667/3000\n",
      "\n",
      "Epoch 667: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2467e-04 - mae: 0.0068 - val_loss: 4.4959e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 668/3000\n",
      "\n",
      "Epoch 668: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2496e-04 - mae: 0.0069 - val_loss: 4.3885e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 669/3000\n",
      "\n",
      "Epoch 669: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2483e-04 - mae: 0.0069 - val_loss: 4.0791e-04 - val_mae: 0.0148 - 3s/epoch - 21ms/step\n",
      "Epoch 670/3000\n",
      "\n",
      "Epoch 670: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2451e-04 - mae: 0.0068 - val_loss: 4.4910e-04 - val_mae: 0.0155 - 3s/epoch - 21ms/step\n",
      "Epoch 671/3000\n",
      "\n",
      "Epoch 671: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2461e-04 - mae: 0.0068 - val_loss: 4.2382e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 672/3000\n",
      "\n",
      "Epoch 672: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2480e-04 - mae: 0.0069 - val_loss: 3.9275e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 673/3000\n",
      "\n",
      "Epoch 673: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2493e-04 - mae: 0.0069 - val_loss: 4.4293e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 674/3000\n",
      "\n",
      "Epoch 674: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2445e-04 - mae: 0.0068 - val_loss: 4.0873e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 675/3000\n",
      "\n",
      "Epoch 675: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2485e-04 - mae: 0.0069 - val_loss: 4.3193e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 676/3000\n",
      "\n",
      "Epoch 676: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2446e-04 - mae: 0.0068 - val_loss: 4.2168e-04 - val_mae: 0.0151 - 3s/epoch - 21ms/step\n",
      "Epoch 677/3000\n",
      "\n",
      "Epoch 677: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2449e-04 - mae: 0.0068 - val_loss: 4.2363e-04 - val_mae: 0.0151 - 3s/epoch - 21ms/step\n",
      "Epoch 678/3000\n",
      "\n",
      "Epoch 678: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2476e-04 - mae: 0.0069 - val_loss: 4.0934e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 679/3000\n",
      "\n",
      "Epoch 679: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2488e-04 - mae: 0.0069 - val_loss: 4.6798e-04 - val_mae: 0.0158 - 3s/epoch - 23ms/step\n",
      "Epoch 680/3000\n",
      "\n",
      "Epoch 680: val_mae did not improve from 0.01438\n",
      "151/151 - 4s - loss: 1.2472e-04 - mae: 0.0069 - val_loss: 4.4664e-04 - val_mae: 0.0154 - 4s/epoch - 23ms/step\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 681/3000\n",
      "\n",
      "Epoch 681: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2438e-04 - mae: 0.0068 - val_loss: 4.4999e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 682/3000\n",
      "\n",
      "Epoch 682: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2470e-04 - mae: 0.0069 - val_loss: 4.1605e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 683/3000\n",
      "\n",
      "Epoch 683: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2425e-04 - mae: 0.0068 - val_loss: 4.3341e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 684/3000\n",
      "\n",
      "Epoch 684: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2468e-04 - mae: 0.0069 - val_loss: 4.2188e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 685/3000\n",
      "\n",
      "Epoch 685: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2449e-04 - mae: 0.0068 - val_loss: 4.3997e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 686/3000\n",
      "\n",
      "Epoch 686: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2393e-04 - mae: 0.0068 - val_loss: 4.2707e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 687/3000\n",
      "\n",
      "Epoch 687: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2413e-04 - mae: 0.0068 - val_loss: 4.2621e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 688/3000\n",
      "\n",
      "Epoch 688: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2399e-04 - mae: 0.0068 - val_loss: 4.3415e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 689/3000\n",
      "\n",
      "Epoch 689: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2445e-04 - mae: 0.0068 - val_loss: 4.3606e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 690/3000\n",
      "\n",
      "Epoch 690: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2404e-04 - mae: 0.0068 - val_loss: 4.4888e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 691/3000\n",
      "\n",
      "Epoch 691: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2441e-04 - mae: 0.0068 - val_loss: 3.9915e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 692/3000\n",
      "\n",
      "Epoch 692: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2419e-04 - mae: 0.0068 - val_loss: 4.5191e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 693/3000\n",
      "\n",
      "Epoch 693: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2442e-04 - mae: 0.0068 - val_loss: 4.4990e-04 - val_mae: 0.0155 - 3s/epoch - 21ms/step\n",
      "Epoch 694/3000\n",
      "\n",
      "Epoch 694: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2417e-04 - mae: 0.0068 - val_loss: 4.1954e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 695/3000\n",
      "\n",
      "Epoch 695: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2425e-04 - mae: 0.0068 - val_loss: 4.0334e-04 - val_mae: 0.0148 - 3s/epoch - 21ms/step\n",
      "Epoch 696/3000\n",
      "\n",
      "Epoch 696: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2409e-04 - mae: 0.0068 - val_loss: 4.2735e-04 - val_mae: 0.0153 - 3s/epoch - 21ms/step\n",
      "Epoch 697/3000\n",
      "\n",
      "Epoch 697: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2410e-04 - mae: 0.0068 - val_loss: 4.3615e-04 - val_mae: 0.0154 - 3s/epoch - 21ms/step\n",
      "Epoch 698/3000\n",
      "\n",
      "Epoch 698: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2407e-04 - mae: 0.0068 - val_loss: 4.3459e-04 - val_mae: 0.0154 - 3s/epoch - 21ms/step\n",
      "Epoch 699/3000\n",
      "\n",
      "Epoch 699: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2432e-04 - mae: 0.0068 - val_loss: 4.3318e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 700/3000\n",
      "\n",
      "Epoch 700: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2383e-04 - mae: 0.0068 - val_loss: 4.2194e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 701/3000\n",
      "\n",
      "Epoch 701: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2401e-04 - mae: 0.0068 - val_loss: 4.0950e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 702/3000\n",
      "\n",
      "Epoch 702: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2406e-04 - mae: 0.0068 - val_loss: 3.9402e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 703/3000\n",
      "\n",
      "Epoch 703: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2375e-04 - mae: 0.0068 - val_loss: 4.1891e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 704/3000\n",
      "\n",
      "Epoch 704: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2360e-04 - mae: 0.0068 - val_loss: 4.1528e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 705/3000\n",
      "\n",
      "Epoch 705: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2413e-04 - mae: 0.0068 - val_loss: 4.3210e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 706/3000\n",
      "\n",
      "Epoch 706: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2385e-04 - mae: 0.0068 - val_loss: 4.2269e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 707/3000\n",
      "\n",
      "Epoch 707: val_mae did not improve from 0.01438\n",
      "151/151 - 4s - loss: 1.2383e-04 - mae: 0.0068 - val_loss: 4.0077e-04 - val_mae: 0.0147 - 4s/epoch - 23ms/step\n",
      "Epoch 708/3000\n",
      "\n",
      "Epoch 708: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2379e-04 - mae: 0.0068 - val_loss: 4.0653e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 709/3000\n",
      "\n",
      "Epoch 709: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2387e-04 - mae: 0.0068 - val_loss: 4.2342e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 710/3000\n",
      "\n",
      "Epoch 710: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2380e-04 - mae: 0.0068 - val_loss: 4.3943e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 711/3000\n",
      "\n",
      "Epoch 711: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2387e-04 - mae: 0.0068 - val_loss: 4.1044e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 712/3000\n",
      "\n",
      "Epoch 712: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2353e-04 - mae: 0.0068 - val_loss: 4.2881e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 713/3000\n",
      "\n",
      "Epoch 713: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2382e-04 - mae: 0.0068 - val_loss: 4.3887e-04 - val_mae: 0.0155 - 3s/epoch - 21ms/step\n",
      "Epoch 714/3000\n",
      "\n",
      "Epoch 714: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2391e-04 - mae: 0.0068 - val_loss: 4.3717e-04 - val_mae: 0.0153 - 3s/epoch - 21ms/step\n",
      "Epoch 715/3000\n",
      "\n",
      "Epoch 715: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2342e-04 - mae: 0.0068 - val_loss: 4.2115e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 716/3000\n",
      "\n",
      "Epoch 716: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2405e-04 - mae: 0.0068 - val_loss: 4.2197e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 717/3000\n",
      "\n",
      "Epoch 717: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2329e-04 - mae: 0.0068 - val_loss: 4.2989e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 718/3000\n",
      "\n",
      "Epoch 718: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2340e-04 - mae: 0.0068 - val_loss: 4.3555e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 719/3000\n",
      "\n",
      "Epoch 719: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2385e-04 - mae: 0.0068 - val_loss: 4.0753e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 720/3000\n",
      "\n",
      "Epoch 720: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2360e-04 - mae: 0.0068 - val_loss: 4.3951e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 721/3000\n",
      "\n",
      "Epoch 721: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2317e-04 - mae: 0.0068 - val_loss: 4.5678e-04 - val_mae: 0.0157 - 3s/epoch - 21ms/step\n",
      "Epoch 722/3000\n",
      "\n",
      "Epoch 722: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2382e-04 - mae: 0.0068 - val_loss: 4.2697e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 723/3000\n",
      "\n",
      "Epoch 723: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2356e-04 - mae: 0.0068 - val_loss: 4.1166e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 724/3000\n",
      "\n",
      "Epoch 724: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2323e-04 - mae: 0.0068 - val_loss: 4.1756e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 725/3000\n",
      "\n",
      "Epoch 725: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2329e-04 - mae: 0.0068 - val_loss: 4.0895e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 726/3000\n",
      "\n",
      "Epoch 726: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2331e-04 - mae: 0.0068 - val_loss: 4.0984e-04 - val_mae: 0.0149 - 3s/epoch - 21ms/step\n",
      "Epoch 727/3000\n",
      "\n",
      "Epoch 727: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2325e-04 - mae: 0.0068 - val_loss: 4.0372e-04 - val_mae: 0.0147 - 3s/epoch - 21ms/step\n",
      "Epoch 728/3000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 728: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2304e-04 - mae: 0.0068 - val_loss: 4.1032e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 729/3000\n",
      "\n",
      "Epoch 729: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2351e-04 - mae: 0.0068 - val_loss: 4.3423e-04 - val_mae: 0.0153 - 3s/epoch - 21ms/step\n",
      "Epoch 730/3000\n",
      "\n",
      "Epoch 730: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2313e-04 - mae: 0.0068 - val_loss: 4.0214e-04 - val_mae: 0.0147 - 3s/epoch - 21ms/step\n",
      "Epoch 731/3000\n",
      "\n",
      "Epoch 731: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2305e-04 - mae: 0.0068 - val_loss: 4.1889e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 732/3000\n",
      "\n",
      "Epoch 732: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2317e-04 - mae: 0.0068 - val_loss: 3.9428e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 733/3000\n",
      "\n",
      "Epoch 733: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2332e-04 - mae: 0.0068 - val_loss: 4.4774e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 734/3000\n",
      "\n",
      "Epoch 734: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2313e-04 - mae: 0.0068 - val_loss: 4.1806e-04 - val_mae: 0.0150 - 3s/epoch - 21ms/step\n",
      "Epoch 735/3000\n",
      "\n",
      "Epoch 735: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2319e-04 - mae: 0.0068 - val_loss: 4.1215e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 736/3000\n",
      "\n",
      "Epoch 736: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2319e-04 - mae: 0.0068 - val_loss: 4.2823e-04 - val_mae: 0.0152 - 3s/epoch - 21ms/step\n",
      "Epoch 737/3000\n",
      "\n",
      "Epoch 737: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2305e-04 - mae: 0.0068 - val_loss: 4.1668e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 738/3000\n",
      "\n",
      "Epoch 738: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2329e-04 - mae: 0.0068 - val_loss: 4.5596e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 739/3000\n",
      "\n",
      "Epoch 739: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2281e-04 - mae: 0.0068 - val_loss: 4.5286e-04 - val_mae: 0.0157 - 3s/epoch - 22ms/step\n",
      "Epoch 740/3000\n",
      "\n",
      "Epoch 740: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2279e-04 - mae: 0.0068 - val_loss: 4.3846e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 741/3000\n",
      "\n",
      "Epoch 741: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2327e-04 - mae: 0.0068 - val_loss: 4.2919e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 742/3000\n",
      "\n",
      "Epoch 742: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2277e-04 - mae: 0.0068 - val_loss: 4.2728e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 743/3000\n",
      "\n",
      "Epoch 743: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2325e-04 - mae: 0.0068 - val_loss: 4.1857e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 744/3000\n",
      "\n",
      "Epoch 744: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2311e-04 - mae: 0.0068 - val_loss: 4.1645e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 745/3000\n",
      "\n",
      "Epoch 745: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2328e-04 - mae: 0.0068 - val_loss: 4.2175e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 746/3000\n",
      "\n",
      "Epoch 746: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2312e-04 - mae: 0.0068 - val_loss: 4.2875e-04 - val_mae: 0.0151 - 3s/epoch - 21ms/step\n",
      "Epoch 747/3000\n",
      "\n",
      "Epoch 747: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2304e-04 - mae: 0.0068 - val_loss: 4.0910e-04 - val_mae: 0.0148 - 3s/epoch - 21ms/step\n",
      "Epoch 748/3000\n",
      "\n",
      "Epoch 748: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2272e-04 - mae: 0.0068 - val_loss: 4.5384e-04 - val_mae: 0.0156 - 3s/epoch - 21ms/step\n",
      "Epoch 749/3000\n",
      "\n",
      "Epoch 749: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2239e-04 - mae: 0.0068 - val_loss: 4.2277e-04 - val_mae: 0.0150 - 3s/epoch - 21ms/step\n",
      "Epoch 750/3000\n",
      "\n",
      "Epoch 750: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2284e-04 - mae: 0.0068 - val_loss: 4.1538e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 751/3000\n",
      "\n",
      "Epoch 751: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2292e-04 - mae: 0.0068 - val_loss: 4.3175e-04 - val_mae: 0.0152 - 3s/epoch - 21ms/step\n",
      "Epoch 752/3000\n",
      "\n",
      "Epoch 752: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2306e-04 - mae: 0.0068 - val_loss: 3.9975e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 753/3000\n",
      "\n",
      "Epoch 753: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2232e-04 - mae: 0.0068 - val_loss: 4.0865e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 754/3000\n",
      "\n",
      "Epoch 754: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2278e-04 - mae: 0.0068 - val_loss: 4.0752e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 755/3000\n",
      "\n",
      "Epoch 755: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2235e-04 - mae: 0.0068 - val_loss: 4.4715e-04 - val_mae: 0.0154 - 3s/epoch - 21ms/step\n",
      "Epoch 756/3000\n",
      "\n",
      "Epoch 756: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2227e-04 - mae: 0.0068 - val_loss: 4.0499e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 757/3000\n",
      "\n",
      "Epoch 757: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2288e-04 - mae: 0.0068 - val_loss: 4.2871e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 758/3000\n",
      "\n",
      "Epoch 758: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2267e-04 - mae: 0.0068 - val_loss: 4.1497e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 759/3000\n",
      "\n",
      "Epoch 759: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2252e-04 - mae: 0.0068 - val_loss: 4.4875e-04 - val_mae: 0.0155 - 3s/epoch - 21ms/step\n",
      "Epoch 760/3000\n",
      "\n",
      "Epoch 760: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2251e-04 - mae: 0.0068 - val_loss: 3.8498e-04 - val_mae: 0.0145 - 3s/epoch - 21ms/step\n",
      "Epoch 761/3000\n",
      "\n",
      "Epoch 761: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2254e-04 - mae: 0.0068 - val_loss: 4.2953e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 762/3000\n",
      "\n",
      "Epoch 762: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2254e-04 - mae: 0.0068 - val_loss: 3.9683e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 763/3000\n",
      "\n",
      "Epoch 763: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2243e-04 - mae: 0.0068 - val_loss: 4.1208e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 764/3000\n",
      "\n",
      "Epoch 764: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2240e-04 - mae: 0.0068 - val_loss: 4.1003e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 765/3000\n",
      "\n",
      "Epoch 765: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2256e-04 - mae: 0.0068 - val_loss: 4.1792e-04 - val_mae: 0.0150 - 3s/epoch - 21ms/step\n",
      "Epoch 766/3000\n",
      "\n",
      "Epoch 766: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2211e-04 - mae: 0.0067 - val_loss: 4.0917e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 767/3000\n",
      "\n",
      "Epoch 767: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2233e-04 - mae: 0.0068 - val_loss: 4.2262e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 768/3000\n",
      "\n",
      "Epoch 768: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2243e-04 - mae: 0.0068 - val_loss: 4.3205e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 769/3000\n",
      "\n",
      "Epoch 769: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2276e-04 - mae: 0.0068 - val_loss: 4.3383e-04 - val_mae: 0.0153 - 3s/epoch - 21ms/step\n",
      "Epoch 770/3000\n",
      "\n",
      "Epoch 770: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2244e-04 - mae: 0.0068 - val_loss: 4.1386e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 771/3000\n",
      "\n",
      "Epoch 771: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2226e-04 - mae: 0.0068 - val_loss: 4.2661e-04 - val_mae: 0.0151 - 3s/epoch - 21ms/step\n",
      "Epoch 772/3000\n",
      "\n",
      "Epoch 772: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2234e-04 - mae: 0.0068 - val_loss: 4.1544e-04 - val_mae: 0.0149 - 3s/epoch - 21ms/step\n",
      "Epoch 773/3000\n",
      "\n",
      "Epoch 773: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2241e-04 - mae: 0.0068 - val_loss: 4.3151e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 774/3000\n",
      "\n",
      "Epoch 774: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2238e-04 - mae: 0.0068 - val_loss: 4.2915e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 775/3000\n",
      "\n",
      "Epoch 775: val_mae did not improve from 0.01438\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "151/151 - 3s - loss: 1.2215e-04 - mae: 0.0068 - val_loss: 4.3322e-04 - val_mae: 0.0152 - 3s/epoch - 21ms/step\n",
      "Epoch 776/3000\n",
      "\n",
      "Epoch 776: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2208e-04 - mae: 0.0068 - val_loss: 4.2667e-04 - val_mae: 0.0151 - 3s/epoch - 21ms/step\n",
      "Epoch 777/3000\n",
      "\n",
      "Epoch 777: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2201e-04 - mae: 0.0068 - val_loss: 4.3022e-04 - val_mae: 0.0153 - 3s/epoch - 21ms/step\n",
      "Epoch 778/3000\n",
      "\n",
      "Epoch 778: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2214e-04 - mae: 0.0068 - val_loss: 4.2237e-04 - val_mae: 0.0151 - 3s/epoch - 21ms/step\n",
      "Epoch 779/3000\n",
      "\n",
      "Epoch 779: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2212e-04 - mae: 0.0068 - val_loss: 4.0663e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 780/3000\n",
      "\n",
      "Epoch 780: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2198e-04 - mae: 0.0068 - val_loss: 4.3276e-04 - val_mae: 0.0152 - 3s/epoch - 21ms/step\n",
      "Epoch 781/3000\n",
      "\n",
      "Epoch 781: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2198e-04 - mae: 0.0067 - val_loss: 4.3499e-04 - val_mae: 0.0153 - 3s/epoch - 21ms/step\n",
      "Epoch 782/3000\n",
      "\n",
      "Epoch 782: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2185e-04 - mae: 0.0067 - val_loss: 4.1575e-04 - val_mae: 0.0149 - 3s/epoch - 21ms/step\n",
      "Epoch 783/3000\n",
      "\n",
      "Epoch 783: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2232e-04 - mae: 0.0068 - val_loss: 4.2434e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 784/3000\n",
      "\n",
      "Epoch 784: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2197e-04 - mae: 0.0068 - val_loss: 4.2491e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 785/3000\n",
      "\n",
      "Epoch 785: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2172e-04 - mae: 0.0067 - val_loss: 4.1872e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 786/3000\n",
      "\n",
      "Epoch 786: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2210e-04 - mae: 0.0068 - val_loss: 4.3409e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 787/3000\n",
      "\n",
      "Epoch 787: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2184e-04 - mae: 0.0067 - val_loss: 4.1978e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 788/3000\n",
      "\n",
      "Epoch 788: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2169e-04 - mae: 0.0067 - val_loss: 3.9850e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 789/3000\n",
      "\n",
      "Epoch 789: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2176e-04 - mae: 0.0067 - val_loss: 4.3628e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 790/3000\n",
      "\n",
      "Epoch 790: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2177e-04 - mae: 0.0067 - val_loss: 4.3001e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 791/3000\n",
      "\n",
      "Epoch 791: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2195e-04 - mae: 0.0068 - val_loss: 4.4129e-04 - val_mae: 0.0155 - 3s/epoch - 23ms/step\n",
      "Epoch 792/3000\n",
      "\n",
      "Epoch 792: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2218e-04 - mae: 0.0068 - val_loss: 4.2877e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 793/3000\n",
      "\n",
      "Epoch 793: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2155e-04 - mae: 0.0067 - val_loss: 4.2630e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 794/3000\n",
      "\n",
      "Epoch 794: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2178e-04 - mae: 0.0067 - val_loss: 4.3390e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 795/3000\n",
      "\n",
      "Epoch 795: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2198e-04 - mae: 0.0068 - val_loss: 4.1077e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 796/3000\n",
      "\n",
      "Epoch 796: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2174e-04 - mae: 0.0067 - val_loss: 4.2609e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 797/3000\n",
      "\n",
      "Epoch 797: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2156e-04 - mae: 0.0067 - val_loss: 4.0186e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 798/3000\n",
      "\n",
      "Epoch 798: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2182e-04 - mae: 0.0068 - val_loss: 4.0265e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 799/3000\n",
      "\n",
      "Epoch 799: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2193e-04 - mae: 0.0068 - val_loss: 3.9834e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 800/3000\n",
      "\n",
      "Epoch 800: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2186e-04 - mae: 0.0068 - val_loss: 4.4746e-04 - val_mae: 0.0156 - 3s/epoch - 21ms/step\n",
      "Epoch 801/3000\n",
      "\n",
      "Epoch 801: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2140e-04 - mae: 0.0067 - val_loss: 4.2246e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 802/3000\n",
      "\n",
      "Epoch 802: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2154e-04 - mae: 0.0067 - val_loss: 4.4067e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 803/3000\n",
      "\n",
      "Epoch 803: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2171e-04 - mae: 0.0067 - val_loss: 4.2107e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 804/3000\n",
      "\n",
      "Epoch 804: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2116e-04 - mae: 0.0067 - val_loss: 4.1475e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 805/3000\n",
      "\n",
      "Epoch 805: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2147e-04 - mae: 0.0067 - val_loss: 4.4017e-04 - val_mae: 0.0154 - 3s/epoch - 21ms/step\n",
      "Epoch 806/3000\n",
      "\n",
      "Epoch 806: val_mae did not improve from 0.01438\n",
      "151/151 - 3s - loss: 1.2152e-04 - mae: 0.0067 - val_loss: 4.1902e-04 - val_mae: 0.0150 - 3s/epoch - 21ms/step\n",
      "Epoch 807/3000\n",
      "\n",
      "Epoch 807: val_mae improved from 0.01438 to 0.01432, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 1.2135e-04 - mae: 0.0067 - val_loss: 3.7972e-04 - val_mae: 0.0143 - 3s/epoch - 22ms/step\n",
      "Epoch 808/3000\n",
      "\n",
      "Epoch 808: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2169e-04 - mae: 0.0067 - val_loss: 3.9453e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 809/3000\n",
      "\n",
      "Epoch 809: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2145e-04 - mae: 0.0067 - val_loss: 4.4776e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 810/3000\n",
      "\n",
      "Epoch 810: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2152e-04 - mae: 0.0067 - val_loss: 4.1418e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 811/3000\n",
      "\n",
      "Epoch 811: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2165e-04 - mae: 0.0067 - val_loss: 4.1183e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 812/3000\n",
      "\n",
      "Epoch 812: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2115e-04 - mae: 0.0067 - val_loss: 4.4140e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 813/3000\n",
      "\n",
      "Epoch 813: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2146e-04 - mae: 0.0067 - val_loss: 4.1220e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 814/3000\n",
      "\n",
      "Epoch 814: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2141e-04 - mae: 0.0067 - val_loss: 4.0215e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 815/3000\n",
      "\n",
      "Epoch 815: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2124e-04 - mae: 0.0067 - val_loss: 4.4176e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 816/3000\n",
      "\n",
      "Epoch 816: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2140e-04 - mae: 0.0067 - val_loss: 4.3260e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 817/3000\n",
      "\n",
      "Epoch 817: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2095e-04 - mae: 0.0067 - val_loss: 4.1938e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 818/3000\n",
      "\n",
      "Epoch 818: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2118e-04 - mae: 0.0067 - val_loss: 3.9761e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 819/3000\n",
      "\n",
      "Epoch 819: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2146e-04 - mae: 0.0067 - val_loss: 4.1225e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 820/3000\n",
      "\n",
      "Epoch 820: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2129e-04 - mae: 0.0067 - val_loss: 4.1732e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 821/3000\n",
      "\n",
      "Epoch 821: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2123e-04 - mae: 0.0067 - val_loss: 4.1808e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 822/3000\n",
      "\n",
      "Epoch 822: val_mae did not improve from 0.01432\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "151/151 - 4s - loss: 1.2093e-04 - mae: 0.0067 - val_loss: 4.2878e-04 - val_mae: 0.0151 - 4s/epoch - 24ms/step\n",
      "Epoch 823/3000\n",
      "\n",
      "Epoch 823: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2150e-04 - mae: 0.0067 - val_loss: 4.3700e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 824/3000\n",
      "\n",
      "Epoch 824: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2142e-04 - mae: 0.0067 - val_loss: 4.3605e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 825/3000\n",
      "\n",
      "Epoch 825: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2127e-04 - mae: 0.0067 - val_loss: 4.3393e-04 - val_mae: 0.0151 - 3s/epoch - 21ms/step\n",
      "Epoch 826/3000\n",
      "\n",
      "Epoch 826: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2111e-04 - mae: 0.0067 - val_loss: 4.2526e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 827/3000\n",
      "\n",
      "Epoch 827: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2123e-04 - mae: 0.0067 - val_loss: 4.1584e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 828/3000\n",
      "\n",
      "Epoch 828: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2093e-04 - mae: 0.0067 - val_loss: 4.2069e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 829/3000\n",
      "\n",
      "Epoch 829: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2108e-04 - mae: 0.0067 - val_loss: 4.3267e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 830/3000\n",
      "\n",
      "Epoch 830: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2102e-04 - mae: 0.0067 - val_loss: 4.3463e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 831/3000\n",
      "\n",
      "Epoch 831: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2135e-04 - mae: 0.0067 - val_loss: 3.8925e-04 - val_mae: 0.0145 - 3s/epoch - 23ms/step\n",
      "Epoch 832/3000\n",
      "\n",
      "Epoch 832: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2111e-04 - mae: 0.0067 - val_loss: 3.9086e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 833/3000\n",
      "\n",
      "Epoch 833: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2098e-04 - mae: 0.0067 - val_loss: 4.0254e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 834/3000\n",
      "\n",
      "Epoch 834: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2140e-04 - mae: 0.0067 - val_loss: 4.2044e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 835/3000\n",
      "\n",
      "Epoch 835: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2090e-04 - mae: 0.0067 - val_loss: 4.2546e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 836/3000\n",
      "\n",
      "Epoch 836: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2058e-04 - mae: 0.0067 - val_loss: 4.0432e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 837/3000\n",
      "\n",
      "Epoch 837: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2079e-04 - mae: 0.0067 - val_loss: 4.1712e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 838/3000\n",
      "\n",
      "Epoch 838: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2063e-04 - mae: 0.0067 - val_loss: 4.2694e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 839/3000\n",
      "\n",
      "Epoch 839: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2088e-04 - mae: 0.0067 - val_loss: 4.3080e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 840/3000\n",
      "\n",
      "Epoch 840: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2077e-04 - mae: 0.0067 - val_loss: 3.9809e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 841/3000\n",
      "\n",
      "Epoch 841: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2067e-04 - mae: 0.0067 - val_loss: 4.1028e-04 - val_mae: 0.0148 - 3s/epoch - 21ms/step\n",
      "Epoch 842/3000\n",
      "\n",
      "Epoch 842: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2089e-04 - mae: 0.0067 - val_loss: 4.0873e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 843/3000\n",
      "\n",
      "Epoch 843: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2058e-04 - mae: 0.0067 - val_loss: 4.1676e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 844/3000\n",
      "\n",
      "Epoch 844: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2115e-04 - mae: 0.0067 - val_loss: 4.3268e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 845/3000\n",
      "\n",
      "Epoch 845: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2088e-04 - mae: 0.0067 - val_loss: 3.9534e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 846/3000\n",
      "\n",
      "Epoch 846: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2077e-04 - mae: 0.0067 - val_loss: 4.2020e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 847/3000\n",
      "\n",
      "Epoch 847: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2051e-04 - mae: 0.0067 - val_loss: 4.1109e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 848/3000\n",
      "\n",
      "Epoch 848: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2059e-04 - mae: 0.0067 - val_loss: 4.1611e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 849/3000\n",
      "\n",
      "Epoch 849: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2086e-04 - mae: 0.0067 - val_loss: 4.1017e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 850/3000\n",
      "\n",
      "Epoch 850: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2048e-04 - mae: 0.0067 - val_loss: 4.3776e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 851/3000\n",
      "\n",
      "Epoch 851: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2050e-04 - mae: 0.0067 - val_loss: 4.1543e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 852/3000\n",
      "\n",
      "Epoch 852: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2036e-04 - mae: 0.0067 - val_loss: 4.0778e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 853/3000\n",
      "\n",
      "Epoch 853: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2043e-04 - mae: 0.0067 - val_loss: 4.0551e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 854/3000\n",
      "\n",
      "Epoch 854: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2049e-04 - mae: 0.0067 - val_loss: 4.0925e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 855/3000\n",
      "\n",
      "Epoch 855: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2039e-04 - mae: 0.0067 - val_loss: 4.2396e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 856/3000\n",
      "\n",
      "Epoch 856: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2025e-04 - mae: 0.0067 - val_loss: 4.1282e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 857/3000\n",
      "\n",
      "Epoch 857: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2039e-04 - mae: 0.0067 - val_loss: 4.2972e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 858/3000\n",
      "\n",
      "Epoch 858: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2044e-04 - mae: 0.0067 - val_loss: 4.3819e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 859/3000\n",
      "\n",
      "Epoch 859: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2064e-04 - mae: 0.0067 - val_loss: 4.2181e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 860/3000\n",
      "\n",
      "Epoch 860: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2058e-04 - mae: 0.0067 - val_loss: 4.1791e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 861/3000\n",
      "\n",
      "Epoch 861: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2018e-04 - mae: 0.0067 - val_loss: 4.1137e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 862/3000\n",
      "\n",
      "Epoch 862: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2017e-04 - mae: 0.0067 - val_loss: 4.2355e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 863/3000\n",
      "\n",
      "Epoch 863: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2037e-04 - mae: 0.0067 - val_loss: 4.1514e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 864/3000\n",
      "\n",
      "Epoch 864: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2037e-04 - mae: 0.0067 - val_loss: 3.8479e-04 - val_mae: 0.0144 - 3s/epoch - 22ms/step\n",
      "Epoch 865/3000\n",
      "\n",
      "Epoch 865: val_mae did not improve from 0.01432\n",
      "151/151 - 4s - loss: 1.2025e-04 - mae: 0.0067 - val_loss: 4.0486e-04 - val_mae: 0.0147 - 4s/epoch - 23ms/step\n",
      "Epoch 866/3000\n",
      "\n",
      "Epoch 866: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2028e-04 - mae: 0.0067 - val_loss: 4.4305e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 867/3000\n",
      "\n",
      "Epoch 867: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2033e-04 - mae: 0.0067 - val_loss: 4.1599e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 868/3000\n",
      "\n",
      "Epoch 868: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2029e-04 - mae: 0.0067 - val_loss: 4.0911e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 869/3000\n",
      "\n",
      "Epoch 869: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2024e-04 - mae: 0.0067 - val_loss: 4.2208e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 870/3000\n",
      "\n",
      "Epoch 870: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2032e-04 - mae: 0.0067 - val_loss: 4.0330e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 871/3000\n",
      "\n",
      "Epoch 871: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2046e-04 - mae: 0.0067 - val_loss: 4.0715e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 872/3000\n",
      "\n",
      "Epoch 872: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2025e-04 - mae: 0.0067 - val_loss: 4.3271e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 873/3000\n",
      "\n",
      "Epoch 873: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2003e-04 - mae: 0.0067 - val_loss: 4.2397e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 874/3000\n",
      "\n",
      "Epoch 874: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2019e-04 - mae: 0.0067 - val_loss: 4.2365e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 875/3000\n",
      "\n",
      "Epoch 875: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1996e-04 - mae: 0.0067 - val_loss: 4.1533e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 876/3000\n",
      "\n",
      "Epoch 876: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2018e-04 - mae: 0.0067 - val_loss: 4.4680e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 877/3000\n",
      "\n",
      "Epoch 877: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2001e-04 - mae: 0.0067 - val_loss: 4.0938e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 878/3000\n",
      "\n",
      "Epoch 878: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2016e-04 - mae: 0.0067 - val_loss: 4.2092e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 879/3000\n",
      "\n",
      "Epoch 879: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1979e-04 - mae: 0.0067 - val_loss: 4.6475e-04 - val_mae: 0.0159 - 3s/epoch - 22ms/step\n",
      "Epoch 880/3000\n",
      "\n",
      "Epoch 880: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2037e-04 - mae: 0.0067 - val_loss: 4.0939e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 881/3000\n",
      "\n",
      "Epoch 881: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2015e-04 - mae: 0.0067 - val_loss: 4.0740e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 882/3000\n",
      "\n",
      "Epoch 882: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2001e-04 - mae: 0.0067 - val_loss: 4.1945e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 883/3000\n",
      "\n",
      "Epoch 883: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1959e-04 - mae: 0.0067 - val_loss: 4.0685e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 884/3000\n",
      "\n",
      "Epoch 884: val_mae did not improve from 0.01432\n",
      "151/151 - 4s - loss: 1.2029e-04 - mae: 0.0067 - val_loss: 3.8892e-04 - val_mae: 0.0145 - 4s/epoch - 24ms/step\n",
      "Epoch 885/3000\n",
      "\n",
      "Epoch 885: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1999e-04 - mae: 0.0067 - val_loss: 4.2430e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 886/3000\n",
      "\n",
      "Epoch 886: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1972e-04 - mae: 0.0067 - val_loss: 4.1842e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 887/3000\n",
      "\n",
      "Epoch 887: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2012e-04 - mae: 0.0067 - val_loss: 3.9997e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 888/3000\n",
      "\n",
      "Epoch 888: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1994e-04 - mae: 0.0067 - val_loss: 4.2555e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 889/3000\n",
      "\n",
      "Epoch 889: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2007e-04 - mae: 0.0067 - val_loss: 4.1485e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 890/3000\n",
      "\n",
      "Epoch 890: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2000e-04 - mae: 0.0067 - val_loss: 4.0842e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 891/3000\n",
      "\n",
      "Epoch 891: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2005e-04 - mae: 0.0067 - val_loss: 4.2874e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 892/3000\n",
      "\n",
      "Epoch 892: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1970e-04 - mae: 0.0067 - val_loss: 3.9782e-04 - val_mae: 0.0146 - 3s/epoch - 23ms/step\n",
      "Epoch 893/3000\n",
      "\n",
      "Epoch 893: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1981e-04 - mae: 0.0067 - val_loss: 4.0784e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 894/3000\n",
      "\n",
      "Epoch 894: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1957e-04 - mae: 0.0067 - val_loss: 4.0004e-04 - val_mae: 0.0146 - 3s/epoch - 23ms/step\n",
      "Epoch 895/3000\n",
      "\n",
      "Epoch 895: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1967e-04 - mae: 0.0067 - val_loss: 4.2429e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 896/3000\n",
      "\n",
      "Epoch 896: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1958e-04 - mae: 0.0067 - val_loss: 4.2180e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 897/3000\n",
      "\n",
      "Epoch 897: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1981e-04 - mae: 0.0067 - val_loss: 4.0580e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 898/3000\n",
      "\n",
      "Epoch 898: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.2017e-04 - mae: 0.0067 - val_loss: 4.2081e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 899/3000\n",
      "\n",
      "Epoch 899: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1985e-04 - mae: 0.0067 - val_loss: 4.1708e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 900/3000\n",
      "\n",
      "Epoch 900: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1988e-04 - mae: 0.0067 - val_loss: 4.3950e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 901/3000\n",
      "\n",
      "Epoch 901: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1970e-04 - mae: 0.0067 - val_loss: 4.0938e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 902/3000\n",
      "\n",
      "Epoch 902: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1941e-04 - mae: 0.0067 - val_loss: 4.0097e-04 - val_mae: 0.0147 - 3s/epoch - 23ms/step\n",
      "Epoch 903/3000\n",
      "\n",
      "Epoch 903: val_mae did not improve from 0.01432\n",
      "151/151 - 4s - loss: 1.1954e-04 - mae: 0.0067 - val_loss: 3.9574e-04 - val_mae: 0.0146 - 4s/epoch - 24ms/step\n",
      "Epoch 904/3000\n",
      "\n",
      "Epoch 904: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1975e-04 - mae: 0.0067 - val_loss: 4.2170e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 905/3000\n",
      "\n",
      "Epoch 905: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1945e-04 - mae: 0.0067 - val_loss: 4.4459e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 906/3000\n",
      "\n",
      "Epoch 906: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1965e-04 - mae: 0.0067 - val_loss: 4.3362e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 907/3000\n",
      "\n",
      "Epoch 907: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1965e-04 - mae: 0.0067 - val_loss: 3.9467e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 908/3000\n",
      "\n",
      "Epoch 908: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1968e-04 - mae: 0.0067 - val_loss: 4.1556e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 909/3000\n",
      "\n",
      "Epoch 909: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1934e-04 - mae: 0.0067 - val_loss: 4.4212e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 910/3000\n",
      "\n",
      "Epoch 910: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1959e-04 - mae: 0.0067 - val_loss: 4.0835e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 911/3000\n",
      "\n",
      "Epoch 911: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1980e-04 - mae: 0.0067 - val_loss: 4.3070e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 912/3000\n",
      "\n",
      "Epoch 912: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1939e-04 - mae: 0.0067 - val_loss: 4.2856e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 913/3000\n",
      "\n",
      "Epoch 913: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1959e-04 - mae: 0.0067 - val_loss: 3.8789e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 914/3000\n",
      "\n",
      "Epoch 914: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1940e-04 - mae: 0.0067 - val_loss: 3.9634e-04 - val_mae: 0.0146 - 3s/epoch - 23ms/step\n",
      "Epoch 915/3000\n",
      "\n",
      "Epoch 915: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1935e-04 - mae: 0.0067 - val_loss: 4.0380e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 916/3000\n",
      "\n",
      "Epoch 916: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1924e-04 - mae: 0.0067 - val_loss: 4.1355e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 917/3000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 917: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1951e-04 - mae: 0.0067 - val_loss: 4.1183e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 918/3000\n",
      "\n",
      "Epoch 918: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1934e-04 - mae: 0.0067 - val_loss: 4.2485e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 919/3000\n",
      "\n",
      "Epoch 919: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1959e-04 - mae: 0.0067 - val_loss: 4.3212e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 920/3000\n",
      "\n",
      "Epoch 920: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1901e-04 - mae: 0.0066 - val_loss: 4.0732e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 921/3000\n",
      "\n",
      "Epoch 921: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1964e-04 - mae: 0.0067 - val_loss: 3.9848e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 922/3000\n",
      "\n",
      "Epoch 922: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1926e-04 - mae: 0.0067 - val_loss: 4.1964e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 923/3000\n",
      "\n",
      "Epoch 923: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1926e-04 - mae: 0.0067 - val_loss: 4.0454e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 924/3000\n",
      "\n",
      "Epoch 924: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1957e-04 - mae: 0.0067 - val_loss: 4.3252e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 925/3000\n",
      "\n",
      "Epoch 925: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1919e-04 - mae: 0.0067 - val_loss: 4.4348e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 926/3000\n",
      "\n",
      "Epoch 926: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1945e-04 - mae: 0.0067 - val_loss: 4.0605e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 927/3000\n",
      "\n",
      "Epoch 927: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1952e-04 - mae: 0.0067 - val_loss: 4.0970e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 928/3000\n",
      "\n",
      "Epoch 928: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1900e-04 - mae: 0.0066 - val_loss: 4.3585e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 929/3000\n",
      "\n",
      "Epoch 929: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1906e-04 - mae: 0.0066 - val_loss: 4.0473e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 930/3000\n",
      "\n",
      "Epoch 930: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1923e-04 - mae: 0.0067 - val_loss: 4.4007e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 931/3000\n",
      "\n",
      "Epoch 931: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1890e-04 - mae: 0.0066 - val_loss: 4.4226e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 932/3000\n",
      "\n",
      "Epoch 932: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1945e-04 - mae: 0.0067 - val_loss: 4.2397e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 933/3000\n",
      "\n",
      "Epoch 933: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1918e-04 - mae: 0.0067 - val_loss: 4.1670e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 934/3000\n",
      "\n",
      "Epoch 934: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1974e-04 - mae: 0.0067 - val_loss: 4.0678e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 935/3000\n",
      "\n",
      "Epoch 935: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1927e-04 - mae: 0.0067 - val_loss: 4.3593e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 936/3000\n",
      "\n",
      "Epoch 936: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1916e-04 - mae: 0.0067 - val_loss: 3.9098e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 937/3000\n",
      "\n",
      "Epoch 937: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1916e-04 - mae: 0.0067 - val_loss: 3.9879e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 938/3000\n",
      "\n",
      "Epoch 938: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1912e-04 - mae: 0.0066 - val_loss: 4.4864e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 939/3000\n",
      "\n",
      "Epoch 939: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1884e-04 - mae: 0.0066 - val_loss: 4.1525e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 940/3000\n",
      "\n",
      "Epoch 940: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1898e-04 - mae: 0.0066 - val_loss: 4.2147e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 941/3000\n",
      "\n",
      "Epoch 941: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1885e-04 - mae: 0.0066 - val_loss: 4.1234e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 942/3000\n",
      "\n",
      "Epoch 942: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1908e-04 - mae: 0.0067 - val_loss: 3.8045e-04 - val_mae: 0.0143 - 3s/epoch - 22ms/step\n",
      "Epoch 943/3000\n",
      "\n",
      "Epoch 943: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1932e-04 - mae: 0.0067 - val_loss: 3.9252e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 944/3000\n",
      "\n",
      "Epoch 944: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1899e-04 - mae: 0.0066 - val_loss: 4.0888e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 945/3000\n",
      "\n",
      "Epoch 945: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1887e-04 - mae: 0.0066 - val_loss: 3.9962e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 946/3000\n",
      "\n",
      "Epoch 946: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1896e-04 - mae: 0.0066 - val_loss: 4.0583e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 947/3000\n",
      "\n",
      "Epoch 947: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1930e-04 - mae: 0.0067 - val_loss: 4.3219e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 948/3000\n",
      "\n",
      "Epoch 948: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1925e-04 - mae: 0.0067 - val_loss: 4.0828e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 949/3000\n",
      "\n",
      "Epoch 949: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1895e-04 - mae: 0.0066 - val_loss: 4.2274e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 950/3000\n",
      "\n",
      "Epoch 950: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1877e-04 - mae: 0.0066 - val_loss: 4.1303e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 951/3000\n",
      "\n",
      "Epoch 951: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1887e-04 - mae: 0.0066 - val_loss: 4.0522e-04 - val_mae: 0.0147 - 3s/epoch - 23ms/step\n",
      "Epoch 952/3000\n",
      "\n",
      "Epoch 952: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1865e-04 - mae: 0.0066 - val_loss: 3.9335e-04 - val_mae: 0.0146 - 3s/epoch - 23ms/step\n",
      "Epoch 953/3000\n",
      "\n",
      "Epoch 953: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1876e-04 - mae: 0.0066 - val_loss: 3.9596e-04 - val_mae: 0.0146 - 3s/epoch - 23ms/step\n",
      "Epoch 954/3000\n",
      "\n",
      "Epoch 954: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1855e-04 - mae: 0.0066 - val_loss: 4.1594e-04 - val_mae: 0.0149 - 3s/epoch - 21ms/step\n",
      "Epoch 955/3000\n",
      "\n",
      "Epoch 955: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1870e-04 - mae: 0.0066 - val_loss: 4.0112e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 956/3000\n",
      "\n",
      "Epoch 956: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1908e-04 - mae: 0.0067 - val_loss: 4.0811e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 957/3000\n",
      "\n",
      "Epoch 957: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1860e-04 - mae: 0.0066 - val_loss: 4.2372e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 958/3000\n",
      "\n",
      "Epoch 958: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1887e-04 - mae: 0.0066 - val_loss: 4.1979e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 959/3000\n",
      "\n",
      "Epoch 959: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1870e-04 - mae: 0.0066 - val_loss: 4.1371e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 960/3000\n",
      "\n",
      "Epoch 960: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1886e-04 - mae: 0.0066 - val_loss: 4.1788e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 961/3000\n",
      "\n",
      "Epoch 961: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1902e-04 - mae: 0.0066 - val_loss: 4.1048e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 962/3000\n",
      "\n",
      "Epoch 962: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1881e-04 - mae: 0.0066 - val_loss: 4.0698e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 963/3000\n",
      "\n",
      "Epoch 963: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1839e-04 - mae: 0.0066 - val_loss: 4.3507e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 964/3000\n",
      "\n",
      "Epoch 964: val_mae did not improve from 0.01432\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "151/151 - 3s - loss: 1.1865e-04 - mae: 0.0066 - val_loss: 4.2058e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 965/3000\n",
      "\n",
      "Epoch 965: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1881e-04 - mae: 0.0066 - val_loss: 4.3000e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 966/3000\n",
      "\n",
      "Epoch 966: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1902e-04 - mae: 0.0067 - val_loss: 4.2839e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 967/3000\n",
      "\n",
      "Epoch 967: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1864e-04 - mae: 0.0066 - val_loss: 4.2315e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 968/3000\n",
      "\n",
      "Epoch 968: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1847e-04 - mae: 0.0066 - val_loss: 4.0465e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 969/3000\n",
      "\n",
      "Epoch 969: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1868e-04 - mae: 0.0066 - val_loss: 4.0911e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 970/3000\n",
      "\n",
      "Epoch 970: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1895e-04 - mae: 0.0067 - val_loss: 4.1453e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 971/3000\n",
      "\n",
      "Epoch 971: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1848e-04 - mae: 0.0066 - val_loss: 4.2431e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 972/3000\n",
      "\n",
      "Epoch 972: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1857e-04 - mae: 0.0066 - val_loss: 4.3167e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 973/3000\n",
      "\n",
      "Epoch 973: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1849e-04 - mae: 0.0066 - val_loss: 4.1426e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 974/3000\n",
      "\n",
      "Epoch 974: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1826e-04 - mae: 0.0066 - val_loss: 4.1296e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 975/3000\n",
      "\n",
      "Epoch 975: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1834e-04 - mae: 0.0066 - val_loss: 4.0011e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 976/3000\n",
      "\n",
      "Epoch 976: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1875e-04 - mae: 0.0066 - val_loss: 4.4764e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 977/3000\n",
      "\n",
      "Epoch 977: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1827e-04 - mae: 0.0066 - val_loss: 4.1444e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 978/3000\n",
      "\n",
      "Epoch 978: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1840e-04 - mae: 0.0066 - val_loss: 4.0735e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 979/3000\n",
      "\n",
      "Epoch 979: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1867e-04 - mae: 0.0066 - val_loss: 4.5419e-04 - val_mae: 0.0156 - 3s/epoch - 22ms/step\n",
      "Epoch 980/3000\n",
      "\n",
      "Epoch 980: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1881e-04 - mae: 0.0066 - val_loss: 4.1495e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 981/3000\n",
      "\n",
      "Epoch 981: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1839e-04 - mae: 0.0066 - val_loss: 4.0734e-04 - val_mae: 0.0147 - 3s/epoch - 23ms/step\n",
      "Epoch 982/3000\n",
      "\n",
      "Epoch 982: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1851e-04 - mae: 0.0066 - val_loss: 3.8849e-04 - val_mae: 0.0144 - 3s/epoch - 23ms/step\n",
      "Epoch 983/3000\n",
      "\n",
      "Epoch 983: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1824e-04 - mae: 0.0066 - val_loss: 4.1179e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 984/3000\n",
      "\n",
      "Epoch 984: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1832e-04 - mae: 0.0066 - val_loss: 4.3125e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 985/3000\n",
      "\n",
      "Epoch 985: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1817e-04 - mae: 0.0066 - val_loss: 4.3040e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 986/3000\n",
      "\n",
      "Epoch 986: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1821e-04 - mae: 0.0066 - val_loss: 4.0996e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 987/3000\n",
      "\n",
      "Epoch 987: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1840e-04 - mae: 0.0066 - val_loss: 4.2029e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 988/3000\n",
      "\n",
      "Epoch 988: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1867e-04 - mae: 0.0066 - val_loss: 4.3747e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 989/3000\n",
      "\n",
      "Epoch 989: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1845e-04 - mae: 0.0066 - val_loss: 4.1439e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 990/3000\n",
      "\n",
      "Epoch 990: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1849e-04 - mae: 0.0066 - val_loss: 4.1193e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 991/3000\n",
      "\n",
      "Epoch 991: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1837e-04 - mae: 0.0066 - val_loss: 4.0476e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 992/3000\n",
      "\n",
      "Epoch 992: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1832e-04 - mae: 0.0066 - val_loss: 4.1250e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 993/3000\n",
      "\n",
      "Epoch 993: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1837e-04 - mae: 0.0066 - val_loss: 3.9370e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 994/3000\n",
      "\n",
      "Epoch 994: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1830e-04 - mae: 0.0066 - val_loss: 4.0060e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 995/3000\n",
      "\n",
      "Epoch 995: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1786e-04 - mae: 0.0066 - val_loss: 4.1413e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 996/3000\n",
      "\n",
      "Epoch 996: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1806e-04 - mae: 0.0066 - val_loss: 3.9725e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 997/3000\n",
      "\n",
      "Epoch 997: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1829e-04 - mae: 0.0066 - val_loss: 4.0838e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 998/3000\n",
      "\n",
      "Epoch 998: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1801e-04 - mae: 0.0066 - val_loss: 4.2959e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 999/3000\n",
      "\n",
      "Epoch 999: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1801e-04 - mae: 0.0066 - val_loss: 4.1436e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1000/3000\n",
      "\n",
      "Epoch 1000: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1810e-04 - mae: 0.0066 - val_loss: 4.3338e-04 - val_mae: 0.0152 - 3s/epoch - 22ms/step\n",
      "Epoch 1001/3000\n",
      "\n",
      "Epoch 1001: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1822e-04 - mae: 0.0066 - val_loss: 4.3401e-04 - val_mae: 0.0153 - 3s/epoch - 22ms/step\n",
      "Epoch 1002/3000\n",
      "\n",
      "Epoch 1002: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1803e-04 - mae: 0.0066 - val_loss: 4.1884e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1003/3000\n",
      "\n",
      "Epoch 1003: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1812e-04 - mae: 0.0066 - val_loss: 4.0698e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1004/3000\n",
      "\n",
      "Epoch 1004: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1827e-04 - mae: 0.0066 - val_loss: 4.1737e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1005/3000\n",
      "\n",
      "Epoch 1005: val_mae did not improve from 0.01432\n",
      "151/151 - 4s - loss: 1.1828e-04 - mae: 0.0066 - val_loss: 3.9725e-04 - val_mae: 0.0146 - 4s/epoch - 23ms/step\n",
      "Epoch 1006/3000\n",
      "\n",
      "Epoch 1006: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1788e-04 - mae: 0.0066 - val_loss: 4.5384e-04 - val_mae: 0.0155 - 3s/epoch - 22ms/step\n",
      "Epoch 1007/3000\n",
      "\n",
      "Epoch 1007: val_mae did not improve from 0.01432\n",
      "151/151 - 4s - loss: 1.1802e-04 - mae: 0.0066 - val_loss: 4.1709e-04 - val_mae: 0.0149 - 4s/epoch - 23ms/step\n",
      "Epoch 1008/3000\n",
      "\n",
      "Epoch 1008: val_mae did not improve from 0.01432\n",
      "151/151 - 4s - loss: 1.1772e-04 - mae: 0.0066 - val_loss: 3.9505e-04 - val_mae: 0.0146 - 4s/epoch - 23ms/step\n",
      "Epoch 1009/3000\n",
      "\n",
      "Epoch 1009: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1812e-04 - mae: 0.0066 - val_loss: 4.4169e-04 - val_mae: 0.0154 - 3s/epoch - 23ms/step\n",
      "Epoch 1010/3000\n",
      "\n",
      "Epoch 1010: val_mae did not improve from 0.01432\n",
      "151/151 - 4s - loss: 1.1793e-04 - mae: 0.0066 - val_loss: 3.9558e-04 - val_mae: 0.0146 - 4s/epoch - 23ms/step\n",
      "Epoch 1011/3000\n",
      "\n",
      "Epoch 1011: val_mae did not improve from 0.01432\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "151/151 - 3s - loss: 1.1778e-04 - mae: 0.0066 - val_loss: 4.0546e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 1012/3000\n",
      "\n",
      "Epoch 1012: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1810e-04 - mae: 0.0066 - val_loss: 4.1357e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1013/3000\n",
      "\n",
      "Epoch 1013: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1783e-04 - mae: 0.0066 - val_loss: 4.2478e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1014/3000\n",
      "\n",
      "Epoch 1014: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1795e-04 - mae: 0.0066 - val_loss: 4.1182e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1015/3000\n",
      "\n",
      "Epoch 1015: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1787e-04 - mae: 0.0066 - val_loss: 3.9580e-04 - val_mae: 0.0145 - 3s/epoch - 23ms/step\n",
      "Epoch 1016/3000\n",
      "\n",
      "Epoch 1016: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1799e-04 - mae: 0.0066 - val_loss: 4.1437e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1017/3000\n",
      "\n",
      "Epoch 1017: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1762e-04 - mae: 0.0066 - val_loss: 3.9323e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 1018/3000\n",
      "\n",
      "Epoch 1018: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1807e-04 - mae: 0.0066 - val_loss: 4.1398e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1019/3000\n",
      "\n",
      "Epoch 1019: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1771e-04 - mae: 0.0066 - val_loss: 4.0528e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 1020/3000\n",
      "\n",
      "Epoch 1020: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1790e-04 - mae: 0.0066 - val_loss: 3.8677e-04 - val_mae: 0.0144 - 3s/epoch - 22ms/step\n",
      "Epoch 1021/3000\n",
      "\n",
      "Epoch 1021: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1815e-04 - mae: 0.0066 - val_loss: 3.9748e-04 - val_mae: 0.0146 - 3s/epoch - 23ms/step\n",
      "Epoch 1022/3000\n",
      "\n",
      "Epoch 1022: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1809e-04 - mae: 0.0066 - val_loss: 4.2064e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1023/3000\n",
      "\n",
      "Epoch 1023: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1765e-04 - mae: 0.0066 - val_loss: 3.8558e-04 - val_mae: 0.0144 - 3s/epoch - 23ms/step\n",
      "Epoch 1024/3000\n",
      "\n",
      "Epoch 1024: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1776e-04 - mae: 0.0066 - val_loss: 4.1233e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1025/3000\n",
      "\n",
      "Epoch 1025: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1763e-04 - mae: 0.0066 - val_loss: 4.0768e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1026/3000\n",
      "\n",
      "Epoch 1026: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1776e-04 - mae: 0.0066 - val_loss: 4.1437e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1027/3000\n",
      "\n",
      "Epoch 1027: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1779e-04 - mae: 0.0066 - val_loss: 4.0268e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 1028/3000\n",
      "\n",
      "Epoch 1028: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1763e-04 - mae: 0.0066 - val_loss: 4.0132e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 1029/3000\n",
      "\n",
      "Epoch 1029: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1758e-04 - mae: 0.0066 - val_loss: 4.0179e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 1030/3000\n",
      "\n",
      "Epoch 1030: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1759e-04 - mae: 0.0066 - val_loss: 4.1801e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1031/3000\n",
      "\n",
      "Epoch 1031: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1808e-04 - mae: 0.0066 - val_loss: 4.0599e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1032/3000\n",
      "\n",
      "Epoch 1032: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1781e-04 - mae: 0.0066 - val_loss: 4.1706e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1033/3000\n",
      "\n",
      "Epoch 1033: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1758e-04 - mae: 0.0066 - val_loss: 4.0909e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1034/3000\n",
      "\n",
      "Epoch 1034: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1773e-04 - mae: 0.0066 - val_loss: 4.0690e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1035/3000\n",
      "\n",
      "Epoch 1035: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1783e-04 - mae: 0.0066 - val_loss: 4.1856e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1036/3000\n",
      "\n",
      "Epoch 1036: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1790e-04 - mae: 0.0066 - val_loss: 4.1382e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1037/3000\n",
      "\n",
      "Epoch 1037: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1752e-04 - mae: 0.0066 - val_loss: 4.2628e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1038/3000\n",
      "\n",
      "Epoch 1038: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1743e-04 - mae: 0.0066 - val_loss: 4.3839e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 1039/3000\n",
      "\n",
      "Epoch 1039: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1749e-04 - mae: 0.0066 - val_loss: 4.0206e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 1040/3000\n",
      "\n",
      "Epoch 1040: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1750e-04 - mae: 0.0066 - val_loss: 4.1778e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1041/3000\n",
      "\n",
      "Epoch 1041: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1742e-04 - mae: 0.0066 - val_loss: 4.3987e-04 - val_mae: 0.0154 - 3s/epoch - 22ms/step\n",
      "Epoch 1042/3000\n",
      "\n",
      "Epoch 1042: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1745e-04 - mae: 0.0066 - val_loss: 4.0646e-04 - val_mae: 0.0147 - 3s/epoch - 23ms/step\n",
      "Epoch 1043/3000\n",
      "\n",
      "Epoch 1043: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1740e-04 - mae: 0.0066 - val_loss: 3.9405e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 1044/3000\n",
      "\n",
      "Epoch 1044: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1820e-04 - mae: 0.0066 - val_loss: 4.2379e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 1045/3000\n",
      "\n",
      "Epoch 1045: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1757e-04 - mae: 0.0066 - val_loss: 4.1326e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1046/3000\n",
      "\n",
      "Epoch 1046: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1760e-04 - mae: 0.0066 - val_loss: 4.2424e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 1047/3000\n",
      "\n",
      "Epoch 1047: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1741e-04 - mae: 0.0066 - val_loss: 4.1571e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1048/3000\n",
      "\n",
      "Epoch 1048: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1761e-04 - mae: 0.0066 - val_loss: 4.0817e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1049/3000\n",
      "\n",
      "Epoch 1049: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1735e-04 - mae: 0.0066 - val_loss: 4.0171e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 1050/3000\n",
      "\n",
      "Epoch 1050: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1748e-04 - mae: 0.0066 - val_loss: 4.1582e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 1051/3000\n",
      "\n",
      "Epoch 1051: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1743e-04 - mae: 0.0066 - val_loss: 4.0368e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 1052/3000\n",
      "\n",
      "Epoch 1052: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1745e-04 - mae: 0.0066 - val_loss: 4.1314e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1053/3000\n",
      "\n",
      "Epoch 1053: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1725e-04 - mae: 0.0066 - val_loss: 4.0515e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1054/3000\n",
      "\n",
      "Epoch 1054: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1729e-04 - mae: 0.0066 - val_loss: 4.1839e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 1055/3000\n",
      "\n",
      "Epoch 1055: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1728e-04 - mae: 0.0066 - val_loss: 4.1995e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 1056/3000\n",
      "\n",
      "Epoch 1056: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1723e-04 - mae: 0.0066 - val_loss: 4.3843e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n",
      "Epoch 1057/3000\n",
      "\n",
      "Epoch 1057: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1734e-04 - mae: 0.0066 - val_loss: 4.4050e-04 - val_mae: 0.0153 - 3s/epoch - 23ms/step\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1058/3000\n",
      "\n",
      "Epoch 1058: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1739e-04 - mae: 0.0066 - val_loss: 4.0176e-04 - val_mae: 0.0147 - 3s/epoch - 23ms/step\n",
      "Epoch 1059/3000\n",
      "\n",
      "Epoch 1059: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1723e-04 - mae: 0.0066 - val_loss: 4.1149e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1060/3000\n",
      "\n",
      "Epoch 1060: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1739e-04 - mae: 0.0066 - val_loss: 3.9260e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 1061/3000\n",
      "\n",
      "Epoch 1061: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1737e-04 - mae: 0.0066 - val_loss: 3.8833e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 1062/3000\n",
      "\n",
      "Epoch 1062: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1746e-04 - mae: 0.0066 - val_loss: 3.8496e-04 - val_mae: 0.0143 - 3s/epoch - 23ms/step\n",
      "Epoch 1063/3000\n",
      "\n",
      "Epoch 1063: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1722e-04 - mae: 0.0066 - val_loss: 4.0457e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1064/3000\n",
      "\n",
      "Epoch 1064: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1704e-04 - mae: 0.0066 - val_loss: 3.9294e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 1065/3000\n",
      "\n",
      "Epoch 1065: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1733e-04 - mae: 0.0066 - val_loss: 4.1841e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1066/3000\n",
      "\n",
      "Epoch 1066: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1728e-04 - mae: 0.0066 - val_loss: 4.2947e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 1067/3000\n",
      "\n",
      "Epoch 1067: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1740e-04 - mae: 0.0066 - val_loss: 3.9102e-04 - val_mae: 0.0145 - 3s/epoch - 23ms/step\n",
      "Epoch 1068/3000\n",
      "\n",
      "Epoch 1068: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1739e-04 - mae: 0.0066 - val_loss: 4.0162e-04 - val_mae: 0.0147 - 3s/epoch - 23ms/step\n",
      "Epoch 1069/3000\n",
      "\n",
      "Epoch 1069: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1698e-04 - mae: 0.0066 - val_loss: 4.0052e-04 - val_mae: 0.0147 - 3s/epoch - 23ms/step\n",
      "Epoch 1070/3000\n",
      "\n",
      "Epoch 1070: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1721e-04 - mae: 0.0066 - val_loss: 4.1100e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 1071/3000\n",
      "\n",
      "Epoch 1071: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1709e-04 - mae: 0.0066 - val_loss: 4.0611e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 1072/3000\n",
      "\n",
      "Epoch 1072: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1710e-04 - mae: 0.0066 - val_loss: 4.0577e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 1073/3000\n",
      "\n",
      "Epoch 1073: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1719e-04 - mae: 0.0066 - val_loss: 4.0671e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1074/3000\n",
      "\n",
      "Epoch 1074: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1715e-04 - mae: 0.0066 - val_loss: 4.2055e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1075/3000\n",
      "\n",
      "Epoch 1075: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1707e-04 - mae: 0.0066 - val_loss: 4.2459e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1076/3000\n",
      "\n",
      "Epoch 1076: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1690e-04 - mae: 0.0066 - val_loss: 4.1024e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1077/3000\n",
      "\n",
      "Epoch 1077: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1704e-04 - mae: 0.0066 - val_loss: 3.9637e-04 - val_mae: 0.0146 - 3s/epoch - 23ms/step\n",
      "Epoch 1078/3000\n",
      "\n",
      "Epoch 1078: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1736e-04 - mae: 0.0066 - val_loss: 4.2354e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1079/3000\n",
      "\n",
      "Epoch 1079: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1688e-04 - mae: 0.0066 - val_loss: 3.8566e-04 - val_mae: 0.0143 - 3s/epoch - 22ms/step\n",
      "Epoch 1080/3000\n",
      "\n",
      "Epoch 1080: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1724e-04 - mae: 0.0066 - val_loss: 4.2652e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 1081/3000\n",
      "\n",
      "Epoch 1081: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1706e-04 - mae: 0.0066 - val_loss: 4.0053e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 1082/3000\n",
      "\n",
      "Epoch 1082: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1694e-04 - mae: 0.0066 - val_loss: 4.3065e-04 - val_mae: 0.0151 - 3s/epoch - 22ms/step\n",
      "Epoch 1083/3000\n",
      "\n",
      "Epoch 1083: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1694e-04 - mae: 0.0066 - val_loss: 4.1312e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1084/3000\n",
      "\n",
      "Epoch 1084: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1717e-04 - mae: 0.0066 - val_loss: 4.0764e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1085/3000\n",
      "\n",
      "Epoch 1085: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1729e-04 - mae: 0.0066 - val_loss: 3.9326e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 1086/3000\n",
      "\n",
      "Epoch 1086: val_mae did not improve from 0.01432\n",
      "151/151 - 4s - loss: 1.1707e-04 - mae: 0.0066 - val_loss: 4.0672e-04 - val_mae: 0.0147 - 4s/epoch - 23ms/step\n",
      "Epoch 1087/3000\n",
      "\n",
      "Epoch 1087: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1678e-04 - mae: 0.0066 - val_loss: 3.9901e-04 - val_mae: 0.0146 - 3s/epoch - 23ms/step\n",
      "Epoch 1088/3000\n",
      "\n",
      "Epoch 1088: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1706e-04 - mae: 0.0066 - val_loss: 3.9360e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 1089/3000\n",
      "\n",
      "Epoch 1089: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1702e-04 - mae: 0.0066 - val_loss: 4.0660e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 1090/3000\n",
      "\n",
      "Epoch 1090: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1672e-04 - mae: 0.0066 - val_loss: 4.0450e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1091/3000\n",
      "\n",
      "Epoch 1091: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1700e-04 - mae: 0.0066 - val_loss: 3.9676e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 1092/3000\n",
      "\n",
      "Epoch 1092: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1678e-04 - mae: 0.0066 - val_loss: 3.9838e-04 - val_mae: 0.0146 - 3s/epoch - 21ms/step\n",
      "Epoch 1093/3000\n",
      "\n",
      "Epoch 1093: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1681e-04 - mae: 0.0066 - val_loss: 3.8858e-04 - val_mae: 0.0145 - 3s/epoch - 23ms/step\n",
      "Epoch 1094/3000\n",
      "\n",
      "Epoch 1094: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1678e-04 - mae: 0.0066 - val_loss: 4.2612e-04 - val_mae: 0.0151 - 3s/epoch - 23ms/step\n",
      "Epoch 1095/3000\n",
      "\n",
      "Epoch 1095: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1684e-04 - mae: 0.0066 - val_loss: 4.1892e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1096/3000\n",
      "\n",
      "Epoch 1096: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1738e-04 - mae: 0.0066 - val_loss: 3.8518e-04 - val_mae: 0.0144 - 3s/epoch - 22ms/step\n",
      "Epoch 1097/3000\n",
      "\n",
      "Epoch 1097: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1739e-04 - mae: 0.0066 - val_loss: 4.1377e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1098/3000\n",
      "\n",
      "Epoch 1098: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1660e-04 - mae: 0.0065 - val_loss: 3.9241e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 1099/3000\n",
      "\n",
      "Epoch 1099: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1654e-04 - mae: 0.0065 - val_loss: 4.1168e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1100/3000\n",
      "\n",
      "Epoch 1100: val_mae did not improve from 0.01432\n",
      "151/151 - 3s - loss: 1.1700e-04 - mae: 0.0066 - val_loss: 4.2047e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1101/3000\n",
      "\n",
      "Epoch 1101: val_mae improved from 0.01432 to 0.01431, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 1.1663e-04 - mae: 0.0065 - val_loss: 3.8243e-04 - val_mae: 0.0143 - 3s/epoch - 23ms/step\n",
      "Epoch 1102/3000\n",
      "\n",
      "Epoch 1102: val_mae did not improve from 0.01431\n",
      "151/151 - 4s - loss: 1.1671e-04 - mae: 0.0066 - val_loss: 4.0903e-04 - val_mae: 0.0148 - 4s/epoch - 23ms/step\n",
      "Epoch 1103/3000\n",
      "\n",
      "Epoch 1103: val_mae did not improve from 0.01431\n",
      "151/151 - 4s - loss: 1.1692e-04 - mae: 0.0066 - val_loss: 3.9739e-04 - val_mae: 0.0145 - 4s/epoch - 23ms/step\n",
      "Epoch 1104/3000\n",
      "\n",
      "Epoch 1104: val_mae did not improve from 0.01431\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "151/151 - 3s - loss: 1.1689e-04 - mae: 0.0066 - val_loss: 4.2337e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1105/3000\n",
      "\n",
      "Epoch 1105: val_mae did not improve from 0.01431\n",
      "151/151 - 3s - loss: 1.1675e-04 - mae: 0.0066 - val_loss: 4.2194e-04 - val_mae: 0.0150 - 3s/epoch - 23ms/step\n",
      "Epoch 1106/3000\n",
      "\n",
      "Epoch 1106: val_mae did not improve from 0.01431\n",
      "151/151 - 3s - loss: 1.1656e-04 - mae: 0.0065 - val_loss: 4.1970e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1107/3000\n",
      "\n",
      "Epoch 1107: val_mae did not improve from 0.01431\n",
      "151/151 - 3s - loss: 1.1699e-04 - mae: 0.0066 - val_loss: 4.2173e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1108/3000\n",
      "\n",
      "Epoch 1108: val_mae did not improve from 0.01431\n",
      "151/151 - 3s - loss: 1.1681e-04 - mae: 0.0066 - val_loss: 4.1501e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1109/3000\n",
      "\n",
      "Epoch 1109: val_mae did not improve from 0.01431\n",
      "151/151 - 3s - loss: 1.1677e-04 - mae: 0.0066 - val_loss: 4.0145e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 1110/3000\n",
      "\n",
      "Epoch 1110: val_mae did not improve from 0.01431\n",
      "151/151 - 3s - loss: 1.1652e-04 - mae: 0.0065 - val_loss: 3.8568e-04 - val_mae: 0.0144 - 3s/epoch - 22ms/step\n",
      "Epoch 1111/3000\n",
      "\n",
      "Epoch 1111: val_mae improved from 0.01431 to 0.01429, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 1.1661e-04 - mae: 0.0066 - val_loss: 3.8285e-04 - val_mae: 0.0143 - 3s/epoch - 22ms/step\n",
      "Epoch 1112/3000\n",
      "\n",
      "Epoch 1112: val_mae did not improve from 0.01429\n",
      "151/151 - 3s - loss: 1.1672e-04 - mae: 0.0066 - val_loss: 4.0242e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 1113/3000\n",
      "\n",
      "Epoch 1113: val_mae did not improve from 0.01429\n",
      "151/151 - 3s - loss: 1.1647e-04 - mae: 0.0065 - val_loss: 4.0798e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1114/3000\n",
      "\n",
      "Epoch 1114: val_mae did not improve from 0.01429\n",
      "151/151 - 3s - loss: 1.1655e-04 - mae: 0.0065 - val_loss: 4.0203e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 1115/3000\n",
      "\n",
      "Epoch 1115: val_mae did not improve from 0.01429\n",
      "151/151 - 3s - loss: 1.1645e-04 - mae: 0.0065 - val_loss: 3.8541e-04 - val_mae: 0.0144 - 3s/epoch - 22ms/step\n",
      "Epoch 1116/3000\n",
      "\n",
      "Epoch 1116: val_mae did not improve from 0.01429\n",
      "151/151 - 3s - loss: 1.1647e-04 - mae: 0.0065 - val_loss: 4.1242e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1117/3000\n",
      "\n",
      "Epoch 1117: val_mae did not improve from 0.01429\n",
      "151/151 - 3s - loss: 1.1684e-04 - mae: 0.0066 - val_loss: 4.0922e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 1118/3000\n",
      "\n",
      "Epoch 1118: val_mae did not improve from 0.01429\n",
      "151/151 - 4s - loss: 1.1652e-04 - mae: 0.0065 - val_loss: 4.1997e-04 - val_mae: 0.0150 - 4s/epoch - 24ms/step\n",
      "Epoch 1119/3000\n",
      "\n",
      "Epoch 1119: val_mae did not improve from 0.01429\n",
      "151/151 - 4s - loss: 1.1650e-04 - mae: 0.0065 - val_loss: 4.0657e-04 - val_mae: 0.0147 - 4s/epoch - 23ms/step\n",
      "Epoch 1120/3000\n",
      "\n",
      "Epoch 1120: val_mae did not improve from 0.01429\n",
      "151/151 - 4s - loss: 1.1641e-04 - mae: 0.0065 - val_loss: 3.9421e-04 - val_mae: 0.0146 - 4s/epoch - 24ms/step\n",
      "Epoch 1121/3000\n",
      "\n",
      "Epoch 1121: val_mae did not improve from 0.01429\n",
      "151/151 - 4s - loss: 1.1663e-04 - mae: 0.0066 - val_loss: 3.9860e-04 - val_mae: 0.0146 - 4s/epoch - 23ms/step\n",
      "Epoch 1122/3000\n",
      "\n",
      "Epoch 1122: val_mae did not improve from 0.01429\n",
      "151/151 - 3s - loss: 1.1661e-04 - mae: 0.0066 - val_loss: 3.9374e-04 - val_mae: 0.0145 - 3s/epoch - 23ms/step\n",
      "Epoch 1123/3000\n",
      "\n",
      "Epoch 1123: val_mae did not improve from 0.01429\n",
      "151/151 - 3s - loss: 1.1689e-04 - mae: 0.0066 - val_loss: 3.9318e-04 - val_mae: 0.0145 - 3s/epoch - 23ms/step\n",
      "Epoch 1124/3000\n",
      "\n",
      "Epoch 1124: val_mae did not improve from 0.01429\n",
      "151/151 - 3s - loss: 1.1663e-04 - mae: 0.0066 - val_loss: 4.0960e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1125/3000\n",
      "\n",
      "Epoch 1125: val_mae improved from 0.01429 to 0.01425, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 1.1667e-04 - mae: 0.0066 - val_loss: 3.7730e-04 - val_mae: 0.0142 - 3s/epoch - 23ms/step\n",
      "Epoch 1126/3000\n",
      "\n",
      "Epoch 1126: val_mae did not improve from 0.01425\n",
      "151/151 - 3s - loss: 1.1637e-04 - mae: 0.0065 - val_loss: 4.1562e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 1127/3000\n",
      "\n",
      "Epoch 1127: val_mae did not improve from 0.01425\n",
      "151/151 - 3s - loss: 1.1645e-04 - mae: 0.0065 - val_loss: 4.0694e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 1128/3000\n",
      "\n",
      "Epoch 1128: val_mae did not improve from 0.01425\n",
      "151/151 - 4s - loss: 1.1638e-04 - mae: 0.0065 - val_loss: 4.2176e-04 - val_mae: 0.0150 - 4s/epoch - 24ms/step\n",
      "Epoch 1129/3000\n",
      "\n",
      "Epoch 1129: val_mae did not improve from 0.01425\n",
      "151/151 - 3s - loss: 1.1650e-04 - mae: 0.0065 - val_loss: 4.0740e-04 - val_mae: 0.0148 - 3s/epoch - 22ms/step\n",
      "Epoch 1130/3000\n",
      "\n",
      "Epoch 1130: val_mae did not improve from 0.01425\n",
      "151/151 - 3s - loss: 1.1645e-04 - mae: 0.0065 - val_loss: 4.1701e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1131/3000\n",
      "\n",
      "Epoch 1131: val_mae did not improve from 0.01425\n",
      "151/151 - 3s - loss: 1.1633e-04 - mae: 0.0065 - val_loss: 4.1417e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1132/3000\n",
      "\n",
      "Epoch 1132: val_mae did not improve from 0.01425\n",
      "151/151 - 3s - loss: 1.1674e-04 - mae: 0.0066 - val_loss: 4.1418e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1133/3000\n",
      "\n",
      "Epoch 1133: val_mae did not improve from 0.01425\n",
      "151/151 - 3s - loss: 1.1641e-04 - mae: 0.0065 - val_loss: 4.1488e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1134/3000\n",
      "\n",
      "Epoch 1134: val_mae did not improve from 0.01425\n",
      "151/151 - 3s - loss: 1.1630e-04 - mae: 0.0065 - val_loss: 4.1563e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1135/3000\n",
      "\n",
      "Epoch 1135: val_mae did not improve from 0.01425\n",
      "151/151 - 3s - loss: 1.1635e-04 - mae: 0.0065 - val_loss: 4.0285e-04 - val_mae: 0.0147 - 3s/epoch - 22ms/step\n",
      "Epoch 1136/3000\n",
      "\n",
      "Epoch 1136: val_mae did not improve from 0.01425\n",
      "151/151 - 3s - loss: 1.1626e-04 - mae: 0.0065 - val_loss: 3.9397e-04 - val_mae: 0.0145 - 3s/epoch - 22ms/step\n",
      "Epoch 1137/3000\n",
      "\n",
      "Epoch 1137: val_mae improved from 0.01425 to 0.01408, saving model to saved_models/gcn_gate_pump_nofuture.h5\n",
      "151/151 - 3s - loss: 1.1631e-04 - mae: 0.0065 - val_loss: 3.6792e-04 - val_mae: 0.0141 - 3s/epoch - 23ms/step\n",
      "Epoch 1138/3000\n",
      "\n",
      "Epoch 1138: val_mae did not improve from 0.01408\n",
      "151/151 - 3s - loss: 1.1646e-04 - mae: 0.0065 - val_loss: 3.9352e-04 - val_mae: 0.0145 - 3s/epoch - 23ms/step\n",
      "Epoch 1139/3000\n",
      "\n",
      "Epoch 1139: val_mae did not improve from 0.01408\n",
      "151/151 - 3s - loss: 1.1642e-04 - mae: 0.0065 - val_loss: 3.9858e-04 - val_mae: 0.0146 - 3s/epoch - 23ms/step\n",
      "Epoch 1140/3000\n",
      "\n",
      "Epoch 1140: val_mae did not improve from 0.01408\n",
      "151/151 - 3s - loss: 1.1621e-04 - mae: 0.0065 - val_loss: 4.1604e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1141/3000\n",
      "\n",
      "Epoch 1141: val_mae did not improve from 0.01408\n",
      "151/151 - 3s - loss: 1.1646e-04 - mae: 0.0065 - val_loss: 4.1832e-04 - val_mae: 0.0150 - 3s/epoch - 22ms/step\n",
      "Epoch 1142/3000\n",
      "\n",
      "Epoch 1142: val_mae did not improve from 0.01408\n",
      "151/151 - 3s - loss: 1.1652e-04 - mae: 0.0066 - val_loss: 4.1326e-04 - val_mae: 0.0148 - 3s/epoch - 23ms/step\n",
      "Epoch 1143/3000\n",
      "\n",
      "Epoch 1143: val_mae did not improve from 0.01408\n",
      "151/151 - 3s - loss: 1.1615e-04 - mae: 0.0065 - val_loss: 4.0420e-04 - val_mae: 0.0147 - 3s/epoch - 23ms/step\n",
      "Epoch 1144/3000\n",
      "\n",
      "Epoch 1144: val_mae did not improve from 0.01408\n",
      "151/151 - 3s - loss: 1.1628e-04 - mae: 0.0065 - val_loss: 4.1605e-04 - val_mae: 0.0149 - 3s/epoch - 23ms/step\n",
      "Epoch 1145/3000\n",
      "\n",
      "Epoch 1145: val_mae did not improve from 0.01408\n",
      "151/151 - 3s - loss: 1.1615e-04 - mae: 0.0065 - val_loss: 3.8736e-04 - val_mae: 0.0144 - 3s/epoch - 22ms/step\n",
      "Epoch 1146/3000\n",
      "\n",
      "Epoch 1146: val_mae did not improve from 0.01408\n",
      "151/151 - 3s - loss: 1.1689e-04 - mae: 0.0066 - val_loss: 3.9661e-04 - val_mae: 0.0146 - 3s/epoch - 22ms/step\n",
      "Epoch 1147/3000\n",
      "\n",
      "Epoch 1147: val_mae did not improve from 0.01408\n",
      "151/151 - 3s - loss: 1.1613e-04 - mae: 0.0065 - val_loss: 4.3148e-04 - val_mae: 0.0152 - 3s/epoch - 23ms/step\n",
      "Epoch 1148/3000\n",
      "\n",
      "Epoch 1148: val_mae did not improve from 0.01408\n",
      "151/151 - 3s - loss: 1.1614e-04 - mae: 0.0065 - val_loss: 4.1672e-04 - val_mae: 0.0149 - 3s/epoch - 22ms/step\n",
      "Epoch 1149/3000\n",
      "\n",
      "Epoch 1149: val_mae did not improve from 0.01408\n",
      "151/151 - 3s - loss: 1.1602e-04 - mae: 0.0065 - val_loss: 4.0561e-04 - val_mae: 0.0147 - 3s/epoch - 23ms/step\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1150/3000\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[13], line 21\u001b[0m\n\u001b[1;32m     11\u001b[0m es \u001b[38;5;241m=\u001b[39m EarlyStopping(monitor\u001b[38;5;241m=\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mval_mae\u001b[39m\u001b[38;5;124m'\u001b[39m, mode\u001b[38;5;241m=\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mmin\u001b[39m\u001b[38;5;124m'\u001b[39m, verbose\u001b[38;5;241m=\u001b[39m\u001b[38;5;241m2\u001b[39m, patience\u001b[38;5;241m=\u001b[39mPATIENCE)\n\u001b[1;32m     12\u001b[0m mc \u001b[38;5;241m=\u001b[39m ModelCheckpoint(\u001b[38;5;124m'\u001b[39m\u001b[38;5;124msaved_models/gcn_gate_pump_nofuture.h5\u001b[39m\u001b[38;5;124m'\u001b[39m, \n\u001b[1;32m     13\u001b[0m                      monitor\u001b[38;5;241m=\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mval_mae\u001b[39m\u001b[38;5;124m'\u001b[39m, \n\u001b[1;32m     14\u001b[0m                      mode\u001b[38;5;241m=\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mmin\u001b[39m\u001b[38;5;124m'\u001b[39m, \n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     17\u001b[0m                      save_best_only\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mTrue\u001b[39;00m\n\u001b[1;32m     18\u001b[0m                     )\n\u001b[0;32m---> 21\u001b[0m \u001b[43mmodel\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mfit\u001b[49m\u001b[43m(\u001b[49m\u001b[43m[\u001b[49m\u001b[43mtrain_X_mask\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mtrain_adj_mat\u001b[49m\u001b[43m]\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mtrain_ws_y\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m     22\u001b[0m \u001b[43m          \u001b[49m\u001b[43mvalidation_data\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43m(\u001b[49m\u001b[43m[\u001b[49m\u001b[43mtest_X_mask\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mtest_adj_mat\u001b[49m\u001b[43m]\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mtest_ws_y\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m     23\u001b[0m \u001b[43m          \u001b[49m\u001b[43mepochs\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mEPOCHS\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m     24\u001b[0m \u001b[43m          \u001b[49m\u001b[43mbatch_size\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mBATCH\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m     25\u001b[0m \u001b[43m          \u001b[49m\u001b[43mverbose\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;241;43m2\u001b[39;49m\u001b[43m,\u001b[49m\n\u001b[1;32m     26\u001b[0m \u001b[43m          \u001b[49m\u001b[43mshuffle\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43;01mTrue\u001b[39;49;00m\u001b[43m,\u001b[49m\n\u001b[1;32m     27\u001b[0m \u001b[43m          \u001b[49m\u001b[43mcallbacks\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43m[\u001b[49m\u001b[43mes\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mmc\u001b[49m\u001b[43m]\u001b[49m\n\u001b[1;32m     28\u001b[0m \u001b[43m         \u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/miniconda3/envs/tf213/lib/python3.10/site-packages/keras/src/utils/traceback_utils.py:65\u001b[0m, in \u001b[0;36mfilter_traceback.<locals>.error_handler\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m     63\u001b[0m filtered_tb \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m\n\u001b[1;32m     64\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[0;32m---> 65\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mfn\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     66\u001b[0m \u001b[38;5;28;01mexcept\u001b[39;00m \u001b[38;5;167;01mException\u001b[39;00m \u001b[38;5;28;01mas\u001b[39;00m e:\n\u001b[1;32m     67\u001b[0m     filtered_tb \u001b[38;5;241m=\u001b[39m _process_traceback_frames(e\u001b[38;5;241m.\u001b[39m__traceback__)\n",
      "File \u001b[0;32m~/miniconda3/envs/tf213/lib/python3.10/site-packages/keras/src/engine/training.py:1742\u001b[0m, in \u001b[0;36mModel.fit\u001b[0;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, validation_batch_size, validation_freq, max_queue_size, workers, use_multiprocessing)\u001b[0m\n\u001b[1;32m   1734\u001b[0m \u001b[38;5;28;01mwith\u001b[39;00m tf\u001b[38;5;241m.\u001b[39mprofiler\u001b[38;5;241m.\u001b[39mexperimental\u001b[38;5;241m.\u001b[39mTrace(\n\u001b[1;32m   1735\u001b[0m     \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mtrain\u001b[39m\u001b[38;5;124m\"\u001b[39m,\n\u001b[1;32m   1736\u001b[0m     epoch_num\u001b[38;5;241m=\u001b[39mepoch,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m   1739\u001b[0m     _r\u001b[38;5;241m=\u001b[39m\u001b[38;5;241m1\u001b[39m,\n\u001b[1;32m   1740\u001b[0m ):\n\u001b[1;32m   1741\u001b[0m     callbacks\u001b[38;5;241m.\u001b[39mon_train_batch_begin(step)\n\u001b[0;32m-> 1742\u001b[0m     tmp_logs \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mtrain_function\u001b[49m\u001b[43m(\u001b[49m\u001b[43miterator\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1743\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m data_handler\u001b[38;5;241m.\u001b[39mshould_sync:\n\u001b[1;32m   1744\u001b[0m         context\u001b[38;5;241m.\u001b[39masync_wait()\n",
      "File \u001b[0;32m~/miniconda3/envs/tf213/lib/python3.10/site-packages/tensorflow/python/util/traceback_utils.py:150\u001b[0m, in \u001b[0;36mfilter_traceback.<locals>.error_handler\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    148\u001b[0m filtered_tb \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m\n\u001b[1;32m    149\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[0;32m--> 150\u001b[0m   \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mfn\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    151\u001b[0m \u001b[38;5;28;01mexcept\u001b[39;00m \u001b[38;5;167;01mException\u001b[39;00m \u001b[38;5;28;01mas\u001b[39;00m e:\n\u001b[1;32m    152\u001b[0m   filtered_tb \u001b[38;5;241m=\u001b[39m _process_traceback_frames(e\u001b[38;5;241m.\u001b[39m__traceback__)\n",
      "File \u001b[0;32m~/miniconda3/envs/tf213/lib/python3.10/site-packages/tensorflow/python/eager/polymorphic_function/polymorphic_function.py:825\u001b[0m, in \u001b[0;36mFunction.__call__\u001b[0;34m(self, *args, **kwds)\u001b[0m\n\u001b[1;32m    822\u001b[0m compiler \u001b[38;5;241m=\u001b[39m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mxla\u001b[39m\u001b[38;5;124m\"\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_jit_compile \u001b[38;5;28;01melse\u001b[39;00m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mnonXla\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m    824\u001b[0m \u001b[38;5;28;01mwith\u001b[39;00m OptionalXlaContext(\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_jit_compile):\n\u001b[0;32m--> 825\u001b[0m   result \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_call\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwds\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    827\u001b[0m new_tracing_count \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mexperimental_get_tracing_count()\n\u001b[1;32m    828\u001b[0m without_tracing \u001b[38;5;241m=\u001b[39m (tracing_count \u001b[38;5;241m==\u001b[39m new_tracing_count)\n",
      "File \u001b[0;32m~/miniconda3/envs/tf213/lib/python3.10/site-packages/tensorflow/python/eager/polymorphic_function/polymorphic_function.py:857\u001b[0m, in \u001b[0;36mFunction._call\u001b[0;34m(self, *args, **kwds)\u001b[0m\n\u001b[1;32m    854\u001b[0m   \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_lock\u001b[38;5;241m.\u001b[39mrelease()\n\u001b[1;32m    855\u001b[0m   \u001b[38;5;66;03m# In this case we have created variables on the first call, so we run the\u001b[39;00m\n\u001b[1;32m    856\u001b[0m   \u001b[38;5;66;03m# defunned version which is guaranteed to never create variables.\u001b[39;00m\n\u001b[0;32m--> 857\u001b[0m   \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_no_variable_creation_fn\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwds\u001b[49m\u001b[43m)\u001b[49m  \u001b[38;5;66;03m# pylint: disable=not-callable\u001b[39;00m\n\u001b[1;32m    858\u001b[0m \u001b[38;5;28;01melif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_variable_creation_fn \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n\u001b[1;32m    859\u001b[0m   \u001b[38;5;66;03m# Release the lock early so that multiple threads can perform the call\u001b[39;00m\n\u001b[1;32m    860\u001b[0m   \u001b[38;5;66;03m# in parallel.\u001b[39;00m\n\u001b[1;32m    861\u001b[0m   \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_lock\u001b[38;5;241m.\u001b[39mrelease()\n",
      "File \u001b[0;32m~/miniconda3/envs/tf213/lib/python3.10/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compiler.py:148\u001b[0m, in \u001b[0;36mTracingCompiler.__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m    145\u001b[0m \u001b[38;5;28;01mwith\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_lock:\n\u001b[1;32m    146\u001b[0m   (concrete_function,\n\u001b[1;32m    147\u001b[0m    filtered_flat_args) \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_maybe_define_function(args, kwargs)\n\u001b[0;32m--> 148\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mconcrete_function\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_call_flat\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    149\u001b[0m \u001b[43m    \u001b[49m\u001b[43mfiltered_flat_args\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mcaptured_inputs\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mconcrete_function\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mcaptured_inputs\u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/miniconda3/envs/tf213/lib/python3.10/site-packages/tensorflow/python/eager/polymorphic_function/monomorphic_function.py:1349\u001b[0m, in \u001b[0;36mConcreteFunction._call_flat\u001b[0;34m(self, args, captured_inputs)\u001b[0m\n\u001b[1;32m   1345\u001b[0m possible_gradient_type \u001b[38;5;241m=\u001b[39m gradients_util\u001b[38;5;241m.\u001b[39mPossibleTapeGradientTypes(args)\n\u001b[1;32m   1346\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m (possible_gradient_type \u001b[38;5;241m==\u001b[39m gradients_util\u001b[38;5;241m.\u001b[39mPOSSIBLE_GRADIENT_TYPES_NONE\n\u001b[1;32m   1347\u001b[0m     \u001b[38;5;129;01mand\u001b[39;00m executing_eagerly):\n\u001b[1;32m   1348\u001b[0m   \u001b[38;5;66;03m# No tape is watching; skip to running the function.\u001b[39;00m\n\u001b[0;32m-> 1349\u001b[0m   \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_build_call_outputs(\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_inference_function\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m)\u001b[49m)\n\u001b[1;32m   1350\u001b[0m forward_backward \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_select_forward_and_backward_functions(\n\u001b[1;32m   1351\u001b[0m     args,\n\u001b[1;32m   1352\u001b[0m     possible_gradient_type,\n\u001b[1;32m   1353\u001b[0m     executing_eagerly)\n\u001b[1;32m   1354\u001b[0m forward_function, args_with_tangents \u001b[38;5;241m=\u001b[39m forward_backward\u001b[38;5;241m.\u001b[39mforward()\n",
      "File \u001b[0;32m~/miniconda3/envs/tf213/lib/python3.10/site-packages/tensorflow/python/eager/polymorphic_function/atomic_function.py:196\u001b[0m, in \u001b[0;36mAtomicFunction.__call__\u001b[0;34m(self, *args)\u001b[0m\n\u001b[1;32m    194\u001b[0m \u001b[38;5;28;01mwith\u001b[39;00m record\u001b[38;5;241m.\u001b[39mstop_recording():\n\u001b[1;32m    195\u001b[0m   \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_bound_context\u001b[38;5;241m.\u001b[39mexecuting_eagerly():\n\u001b[0;32m--> 196\u001b[0m     outputs \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_bound_context\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mcall_function\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    197\u001b[0m \u001b[43m        \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mname\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    198\u001b[0m \u001b[43m        \u001b[49m\u001b[38;5;28;43mlist\u001b[39;49m\u001b[43m(\u001b[49m\u001b[43margs\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    199\u001b[0m \u001b[43m        \u001b[49m\u001b[38;5;28;43mlen\u001b[39;49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mfunction_type\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mflat_outputs\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    200\u001b[0m \u001b[43m    \u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    201\u001b[0m   \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[1;32m    202\u001b[0m     outputs \u001b[38;5;241m=\u001b[39m make_call_op_in_graph(\u001b[38;5;28mself\u001b[39m, \u001b[38;5;28mlist\u001b[39m(args))\n",
      "File \u001b[0;32m~/miniconda3/envs/tf213/lib/python3.10/site-packages/tensorflow/python/eager/context.py:1457\u001b[0m, in \u001b[0;36mContext.call_function\u001b[0;34m(self, name, tensor_inputs, num_outputs)\u001b[0m\n\u001b[1;32m   1455\u001b[0m cancellation_context \u001b[38;5;241m=\u001b[39m cancellation\u001b[38;5;241m.\u001b[39mcontext()\n\u001b[1;32m   1456\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m cancellation_context \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n\u001b[0;32m-> 1457\u001b[0m   outputs \u001b[38;5;241m=\u001b[39m \u001b[43mexecute\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mexecute\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m   1458\u001b[0m \u001b[43m      \u001b[49m\u001b[43mname\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mdecode\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mutf-8\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1459\u001b[0m \u001b[43m      \u001b[49m\u001b[43mnum_outputs\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mnum_outputs\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1460\u001b[0m \u001b[43m      \u001b[49m\u001b[43minputs\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mtensor_inputs\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1461\u001b[0m \u001b[43m      \u001b[49m\u001b[43mattrs\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mattrs\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1462\u001b[0m \u001b[43m      \u001b[49m\u001b[43mctx\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1463\u001b[0m \u001b[43m  \u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1464\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[1;32m   1465\u001b[0m   outputs \u001b[38;5;241m=\u001b[39m execute\u001b[38;5;241m.\u001b[39mexecute_with_cancellation(\n\u001b[1;32m   1466\u001b[0m       name\u001b[38;5;241m.\u001b[39mdecode(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mutf-8\u001b[39m\u001b[38;5;124m\"\u001b[39m),\n\u001b[1;32m   1467\u001b[0m       num_outputs\u001b[38;5;241m=\u001b[39mnum_outputs,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m   1471\u001b[0m       cancellation_manager\u001b[38;5;241m=\u001b[39mcancellation_context,\n\u001b[1;32m   1472\u001b[0m   )\n",
      "File \u001b[0;32m~/miniconda3/envs/tf213/lib/python3.10/site-packages/tensorflow/python/eager/execute.py:53\u001b[0m, in \u001b[0;36mquick_execute\u001b[0;34m(op_name, num_outputs, inputs, attrs, ctx, name)\u001b[0m\n\u001b[1;32m     51\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[1;32m     52\u001b[0m   ctx\u001b[38;5;241m.\u001b[39mensure_initialized()\n\u001b[0;32m---> 53\u001b[0m   tensors \u001b[38;5;241m=\u001b[39m \u001b[43mpywrap_tfe\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mTFE_Py_Execute\u001b[49m\u001b[43m(\u001b[49m\u001b[43mctx\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_handle\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mdevice_name\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mop_name\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m     54\u001b[0m \u001b[43m                                      \u001b[49m\u001b[43minputs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mattrs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mnum_outputs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     55\u001b[0m \u001b[38;5;28;01mexcept\u001b[39;00m core\u001b[38;5;241m.\u001b[39m_NotOkStatusException \u001b[38;5;28;01mas\u001b[39;00m e:\n\u001b[1;32m     56\u001b[0m   \u001b[38;5;28;01mif\u001b[39;00m name \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "lr_schedule = keras.optimizers.schedules.ExponentialDecay(initial_learning_rate=learning_rate, \n",
    "                                                          decay_steps=decay_steps,\n",
    "                                                          decay_rate=decay_rate)\n",
    "\n",
    "model.compile(optimizer=Adam(learning_rate=lr_schedule),\n",
    "              loss='mse',\n",
    "              metrics=['mae']\n",
    "             )\n",
    "\n",
    "\n",
    "es = EarlyStopping(monitor='val_mae', mode='min', verbose=2, patience=PATIENCE)\n",
    "mc = ModelCheckpoint('saved_models/gcn_gate_pump_nofuture.h5', \n",
    "                     monitor='val_mae', \n",
    "                     mode='min', \n",
    "                     verbose=2, \n",
    "                     custom_objects={'GCNConv': GCNConv},\n",
    "                     save_best_only=True\n",
    "                    )\n",
    "\n",
    "\n",
    "model.fit([train_X_mask, train_adj_mat], train_ws_y,\n",
    "          validation_data=([test_X_mask, test_adj_mat], test_ws_y),\n",
    "          epochs=EPOCHS,\n",
    "          batch_size=BATCH,\n",
    "          verbose=2,\n",
    "          shuffle=True,\n",
    "          callbacks=[es, mc]\n",
    "         )"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "45d04c0b",
   "metadata": {},
   "source": [
    "### Performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "c60c3dba",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Layer lstm_1 will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU.\n",
      "603/603 [==============================] - 3s 4ms/step\n",
      "inv_y.shape, inv_yhat.shape (19268, 96) (19268, 96)\n",
      "MAE = 0.0747\n",
      "RMSE = 0.102\n"
     ]
    }
   ],
   "source": [
    "saved_model = load_model('saved_models/gcn_gate_pump_nofuture.h5', custom_objects={'GCNConv': GCNConv})\n",
    "\n",
    "yhat = saved_model.predict([test_X_mask, test_adj_mat])\n",
    "\n",
    "inv_yhat = ws_scaler.inverse_transform(yhat)\n",
    "inv_y = ws_scaler.inverse_transform(test_ws_y)\n",
    "\n",
    "print(\"inv_y.shape, inv_yhat.shape\", inv_y.shape, inv_yhat.shape)\n",
    "print('MAE = {}'.format(float(\"{:.4f}\".format(mae(inv_y, inv_yhat)))))\n",
    "print('RMSE = {}'.format(float(\"{:.4f}\".format(sqrt(mse(inv_y, inv_yhat))))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "10108ca4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABRIAAAFQCAYAAADDfoLGAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8pXeV/AAAACXBIWXMAAA9hAAAPYQGoP6dpAADNaElEQVR4nOzdd3wT5R8H8E+6dymjlF323kP2lulPRFygMhRFERQRla2CAi4QFEUFwQEoIKBsZJW99yqz0DJaOuikO78/0ksz7pJLchltP+/XixeQXO6eJJcb3+f7fB+VWq1Wg4iIiIiIiIiIiMgEN2c3gIiIiIiIiIiIiFwfA4lERERERERERERkFgOJREREREREREREZBYDiURERERERERERGQWA4lERERERERERERkFgOJREREREREREREZBYDiURERERERERERGQWA4lERERERERERERkFgOJREREREREREREZBYDiUREREQuSqVSWf1n2bJleOaZZ6BSqdCoUSOT22nbtq32dXv27JFc7s8//9QuFxkZafT89u3bMWTIENSqVQv+/v7w9fVFeHg42rZti9GjR2PlypV48OCBrR+L1q1btzBp0iS0bt0aISEh8PT0RPny5dGkSRMMGjQI33zzDc6cOWP0ury8POzcuRMTJkxA+/btUaZMGXh6eiIkJATt27fHrFmzkJSUZHLbcr6Dl156SfS1y5Ytk3yN8Jk999xz2LJliyKfExEREZFSVGq1Wu3sRhARERGRsbCwMNHH09LSkJ6eDgAoX7686DLz589HfHw8xowZA5VKhdjYWJQrV85oudTUVJQuXRq5ubkAgI8++ggff/yx6DpHjRqFn376CRUrVsSdO3e0j2dlZeHll1/G6tWrtY+5ubmhVKlSSE1NRU5OjvZxU+u3xPLly/H6668jIyND+1hQUBDy8/ORlpamfaxatWqIiorSe+1rr72GxYsX67U1KCgIDx8+1D5Wvnx5rF+/Hm3bthXdvkqlAgCEhITAy8tLdJlnnnkG3333ndHjy5Ytw4gRIwAAZcuWhbu7OwBArVYjMTFR+10AwIgRI/DLL7+Irp+IiIjI0ZiRSEREROSi7t+/L/pnwoQJZpd5/vnn0a1bNwCaAFVERIToNvbv34/c3Fxt0NJURqLwnLBewfvvv68NIr7yyis4ffo0srKykJCQgMzMTFy+fBnffvstOnbsqA3A2eLYsWMYOnQoMjIy0KRJE6xZswZpaWlITk5Gamoq4uLisH79egwfPhz+/v5Gr8/JyUFoaCgmTJiAgwcPIjMzE0lJSUhNTcXixYtRpkwZxMbGon///mYzKNeuXSv5HYgFEcXei7B8bGwssrKycPr0aTz++OMAgKVLl2LlypXWfVBERERECmMgkYiIiKiYatCggTZjcffu3aLLCMHBt956C/7+/jh8+DAyMzONlrt79y6uXLkCQD+QmJqaip9++gmAJmNxyZIlaNq0KTw8PABosv3q1q2LMWPGYN++ffjggw9sfl/ffPMN8vPzERoair1792LQoEF6AcNy5cphwIABWLp0KY4fP270+jfffBNRUVH48ssv0a5dO3h6egIAAgIC8Oqrr2LDhg0AgMTERPz44482t9cSbm5uaNq0KdatW4fg4GAAwPr16x3aBiIiIiIpDCQSERERFWNdu3YFIJ1pKDzes2dPtG/fHllZWTh06JDkcoB+IPHy5cvIysoCAAwYMMBse3x9feU13ITTp08D0Lw3IdhmyfYee+wxk+1o164dGjRoAECTMegM/v7+qFWrFgDoDdUmIiIiciYGEomIiIiKMSHod/HiRcTFxek9l5qaipMnT8Lf3x+tWrVCly5dAIgHHYXHqlatiho1aohuKyYmRrmGy2DP7fn4+ADQTMziDBkZGbh27RoAoG7duk5pAxEREZEhBhKJiIiIijHd7EHDAKFQH7FDhw7w8PDQBhLFhkELjxnWR2zUqBH8/PwAAJ988olDMvjatGkDADh48CC+/vprZGdnK7r++Ph4nD9/HgDQuHFjk8uOHz8eoaGh8PLyQrly5dCjRw8sXLhQbxIYS6jVapw7dw6DBg1CcnIyfH198dZbb1m1LiIiIiKlMZBIREREVIzVqVMHFStWBGAcSBT+LwQQ27RpA19fXxw5cgSPHj3SLnfnzh1tdpxhINHX1xdTpkzRLtemTRvUr18fI0eOxKJFi3DixAm9WYiVMHHiRAQGBgIAJkyYgLCwMAwcOBCfffYZtm7dqjf7sjWmTZuG7OxseHh4YPjw4SaXPXXqFDIyMuDj44P4+Hjs2rULY8aMQfPmzXHp0iWz22rdujXCwsK0f7y9vdGkSRPs2bMHTz31FA4fPoyaNWva9H6IiIiIlMJAIhEREVExJwT/DDMNhf8LdRS9vLzQtm1bZGdn4+DBg0bL6a5L1+TJk7FgwQKULl0agKZu4pIlS/Dmm2+iVatWKFu2LEaOHIkbN24o8n7q1q2LiIgItG7dGgCQlJSE9evXY+rUqejbty/KlCmDrl27WjVJyV9//YVFixYB0MxGLTWseOjQodi0aRMSEhKQlpaGlJQUREdHY9q0afD09MSVK1fQu3dvs0HN+Ph4xMbGav/k5OQAALKysvDw4UPcvXvX4vdAREREZC8MJBIREREVc0Kg8PLly4iNjQVQWB/Rz89PG5ADIFonUfh3jRo1ULVqVdFtjB07FjExMVizZg3eeusttG7dWjuhSXJyMpYsWYLGjRtj48aNiryn5s2b4+jRozh27Bg++eQT9OnTB2FhYQCA/Px8REREYODAgRgxYgTUarWsde7btw8jRowAAHTv3h0zZsyQXPbXX39Fv379tMFTAKhcuTJmzJiBP//8EwAQHR2NuXPnmtzmzZs3oVartX+ysrJw6dIlvP/++9i7dy/69++P77//Xlb7iYiIiOyNgUQiIiKiYk6sTuK+ffuQl5eH9u3bw9PTU/u8qUBi9+7dTW7H19cXgwYNwnfffYejR48iJSUFBw4cwLBhwwBoJhB54YUXcP/+fQXelUarVq0wffp0bNmyBffu3cPNmzfx1VdfoWzZsgCAZcuWYeHChWbXc+jQIfTv3x+PHj1Chw4d8M8//8DDw8OqNj399NPo0KEDAFicFenl5YV69erh888/x7hx45Cfn4933nkH169ft6otREREREpiIJGIiIiomKtZsyaqVKkCoHCYsmF9REHbtm3h7e2No0ePIiMjA9HR0dogltiwZlM8PDzQvn17LFu2DNOnTwcApKenazP27CE8PBzvvfceIiIitBmRixcvNvmaQ4cOoU+fPkhNTUW7du2wZcsWBAQE2NSOdu3aAYBNw7lHjhwJAMjNzcXq1attag8RERGREhhIJCIiIioBhCCgEEAU/haGPQt8fHzQpk0bbZ1E3cxESwOJukaNGqX9d2RkpNXrkatBgwbo2LGj2e0dPHgQvXv3RkpKCtq1a4dt27ZpJ3JxtmrVqmn/ffPmTSe2hIiIiEiDgUQiIiKiEkAIAkZGRiIyMhInT56Er68v2rRpY7SskKW4e/dubQZj3bp1UaFCBau3r5vh5+3tbfV6rNmm1PYOHjyol4m4detWxYKIhw8fBgBUr17d6nXExMRo/+3v729zm4iIiIhsxUAiERERUQmgm0346aefIi8vD+3atYOXl5fRsrp1EoWMRKlsxPj4eJw4ccLs9n/99Vftv1u0aGFJ043s2rVLO7uxlDt37mDHjh2S29MNIrZv3x7btm1DUFCQrO2bm7xl/fr12L9/PwBgwIABstYpZsWKFdp/t2rVyur1EBERESmFgUQiIiKiEqBatWra7LiVK1cCMK6PKBAmYDly5Ih2SK1UIPH+/fto1aoVunTpgkWLFiEyMlIbaMvLy0NkZCTeffddvPvuu9p2DBo0yKb38sEHH6BGjRqYOHEi9u/fj0ePHmmfS0xMxOLFi9GxY0ekpqYCAN577z291x8+fFgbROzQoYPFmYjPPfccJk6ciMOHDyMzM1P7+J07d/DJJ5/g+eefBwBUqlTJaNtyJCUl4ZtvvsGsWbMAaOo+Pv300xavh4iIiEhp1k1FR0RERERFTrdu3XDz5k3k5eUBMK6PKPDz80OrVq1w6NAh7WNSy3p4eEClUmHv3r3Yu3ev9rGgoCAkJydrtwUANWrUwIYNG2wepuvp6YmYmBh8/vnn+Pzzz6FSqRAUFIScnBxkZGRol/Py8sKXX36J/v37671+8uTJ2iDjxYsXUbt2bcltValSBceOHdN77MGDB1izZg0+//xzuLm5ITg4GHl5eUhJSdEuU7duXaxduxYhISEm30vr1q3h7u6u/X9OTg6SkpK0wdiqVati48aN8PHxMfOpEBEREdkfA4lEREREJUS3bt3wyy+/ANBMqvLYY49JLtulSxdtILFhw4YIDQ0VXa5evXqIjo7Gpk2bsG/fPpw9exa3bt1CcnIyvL29ERoaiiZNmuDJJ5/ESy+9pEh9xN27d2Pnzp3YtWsXjh07hqtXryIxMRFqtRply5ZFnTp10K1bN7zyyiuoUaOG0evz8/O1/05KSjK5LbEA3uTJk9G8eXMcOXIE0dHRSEhIQH5+PipWrIhmzZph4MCBeOmll2QF/+Lj4/X+7+HhgTJlyqBRo0Z48skn8dprr9k8gzQRERGRUlRqc0VeiIiIiIiIiIiIqMRjjUQiIiIiIiIiIiIyi4FEIiIiIiIiIiIiMouBRCIiIiIiIiIiIjKLk60QERERkcMdPHgQTz/9tEWvad++PdauXWunFhERERGROQwkEhEREZHDZWdnIzY21qLXJCYm2qk1RERERCQHZ20mIiIiIiIiIiIis5iR6KLy8/Nx9+5dBAYGQqVSObs5RERERERERETkRGq1GqmpqahYsSLc3Jwz7QkDiS7q7t27qFKlirObQURERERERERELiQ6OhqVK1d2yrYZSHRRgYGBADQ7R1BQkJNbQ0REREREREREzpSSkoIqVapoY0bOwECiixKGMwcFBTGQSEREREREREREAODUEnjOGVBNRERERERERERERQoDiURERERERERERGRWiQwkRkZG4ttvv8Xw4cPRuHFjeHh4QKVS4dNPP7VpvTt27EC/fv1QtmxZ+Pr6ol69epgyZQrS0tIUajkREREREREREZFzlMgaiT/88APmz5+v6DrnzZuH8ePHQ6VSoVOnTihfvjz27duHWbNm4e+//8b+/ftRtmxZRbdJRERERERERETkKCUyI7FRo0aYMGECli9fjkuXLuHll1+2aX2nTp3Ce++9B3d3d2zatAkRERFYtWoVrl+/jh49eiAyMhJvvPGGQq0nIiIiIiIiIiJyvBKZkThy5Ei9/7u52RZPnT17NtRqNUaMGIG+fftqH/fz88OSJUtQo0YN/P3337h8+TLq1atn07aIiIiIiIiIiIicoURmJCopOzsbmzZtAgAMGTLE6Plq1aqhQ4cOAIB169Y5tG1ERERERERF3YlbSYhOzHB2M4iICAwk2uzKlSvIyNCc1Fq1aiW6jPD4qVOnHNYuIiIiIiKi4mDQDwfR6Yvdzm4GERGBgUSb3bx5EwBQqlQpBAYGii5TpUoVvWWJiIiIiIiIiIiKmhJZI1FJqampAAB/f3/JZQICAgAAKSkpkstkZWUhKytL+39TyxIRERERERERETkaMxJdxOzZsxEcHKz9I2QxEhERERERERERuQIGEm0kDGdOT0+XXCYtLQ0AEBQUJLnMpEmTkJycrP0THR2tbEOJiIiIiIiIiIhswKHNNgoPDwcAPHz4EKmpqaJ1EoWgoLCsGG9vb3h7e9ujiURERERERERERDZjRqKN6tatCz8/PwDA8ePHRZcRHm/RooXD2kVERERERERERKQkBhJt5OXlhf79+wMAVqxYYfT8rVu3cPDgQQDAwIEDHdo2IiIiIiIiIiIipTCQKNN3332HevXqYejQoUbPTZw4ESqVCkuXLsXWrVu1j2dkZODVV19FXl4eBg0ahHr16jmyyURERERERERERIopkTUST548idGjR2v/f/36dQDAjz/+iI0bN2ofX7duHSpUqAAAiI+PR2RkJMLCwozW16JFC3z99dcYP348+vXrhy5duiA0NBT79u3DvXv3ULduXSxatMjO74qIiIiIiIiIiMh+SmQgMSUlBUeOHDF6PCYmBjExMdr/Z2VlyV7nu+++i8aNG+Prr7/G0aNHkZ6ejqpVq2LSpEmYNGmS6CQsRERERERERERERYVKrVarnd0IMpaSkoLg4GAkJycjKCjI2c0hIiIiIiJyivCJmwAAUXP6O7klRETO5QqxIpsyEjMyMrBv3z5ERETg0KFDuHv3Lh48eIDMzEyUKVMG5cqVQ/369dGlSxd06dIFdevWVardRERERERERERE5EBWBRJPnjyJH3/8EStXrkR6ejoAwDCx8c6dO7hz5w5Onz6NP//8EwDQvHlzjBo1CkOGDIG/v7+NTSciIiIiIiIiIiJHsSiQePr0aUyYMAG7d+/WBg59fHzQokULNG/eHGXLlkXp0qXh6+uLxMREJCYm4ubNmzhy5Ahu376NkydP4o033sAHH3yAyZMn45133oGXl5dd3hgREREREREREREpR3YgcdiwYVi+fDny8/NRtmxZPP/88xgyZAhat24NDw/zq4mNjcW///6LP/74A/v378fEiRPxww8/4Ndff0WnTp1sehNERERERERERERkX25yF/z9999Rr149rFy5Evfu3cO3336Ldu3ayQoiAkD58uXx2muvISIiAjdu3MDrr7+Ou3fvYvfu3VY3noiIiIiIiIiIiBxDdiBx5cqVOHfuHJ5//nm4u7vbtNFq1arhhx9+wLVr19CjRw+b1kVERERERERERET2J3to8/PPP6/4xitXrozKlSsrvl4iIiIiIiIiIiJSluyMRCIiIiIiIiIiIiq5bAokdu/eHc8++6zs5QcPHsyhzEREREREREREREWQ7KHNYvbs2YOwsDDZyx8+fBi3b9+2ZZNERERERERERETkBA4d2pyfnw+VSuXITRIREREREREREZECHBZIzMvLQ1xcHPz9/R21SSIiIiIiIiIiIlKIRUObU1JS8PDhQ73H8vLyEB0dDbVaLfoatVqNhw8fYunSpcjKykKTJk2sbiwRERERERERERE5h0WBxHnz5mHGjBl6j8XHxyM8PFzW61UqFV5++WVLNklEREREREREREQuwKJAolqt1ss8VKlUkpmIhipVqoQ33ngDY8aMsayFRERERERERERE5HQWBRLfffddjBgxAoAmqFijRg2UK1cOR48elXyNm5sbgoKCEBwcbFtLiYiIiIiIiIiIyGlkBxLHjx+P4OBgfPTRR9rHunTpgjJlyqBatWp2aRwRERERERERERG5BtmzNn/zzTf48ccf9R6LiIjAwYMHFW8UERERERERERERuRbZgUSVSoX8/Hx7toWIiIiIiIiIiIhclOxAYnBwMBITE5GZmWnP9hAREREREREREZELkl0jsVmzZoiIiMCwYcMwbNgwBAQEAACys7Oxb98+2bM3A0Dnzp0tbykRERERERERERE5jexA4rhx47Bnzx6sWbMGa9as0T6elJSErl27yt6gSqVCbm6uRY0kIiIiIiIiIiIi55I9tPnJJ5/E6tWr0aZNG/j6+kKtVkOlUkGtVlv0h3UWiYiIiIiIiIiIih7ZgUQAGDRoEA4dOoS0tDTk5+dDrVYjLCwM+fn5Fv0hIiIiIiIiIiKiosWiQCIRERERERERERGVTLJrJIq5efMm3N3dlWoLERERERERERFZKC9fDXc3lbObQSWATRmJ1apVQ+XKlZVqCxERERERERERWeDQ9QTUnLwZN+PTnd0UKgFkBxIzMjLs0gB7rZeIiIiIiIiIqLg7cSsRAHAzPs3JLaGSQHYgsXr16pg7dy4yMzMV2fDx48fxxBNP4Ouvv1ZkfURERERERERERGQ/sgOJ2dnZeP/99xEeHo6JEyfi3LlzFm8sLS0Nv/32G3r27InHHnsMmzdvRnBwsMXrISIiIiIiIiIiIseSPdnKtWvX8PHHH+PHH3/El19+iS+//BINGjRAp06d0KZNGzRt2hTlypVD6dKl4e3tjaSkJCQmJuLGjRs4evQojhw5goiICDx69AhqtRoNGjTAF198gX79+tnz/REREREREREREZECZAcSy5Qpg2+//RZvv/02vv76a6xYsQIXLlzAxYsX8eOPP5p9vVqtBgC0adMGb731Fl588UW4udk01wsREREREREREQEoCLsQ2ZXFkbzatWtj0aJFuHv3Ln766ScMHDgQZcuWhVqtFv3j7u6O1q1b44MPPsCpU6dw+PBhvPzyywwiEhERERERERERFSGyMxINBQQEYOTIkRg5ciQA4ObNm7h79y4ePHiAzMxMlClTBuXKlUPt2rXh7++vWIOJiIiIiIiIiEifSuXsFlBJYHUg0VD16tVRvXp1pVZHRERERERERERELoTji4mIiIiIiIiIiMgsBhKJiIiIiIiIiIo4TrZCjsBAIhEREREREREREZnFQCIRERERERERURHHyVbIEUp0IHH16tXo2rUrQkJC4O/vj6ZNm+KLL75ATk6ORetZtmwZVCqVyT9bt26107sgIipaohMzcPRmorObQURERERUrHBoMzmCYrM2FzXjxo3D/Pnz4eHhge7duyMgIAC7du3Chx9+iA0bNmD79u3w9fW1aJ01a9ZEx44dRZ+rVKmSEs0mIiryun61B3n5akTN6e/sphAREREREZEFSmQgcf369Zg/fz4CAgIQERGBFi1aAADi4+PRvXt37N+/H9OmTcNXX31l0Xo7duyIZcuW2aHFRETFR14+u0qJiIiIiIiKohI5tHnWrFkAgIkTJ2qDiABQtmxZfP/99wCA7777DsnJyU5pHxERERERERERkauxKZD422+/4bfffkNsbKxS7bG7O3fu4NixYwCAIUOGGD3fsWNHVKlSBVlZWdi8ebOjm0dERERELuznvTcwevkJZzeDiIiIyClsGto8fPhweHh44OHDhwo1x/5OnToFAChdujSqV68uukyrVq0QHR2NU6dOYfDgwbLXfe3aNUydOhVxcXEICAhAo0aN8OSTT6Js2bKKtJ2IiIiInOuzzZec3QQiIiIip7EpkFi6dGkAgJ+fnyKNcYSbN28CAKpWrSq5TJUqVfSWlevAgQM4cOCA3mM+Pj74+OOP8eGHH5p8bVZWFrKysrT/T0lJsWjbRERERERERFTyqFQqZzeBShCbhjbXq1cPycnJSEtLU6o9dpeamgoA8Pf3l1wmICAAgPxgXlhYGKZMmYIjR47gwYMHSElJwbFjxzB06FBkZWVh4sSJ2rqMUmbPno3g4GDtHyGYSURUXGXl5jm7CURERERERZ5azckMyXFsCiQOHz4ceXl5WLx4sVLtKZL69OmDTz/9FG3atEHZsmURGBiIVq1a4ddff9XO/DxjxgyTtSQnTZqE5ORk7Z/o6GhHNZ+ISpCo+HRnN0Hry62Rzm4CERERERERWcCmQOLIkSMxaNAgfPjhh/j++++Rm5urVLvsJjAwEACQni59My1kWAYFBdm8vXfeeQdly5ZFVlYWtm/fLrmct7c3goKC9P4QESnpeFQiun61B3si45zdFABAXGqW+YWIiOxgwc6rGLvylLObQURERFTk2FQj8ZVXXkFAQAC8vb0xduxYTJ8+Ha1bt0ZoaCjc3d1FX6NSqbBkyRJbNmuT8PBwADCZ8Sc8JyxrC3d3d9SuXRvx8fGIiYmxeX1ERNaKSXoEALidmOHklmiwlEvxkfwoB60/3YF/xnRA/QrsCCPXN/e/KwCAbwc3d3JLiIiIiIoWmwKJy5Ytg0ql0o7HT0xMxLZt20SXFZZzdiCxeXPNBWNCQgJu3rwpOnPz8ePHAQAtWrRQZJsJCQkACrMhiYiIipPI+6nIzsvHulN3nB5IzM9X46d9NzC0XTX4edl0mUNERERERAZsusIeOnRokZsdqHLlymjdujWOHTuGFStWYMqUKXrP79+/H9HR0fD29ka/fv1s3t7Jkydx5Yqm17tNmzY2r4+IyFpquFYR5qJ19iBTXKnA98nbSZiz5TIysvMw/vE6zm4OEREREVGxYnNGYlE0efJkDBw4EHPmzEHfvn21mYcJCQkYPXo0AGDMmDEIDg7WvmbdunWYNGkSKlWqhJ07d2ofz8jIwNKlSzF06FCjjMO9e/di2LBhAICOHTsykEhELsFVAnhFrSOKzHPUN3ouJhm+Xm6oFWqc6Z+brwlq5uTlO6g1REREREQlR4kc8/PUU0/h7bffxoIFC9C2bVv06NED/v7+2LlzJx4+fIgOHTpg5syZeq9JTk5GZGQkMjMz9R7Pzs7GmDFj8N5776F58+aoWrUqcnNzceXKFZw/fx4A0LhxY6xatcph74+IqChgGLH4cHQ+4v++2w8AiJrT37gtrpMcSUREROQQ7KAnRyqRgUQAmD9/Pjp06ICFCxfi4MGDyMnJQc2aNTFx4kS8++678PLykrUePz8/TJs2DcePH8fly5dx4cIFPHr0CCEhIejZsyeeffZZDB8+XPb6iIjsxeUCLLzeITvi7kVEREQlhSuVmaHiT7FAYmxsLNasWYPjx48jLi4OABAaGorWrVtj0KBBKF++vFKbUsxzzz2H5557Ttayw4cPx/Dhw40e9/LywowZMxRuGREREVnD1WqBEhEVVz/vvYEHaVmY3K++s5tCREQOZHMgMS8vD9OmTcPcuXORk5MDoDAarlKp8Ntvv2H8+PF47733MGPGDLi7u9u6SSIiKgZUzBkrflzoK+UIHyIi+/ps8yUAYCCRyAVwaDM5ks2BxKFDh+LPP/+EWq2Gt7c3WrVqhcqVKwMAYmJicPz4cWRlZWHOnDm4ffs2fv/9d5sbTURF29bz99GwYhCqlPZzdlNKJgsvND7ZcAG3EjLwy/DWijYjIT1L0fWR87jUaBpXagsRERERUTFjUyBx/fr1WLlyJQBg/PjxmDp1KkqVKqW3THJyMj777DN89dVXWLFiBZ599lk8+eSTtmyWiIq4N/44gbAgHxye3MPZTSEZlh6Isst690Q+sMt6yf7+PXMXoYHeaFujjN7jzDIlIiIiIire3Gx58ZIlS6BSqTBlyhR89dVXRkFEAAgODsYXX3yBKVOmQK1W4+eff7Zlk0RUTCRmZDu7CSWOS2WNFWFxKZmYuz2yRBe1fnvlKbzw02FnN8MkBjWJiIiopCnBl6fkQDYFEo8dOwY3NzdMmDDB7LITJkyAm5sbjh07ZssmiYiInGr6PxewYNc1xKVyaLbAlSY4cWZL3lp+Eo/PjXBiC4gst2T/TTT6aJuzm0FERERFhE2BxKSkJAQHByM4ONjsssJySUlJtmySqERLSMvCldhUZzdDGXa42z91Owm/HYoSfe5szEOcuJWo/EaLuZKcdScln5+JJFeq8+2Mtmw6dw9X49Icv2FymEfZedp/F5fz8dztkUjLynV2M4hIxMHr8bh8P8XZzaAixJWuxaj4simQGBISguTkZKSkmD+4JScnIzk5GSEhIbZskqhE67dgH3rN2+vsZrisgd8fxPR/Log+9+R3BzDoh0MObpFrsuT6YvaWy3ZrBxU/DzNynN0EDukhu9K9Qdty7r7zGqIgzvRJ5LqG/HwEfb7Z5+xmUBHAzn9yJJsCia1bt0Z+fj7mzZtndtl58+YhPz8frVq1smWTRCVabAqHUtrL5HXnED5xk7ObYbNRvx9HbwWDzX+fiFFsXVQ8ZeXmYc1xzX6y8uhtJ7cGeGnJEQCWBcyJ5NKNubnSkH5b8LdCRVVWbh4yc/LML0hERIqyKZA4YsQIqNVqzJw5E9OmTUNamvFwntTUVEydOhUzZ86ESqXCq6++assmiQAAx6MSsftynLObQTZwtRuwFUecHwBRwrYLsYiUGG7nWp940cXkHX0/RdzA2lN3nN0MslJ0YgbO30l2djOKpGKT/MFjGhUhuoHDLl/sQYPpW53YGudQq9UY9+cpXC0m5RVIGUJ2ebE5N5FLsymQ+PTTT+O5555Dfn4+Zs2ahfLly6Nr16548cUX8eKLL6JLly4ICwvD7NmzoVar8dxzz2HgwIFKtZ1KsGcWHcKIZZy4h0jKqmPRzm4ClRCprK3mEvLz1XhgxQRAnb7YjSe+3W+HFhV/xWUYGeOIVJScjSns+Lifkol8C36Gl+6l4NeDUco3ysEe5eRh/em7+GTDRWc3hYhKKJsCiQDw+++/Y/z48XB3d8ejR4+wd+9e/Pnnn/jzzz+xb98+PHr0CO7u7njvvffw22+/KdFmIiIyY9+1eEXWY+/b5Jvx6Xbegv0UkxiCzVw2CFHCUkcX7r6G1p/tsHqYX2ZOHlIynV/jsigpLocA1kikosSW3XXg9wfw0b/itbSVlpmTh5G/HsO95EcO2R6RgId0cgSbA4menp746quvEBUVhfnz5+Pll19Gr1690KtXL7z88suYP38+oqKi8OWXX8LT01OJNhNRMeDqQZiWM//DtguOK6SfmpmjaHZLvokueksuMBLTsxVojbRXf2VmsSk9vt6DTWfvObsZpvGC1SUcjdLMSp+dl2/V61/46TCafLxdySYVe9Yeso/cSEDzGduRY+V3RUTWyXfgT+78nWTsuBSHxftu2m0brlYmiFyDq99jUfHgYcuL9+7VFPRv0qQJKlasiLFjxyrSKCIiR1Op9E+8CenZWLDzKno3DLP7tq/FpaHn3AjMfroxBrepqsg684vKVUQRaaazXH+QjpkbL6J/kwrax07cSsSms/cx/X8NnNgyadfi0lArNED28lPXn0P7mmXRr3EF8wubMe7PU9p/M75pmdPRD53dhCJB99Bq7U38z/tuIikjBymPclAmwFuhllmP2StUlBSV3ZW/K/vKz1fjfkomKpbydXZTSEGPsvOQkJ6FyiF+zm4KmWFTRmLXrl3Ro0ePYlMjhojIGXrOjQAAHL2ZqNg600Tq1vFYrQyVg29jDLOWRv56HL8csF+Gg6UMPw9hf5brj8O3MXr5SUXasv70XUXWU5Tx3tVxtpy7j1mbLzm7GTZzlX0mNTMH4RM3Yd/VB85uCpHLSkzPxoFrCQAcfz3iShbuvob2c3aJXu9S0ZSRnYv607ei4+e7nd0UksGmQGJwcDCCg4MREhKiVHuIiEgB+64qUyORlHMzPh3hEzfhWpxlsyzqhn/vJ2ciKYN17IhcwY34dPy094YVr3StTh03F0mdik3JBABsOMMOAZJm0+7qhF1d6T7ckb8ew2u/Hdes28WOJY507o5m0p0MBhKLjc3nHFdSimxnUyCxVq1aSE1NRVaW5bMEEinB2qLypDy1Wm1Rr6CrXfo48tpSrVZj7ckYh9fHmrLuPADX68E+G/MQdx8W/2LkeyLjAADzdly16HW6maQ3HqQp2iYluOqNjIvERhyGCcdFj7MnORnw3X5sOXfP5X4r3Jdd11MLD2DOlsvObkYRYZ8f1p0ScL0kWLj7GmpO3iz6nKe7JoxhbV1gR0hKz+ZoIAt4uLnYyYhMsimQ+MILLyAnJwerVq1Sqj1EFpm89pyzm0AFVh6NRqOPtiE6MUPW8kXhxGqvJp6Kfojxq85gyX7HDk8VLrZc6aZRDeDJ7w6g21d7nN0Uu7tfkG1j6eQperuhznd3K8E1Zrz+McKajKySIzoxA3EF37010rJycfhGAn7ZfxPxaeY7bp0dnKKi40xMssNmsJVHs++6/tVByXU6+iEWRVw3elytVmPOlsuIS7X+WFdc2bOzzVTH8O7LcTh43b6jU6Li7XsdsmDnVeSZmDxQjrdXnnL49TYAxKVkovnM//D3yTsO33ZR5W6HQGJevhrT1p/H/WQem5RmUyDxnXfeQbt27TBmzBhs3izeW0BkTxfvpSiynpikDHT6YhceZth3hlpLxKVm4tD1BGc3Q7YD1zQXK52+MF3XIsnOswArwd5BTiGTNio+HblO6EmVe552RLBXyGLNynXdHmUpaqjR4+s96L9gn/axH/Zcx5Pf7Vd0Ow91hjLrDkHs8uUe7b//PXMXG88WDgdsPmM7fnHAhbMrzzr7jYWZn/bS6YvdaDNrp8llXvvtuOTxfuLfZ/HCT4cxY+NFdp6R4jT36K4RfBYOb0Wgn5EMxCQ9wqKI65ix4aKzm+Iy7NWnoxs8NBWkHLHsGIb8fER7fa60nZdi0fWrPTh1O8ku61fKv2fuYuZGx++X8Wma+52TLv75OIOmzqdjSjDdTszA74dv4YttzKRWmk2zNs+aNQudO3fGuXPn8L///Q8NGzZEhw4dEBoaCnd3d8nXTZ8+3ZbNEmmZuthccyIGVUv7oU310mbX8++Zu4hOfISjNxPRywGz9Mox5OcjuBaXhqg5/Z3dFFnELma2nr+HN/44iUsz+sDXS3NMePXXYwXLK0s3CJyUno0Qfy+r1+Wom5g/j0XDx1P6WGkvcjOWHPE5PEh1/dIYuXn5qDVlC2YMaIjW4aXx8FHhvnb9gX5v/OdbCy9UHmZkw91NhUAfTwC2DSmPvJ+KumGBkmt4e6VmtuInmlQEACRl5OCr7ZGoXtYfXeuWs1uWWu0pW+yyXkslpmejtJnffPjETXi9cw1M7lffQa2S77+LsYi8n4q9H3Qzeu6eTi+6YeD25SVHEF7GHzOfaqR9TOgASMnMga+nu3b4F7me/dfi8Xj98trzozOo1WqX2UdcI5xJ9qRWqzFvx1WMaB9u0XWaMzpdAWDl0dtoUCEITauUsnodSl9L3TeT4X7+TrLe5H1yRwpZ6kbB9Y9hpld+vhoqleOy48/fSYFKpUIlF5u92ZmDA1x9YMLU9eew+dx9RH7aB94ezjv/kW1sCiR+/PHHUKlU2ovW8+fP48IF80MkGEgkR5iw+gwA4MasfnCTmYLlSkPCbtvpxG8Peflq3HlofGGz/pQmQyr5UY72Rsle7ytRJ9Px8Xl7cWxKDyzYeQ0vt6tmNsBgyJHJELpZSI7a+xy9l0cnZmDAwgP4793OKBPgrcg6E9Oz0WLmf1g1qp2szgJb5BYMq9l09h6m/yN/GGCzGf8BAJpUDsa/YzoaXdjl5avx1fZIjO5aUxtslBKdmKEJJFpwjMrIzsOIZcewdERrdKsbKvt1RU1UfDq6frUHP7zYQnIZYRj48sO3XDKQCADZElm5ukN93N30Az77rsZj39V4zHyqkVFnTpOPt+N/TSvi28HNlW+sDbZfuA8fT3d0rlPO2U2xihJBAWEdb688hRdaV8GcQU1kve7DNWeR/CgHz7aqjB71y9veEGjOd8I+plarJY8xyRk52B0Zh6eaV1Jku6bbZPuHrFarMfe/K3i5bTWEBvko0CrX1m72TnSpU072vqS07/dcA2D++un6gzQs2HkVcSmZFrU1NVO3Brf1VzHCK6Pi0xFe1t/s8pMKssCt6dR31h3FgIUHbB4ObM7Rm4m4cFcz2YnhIaPG5M14qW1VfPpUY5u3I+eS540/TsDdTYXrs/rZvD0l6WZYq9VqxKZkISzYMcciV8/qFjpIs3Lz9QKJLnQbTjLY1AXZuXNndO7cGV26dNH+ER4z9YfIkf44csvsMkoecLecu4cz0Q9tX5GLnwR0fb71ssXvWemTnO7NT3xaFmKSHmHejitGwxlikuQHMh3xFUTGWjaDrxJkZyQqtL1N5+4hMT0bx6KUG95xs6Auz/YL9p/h7XrBBCfW7rNnYwoutg0ef+Lb/fhhz3XRelOGhE1bc5HlSiUb5FKr1ThfMCOjOcJMr2dipJcXhoGnZ+chfOImm9tnjqlaPDFJGRZliOh+5XL6xN7444T237rD3V3F67+fwNBfjjq7GS7jngV1m/46Ho2tF+7j1V+PKzbZXL5aDQ93zY41ftUZTP/nvOhyE9eexbi/TuP8nWS8suyY1YGKyPvS5zwlO3MfpGXh213XMGW9+Pspbu4lZ+LPY9F23YapGq0rj8rbtpBYqHs+zctX4/fDt+we/NLV1aAu868Ho5CsU0ZEzLc7r+KdP0/J3oY9khPklJxxRFma5348hPWnNecXsfe56liM3dug2+HgyH1HrsJRKGr8dSwabWfvxL3kkjNRjinaT8aRicaut4sUeTZlJO7Zs0ehZhBZR04Q5pJCdRTlenP5SQDW9V7qcuRMqHn5avy87wZe6VAdXh6W9y/8tFf+ZAu61zc349NRXUaPsByGN9hCFqphlk/Hz3djyzudUL9CkN7jKpVK27iiMBGMpX49GKX9t9xL2+L4OVij/wJNzcOjUYl6j1v68RheawvHplwZF8DCd3HTisLm+U4YESYMxbbWX8eiMXHtOdHfqiEhCJLnjDcq4fXfj+PfMR1Fn+v4uaaOrNxzhO5+4yZxY3o8KhEHrmmym4W/AdfPSnBl/565i14Nytu9/IS5rygvX42E9CyEBtonk0WtLsxIXHdKMynAjAGNjJZLKuiQ+GJbJPZeeYD4tCyUN5Ppl5OXj1sJ6agVqjkWRFx5gGG/HMUfrz6GjrXLGi2/9bxtHUPzd1zFvB1XEDWnv/YmnucxZajVarT6dIfN6xECProjhTacuYtp688jyMcDA5oZZ7wei0rEfxdjbd62lPvJmfjo3ws4E/0Qc59vJrrMiVuJ+Pq/KwCA+S9YluVt733wamwqQgN9EOzniX9O34HhJYW9s7ykzkukoVYDe68+AKCpeV0h2PlDsBPTs3EzPh0tq4U4ZfvCPpNn8Nuwx0+Fe6f92JSR+Ntvv+G3335DbKz9Du5E5pjL7Lh4V34g0fBg8zAjG6mZmh7KQ9cTLOrxev7HQ7KXtVZuXj5iUzIx/Z/zNtWP2XEpFnO2XMafx24r2DrzohScddaw/pxHwUWq2GQQYtlCuq+2922HLbXyrKU7M6ebg8thCRcG1lxr5uer8WPEdQxYeAB/HtXdP4vezaHU9y514aR78yEcej5Yc9bi7eY74Ua69zd7LX6NbjbIrYLjepKMbEphuK+cgKyjPDST2SLYduG+2SwYXbq/oX0FNyYA8Mwi/fPNMYOgd1G18uhtfOyEWYWvxqbi7ZWnsHD3NYdv29A3O66gzWc77TaxUb5arT1fmiIcRvILfmc3HqSj1ac7TO6/n2+5jJ5z92qzJ+8kabJx7j4Uz8rR1pi18qe8XGcEivCWXOiwIEnsu83Ny8e09eeRIGOmdkc4efuh3v8zc/LEJ88z83kL56OVR29r96VHBftHVo74Pv7sokN6Hda2xK3EXiu06VFOHjJz8qBWq3Hxbop2MjgAuCtSvsfstqxupTSx0/nj8/bi+Z8054C5BcFOR2KgRpzuvrb5nKaTxFFB16txaUaPzd0eqR3p8dLiIxj0w0GHtEWMSnt8VvYAfSshHSuP6t/LFoFTQJFl0+3k8OHD8eqrryIw0PqsAyJbmbt5tOUistmM/9Bhzi5ciU3F4J8PY5lOVpc5R25afiOXkZ2L0ctPIDkjR1avzLM/HsJjs3bit0O3cE7mMEAxQtZeTp5jD7fHrPiMpBiem7W1SaxYl9MSGBS+vsiX2PnlBjKV+hiE7Fpr3t6+a/GYvUUzdH6iyIy1xaEjXOp70t8P5X0bUSIZi65U+9WUMStPWvya/Hx1YcDAhSIGsWaK4QtG/X4C7646bXIZ3d/rFp2MLVMBtmd1Aou5eflISs+WnBXaUPjETVZlvpqTkJalV8tWzJ2HjxA+cRM6zNmFudsjMWntOSw7GOXwrDIhsKFfl01Dkfp9FiwrXEvE2WliKrUauBJrfNP5894bemUAhK9A6FBdezIG8WlZaDpjO5oVzBKfnJGDgzozcV4syLoWbhYLz8umP4G1BZmRlhLL3nVGR4ocqZk5WHMiBv+euYvaU7YYBQzP303B74dv4Tsbgtl5+Wr8dey2IsdGw9Ed/565i+Yz/zNabr/BTKzX4tL09gndryNLoi6sPWWKBCuF/SY7Nx/1pm3F0gNR6LdgHxp9tE2RbSp2LaWW/uVcNlEywN4c3TktcNGfthHddjpqXivdif8EC3Zdw/ClmrIi9jjHW0KlwPE5LiUT/ebv0yvfM+yXo9q6psYbtXpTJMGm3bl06dIIDg6Gn5+fUu2hEiItKxcLdl61+OJm1O/H8cJP+pkXSt5gqFSaA5OulMxcpBf0Sgq96XJZ2radl+Kw+dx9PP3DAZMB0q3n7wEATun0ECsRLJi58SKyc/PxKNu49tLOS7FoKXLRaIvv91xHdGIG/j5hWy2VxftuYNhS/ZpbhcOarF+vUvvWzfh0Wdkka09ad/MkJUdiqKfUrtJ3/j4854BMWlMOXotH6892aG9Wh0nUUisqF5C6pJJ+pH7qug8vPyIvW/itFZYH4yx1+Ia8gJTU5CGCM9EPsfZk4W8/xsLjK6CZ+Vz7W7f41fYj5wZZOL7suhwn+nx0YgY+33rZ6Pe6+3KcRTUG9159gFd/PYbBPx9GRrZxYEyM3O/YEi0/3YEWEucQtVqNmKQMnCuoc3nn4SMs2FUYQPlBRh1RW8WlZmrLDQi/SSVi8DM2XES72TttXs80g1p/UsfA3w9F4cQt+fVodbOudC03qC8thDCEoWi6m3+YkYMZGy/irRUnMWTxEaN1abPSC/4vFsyxRGZOnna0iC7dbB/h3656rpi58SImrD6jPQYmGATZlegg+ftkDD78+xz+u2T7yDHD30KuRMdz8qMcpBR8N9cfpKHn3Ai9fUI3cCBccwt1MxN1RgGZbIuljZeQl6/GvP+uaPdH4bi9O1L8mKzr8I0Es6WTlO7Dm7HxImpO3qzsSk3YdTkWk9dJBGV02HOUjSXrFu5dMnPysOXcPaPnf7TjeSQzJ097rygQa7mHs6KuWsatepSdp1jNXUtbIef4nJyRg3Wn9O8T07JysfRgFC7eS8G+q4UdFUInYPjETTh4Xb9Tw6UuEosJm/bmevXqITk5GWlpxj2ZRKYs3H0Nc/+7gpO3LZt8YduFWBy+YVCnDMDZmIdGF72Cfo0rmF3v0gNRADQ9em1m7TSqxSLUD/rn9B3siYzD+lN38Mdh85O4NPxom97wM0FSeja+2XHFKFAlXHRcf2C6p+iNP04aDV1T6jT+zKKDqD99q9HjC3ZdQ0J6tqzg2uJ9N5Cfr8al+5qLLN3PwPBiefjSo3ivYIZtcw7fSBCtk/Pppku4IfGZ7bgUa/HQOCXrU2bn5qPbV3vw5bZIzbrVanyxzbin0JHe+fO06OOX7qXgqE6WqD0nxJHyQ8R1PEjNQrrMgEdRybYDIHlXIdUjm5BemKGie6FkKaWHRL7w02FZy9WZugVtPtshuf0BCw9g/KrC375o4NHMPhiflqU3M6Khb3dexTc75A31OnwjQdGsRrESCr/rnDeeN/gc1VAjMycPLy85grsPH+H9NWfww57rRp/fxLVnsffKA9lHqexcNaILgrSPzdopPhzRgFqtyWQUG7a6/tQdxYdOrzh6Gx0/3437EoXov9gaif1mfgPDfjmKpQduWrX9V5YdQ5vPdqLv/H0ACn+TSgxD++XATaMJVfRKaZg40Oruj1fj9DOOpM5T0/65oPiQtanrz2knyxLatEakA9BwMjOpkQIf2ThcfcB3B9D44+3YcTFWm32SkZ2r3wFrp6FzShGyXaVK5hRmVFq/jYyCwIYSQQLDX4KpdQpBxmUF19a6dL+PPw7fQptZO7WjfeZsuYxuBpOg2NORGwmYv/Mqft6nGTZtmE1pygs/HdYeL37ZfxMPdDKGwyduwgqdzj9LdsHzd5KxRyKQuVTk8zQkti1rfwKvLDuu9z6kjFh2DDFJGdqJ6ZxFyLL7ensk3lx+0uh49KMF9dwNTVh9Bu+tkr5X6Tk3Ag0/2qZ3bhTLwHaXM2OaAyU/ykH96VvR+YvdDt2u8NnIKRk2ed05vPvXGb1AbZ9v9uKHPZrAsO55WjfwfOi65prO0UHSksTmoc15eXlYvHixUu2hEkKo52d4+MjPV2szEuQ6djMRT353AFPWFfbW69ZF9HRX4ee9N0Sz7ATCTHR3Cur2vPbbcb3nhYNUQno2hi89hnF/ncZUkZkADXsnM7Lz8PKSo9qbsdTMHKw/dQezNl/CNzuuIipB/sydhp5dZHnmWPjETVh13PTMemfNfP5yLmo/3XQJm87dw62C9/d+QV23OlO3GC2bYfC9nLqdhI/+Mcy8UOO7XVfxwk+H9b6buw8fyUrPt2RIumZ7+n/bQtjvhGEn+Wr9TFJ7srX9SgVUddthrqapt4dmYoMt5+4h+ZF4ZsKNB2n4+6T9ZwQ0R6nbU6kLqb7f7DP7WsNe8AsiNWHL+HtZ3zgbxaVmSX6Phu48fITI+6n4+N8L2gtEc/Rrmxp/jl//dwXf7Lhqdj0X76bghZ8O6wX6BOETN2HSWstrUxoGfm4lpOtllR01KO0Qm5KFvVceYN/VeL2AmGFGhruFwS3d/Ss1M1d0OKKYHnMj0HTGduTlq/XOn+P+Om3V+ccU4Zxt6pz40hJNVtOuy7EIn7jJKLsy4soDfLLholXbN8wKVWsDicbLJllQ01KK2C9euOE5eC0earUayY9yUGPyZu1+Ep2oH2RVq4GUzByzw8WV8MfhwmCCVAYjYPr7A5TLXBIm2hv523FtPawG07fpBXNcPfvE8GcsnCfj07LwMCNbcjICa7zz52l8ssG24K1hp92MjdK/tRYz/8OaEzES9QgL//3ppktGz8enZRvVGDcOSCuzH41YdgyA+ex5U9KzcjFj40VM/PssPtt0UTsr75L9N6za35/4dj+GLz1mVVsirhgnLhhauPsablgY8LuX/Ag7zEx20/Hz3ejxdUThA06Ilwn3csIxelHEdb2Zxm2J4a05EWPyulMYUfHsokM4oA1IF2ZFt69ZxvqNS9h/Nd7sLND7r8abzLAVfmv2Kp0hxZLSE2K1snVHsOgeDnT/rVYDMzdd1Ab8SXk2BRJHjhyJQYMG4cMPP8T333+P3Fx5GSREhr3zgmUHo/C/7/aL1vmSYlg37dD1BPRbUHjQOHX7IT7bfAk/7DFfZ0bqHPPEt/vNvvbozUTJg1VuwRDTTzZcxLi/TuN+wfBpw+3ZcpEtHDzbzd6J6f8YBzkFi0RS+y25TNU96E9edw7PLhLPfEgxGJ6yOzJO9GLNMOPjzT9O4tdD+jfzWbn5+Gp7YVaREJhoP2eXZO+1EkEwJTIZDNdhLqNTrVbjWpwytW7+uxhr9D0408azd9HJTK+nd8Gs4R/+fU6093f9qTt4ZtEhrDxqOiBurYcZ2Viy/6bFw9oNOxHELtwsvYg1zN797VCU0TJ9vtmLhgZ1nAwDk87O2rTkd/TMooN6gX+zgQmVbcPHTkc/xIdrzmqPKfclahtas78ZHs+7fLnH7Gte//2E2WXcrLgbMnzFoesJSMvK1RZeN3TkZoK2I6jm5M2iWeqWiDRTv0v4DuV0+vx9QlMCwpYAmrnMU+FpsYzEz7dYl1FubsjmxxsuoN60rRiy+AhWHY/G+2Yy9dUAOn2+W2+4uNTMtuETN2Hl0ds2TcomsKkWm4MORUq8T3szPD4IHQ+tPt2BZjP+02ZT2RLk0iUnm80US4+zi/eJZ38ZDlEUc+RmIq7EpuJ4VCJOWThyyRLCUGbD48FpCzp7dxQMG49OysDP+26KjoBRcpSLKcN+OWp2W19ui0R33YCfBN3fULvZuzDyt+NYfuSWaEKAPWw8e9fkZJlqtVqvZjAAbckPIVD6x+HbBkFpZQ5AF++mSNfhg/GxXo3C3481JVykvLTkCAYulM4+f+fPU3hpyRGMEAlMC+1xVm1NUyNJBMJ1uLDMQ4lOaVPf6jaDfYSUZVMg8ZVXXkFAQAC8vb0xduxYhIWFoW/fvhg2bBheeeUV0T+vvvqqUm2nIiozJw8bz2pqV6Rl5uLjfy/gdkIG1p6M0fZwHrqRYFR4Wm5qclyq/o2gEMTLlHExpnuh9O1O81ksuqRuQHUJM3kKwxR3GmRB2HJDfOCapqbVveRM/HbolvSNklpzsxE+cROOmxieJpUlpXvQX3Hktna4k6nlAIieyIDCLFCB2D2yYSDiUxM94YUNML/I/eRM9J2/D0np2aKf/dW4NJvqJD7KzsOVgsyJvVceICcv32yzVh2PRs+5e3FNZMY1S41deQofrLY8k0og961fi0vDMhlDCv85fdfk8xfuJuvN5P1AZLbKcX+d1stEysrJU+xGC9BMYDFz40WLe2eH/HxYb2ICsf1dNyihezyT+7sXm0FWLNBmuM/+e8b0574nMs5oIo7L91MkZ1a1VFZOvuzafIbk1GfacVFzHLXmp/rOn6fwl06WtpJxDluO53rvxWA9wo2I3PerFrm1HPzzYXyw5ozkUDQlstxuPEhDQloW1Gq1yVm8E9KycE/mjKhL9t8snLzJzAd8K0G6Q/IPnVIoYr8r4fwptg2pfTk9KxeNPtomGZz9TCT7Spfu7/RWQga2m8kAErIWdc0zMWPrpLXnUGuKvCDAi4sP2zRiwpCw/xl+mnn5aizYedXkiBFr1JqyBa0+U66mc1ZuHrJyFR4eZzC0b8yKU3pPC0MkhXPFZ5suYvxfp82u9r1VZ7TnxFgLz2Mbz97FXomsNksPZ5fvp4oeo3SzW03pNW8vnll0CAO/V26YvtT1nOE1ZqqJrFtDUqViAN2ZaWWvzi7MbT41M8fo3mnJfuNrugU7ryI7Nx9/HpX3HQKwOvg4ZsUpvaQQQ6bqyOuW79QtFRWv0AzoE1afMZoZWJdwrScWLBv8s7zSMKbcS36krS1v6t5T7Jo7Pi1Lcy/i4H0yJy9f71pd6Egx1dGsHR1WsAd3mLMLzy46iD4G1xO652nXGjhe/HnY8uJly5ZBpVJpD8yJiYnYtk18hithOZVKhSVLltiyWSridFPf31x+Apk5+UZZCJPWnsMPe65j7wfdAEDvBt1SwsHqp703UC7AG691riG5rG7Nn69NXJBba4dBweuZGy/i1Y7Vtf+35QD4+dbLeLNrTe3/a0zejO3vdsZfx6IxtX997YFWDeDDvzXBpWcKhqeNf7yO0fqmrj+H2U83MXpcbnaRpdl8wvFBaOdbK07iq2eawtfL3eiEF5P0yOxwyRsmslrVUOPIjQRMWnsON+LT0Xzmf/B0L/z0dbfX7as92PN+N73Xn7+TjEv3UvBsqyom2zBh9Rls0in4fDU2DbVCA0y+Rpg9MyEty+yycogF48QYBhTy89U4ZDDhwujlJ/D9iy2NXjv458N4kJqF4R2qGz0HyJ+0pv8C85m/gP5EFr8euoXN5+9jxcjHULt8oKzXmyIExi29yJIz3FH39z3y1+OSywHimUWxKfK+S8OmbzhzFwOaVsRnmy+hZjl/LB7WWu95YRhV1Jz+2sf6FAyr1n3MWkIW6tXP+sKzYMrCt1eeEl3W0mPg3qvx2qGflnxlwvHGnlQwn4UmZbHOTZxUQFfurIu7LokPa7qT9AiTTWRV2Kr71xEo5eeJBS80N7lcuzm7ZHcGfL/7GiqF+AIwva/svBSLV389jrWj26NF1RCj5+/qBC6F+rWCn/ZeR9XS/gA0GfxCFv+bXWviwz71JIMCN+PTkZaVi1XHo9GoUjD+PhGDUn6e2ufTLQiWfS9jaL9YM+Tu0rcTMkwOTRU6Jq2VkpmDK/dTjbLuDH9zB67FY+5/VzD3vyt4T+Q6xBY5EpOBWEqtVqPu1K3w8nDDlU/7yn7dnsg45OSp8XiD8nqP34xPx/R/zsPPS1PG46BOJ45u7W3hNyFMBPLzPs0xYe7zzUxu9++TMRjcpgpahZeWVSJi87l76FE/FN4e7tpgpthx35rDpdL1eQXWHrmlzutyy28IxAL2wrVbepbx73zFkdt4t2cdlAv01j6281IsutYNVbxmnjUBol7z9uJecqbe9244IgIovAYxHAlmyvwdmuBjXEom9l+Lx9MtKlvUNmv2O0teolarsfzIbTzbqrK2tI4pcspvCSMLdK/hDY+pJ28noXyQDyqV0pzPtp6/hzf+OInLM/vAx9Md4RM3YUq/+qL3rK/9dhzn7xRma+67+gCdapcz2y7Bwt3XtJOS6Np/NR4da5eVvR5LdP1yD+LTshBZcAyVUyPxp303sPX8fe1IJQCiySsqleY3LNbJ5+wROcWdTYHEoUOH8gsii+lmoJmaue+2mVpqYmZsuIhqZaRnEf9s8yWTgURrZxL8YM0ZrDouPVxD7nld6Z9Tr3maXpv3e9eFj2fhCdJwM2Izz209f180kCj3IsVUZoQYtVp/mOKms/ewN/IBjk3taRSUPHQjAXO3R4qspZC5CSEMJzrQa4vON6ablXE6+iGux6VpJ4cxF0g8f1f/gkOlMj/ERXivjj62/nWsMCtr6/n7uPPwEWYa3GhuPncfTT/ZjjMf9dJ73F43C1I7m+HDD1Kz8HjBvr5gcHM82bSixZu68SANm87e0zs+fbfrKlbbOKO4FN2i7mLftGGdVkuIHTtHFqxPbvBJV2pmDrp+uQcrXmuLumHWB2vz8tUQDkNSWZIpmcYXgidvJ+HHiOv46tmmCPTx1HsuTWd5S7KHq0/ajHMfF+7Hur/L+8mZCPH3NLqpeHvlKbzRpSYaVAwyWl98WhYS0gxuvFTAiyIz2FrK1qFQa0/d0buBFZwxcUMkdvwJn7gJa0e3l7VNYUKXhxk5kjNMLz1wE8PahVuUUZyQnq29wRU6FKatP29U21LYz+8+fGQUSFx3Kga/mhhCPWuz+NDlH/Zcx4d96onuZxvO3NVeewifnKlJxHRrMu67Gm9VZ6nc3V1sdMXC3deMOjaV1OTj7Xr/Fz4zw71KGDECGHfexqdloWyA/n77wZoz2CoyVG32FtPZnqejH6JWaAACvC2/7fmlYEiwsJ9O/+c8BrepivoVjI8Dugw7aKITMzBvxxWsPXlH8jW6tbcLR6wYl0cxd31g6umFu6/hx4jrOPtxb9yMT8fo5SfxVreaeL93PZPrdPVcHzmfi1QH9+5I87UFdc+r802MWJKasOXjDRfw3eDmUKlUOHU7Ca/+ehyfPNkQw9qHiy5/4W4yLt4132Et5zgQeT8Veflqvd+bLrFyU6Yyuq3RZpZm5npLA4m6Jqw+g8FtqqJltRCT79uSy+djUUmYuv48kh/l4K1utQAAV2NTUSbAG6UL6kvrZiO/v+ZMYWKGWo09Vx4gU6KTSBhZtOuy8bH26YJM21c7Vse0JxpgTUHJjiE/H8bfb2rOsxvP3hW9Z00zuE56eclRizp9pepGv7TkCM581As+nm5mg6p7IuNQNsAbjSoFy9qm4egz4TM0la07p6CMiKn7ekBTqmVUQfC2YrCP9nFHlRQoyWzOSCRyJb+IDK/cekH/ojM3Lx+v/nocY7rXQp3ygQj29TR6jaVMBREBTYF7wwtiMctlzI5mjYzsPGwr+BxuxqfrHWgB8ZpbuoffuJRMxBcMkZm87hzmPd8MT3xrunitpQXp89VquEGlN/wzNSsX9aZtxbZxnY2WtybQLNh+wfikrpu9INVD9tTCA3r//37PNfyyPwpfPNMY3eqGGl3EGk6KoFKZv+gTC5It2X8TH/apa9fg4k86s9n9ejBKchiXWM+90ObrD9JQs5xxFqW1QyisGQ609MBNo0BiZk4ePN3djHr+M3PyUG/aVvw8tBW+3HZZm1EAaL4r3bqcYiyfDVweWy/ge8iof2SJyPupSEjPxurj0Zj6RANF1y2HcMG97cJ2XJ/VT+97dLOhQMuQn49o6wAKGZIqFdB29k70axyG719siQs6nQH/nrmLozcT8fVzTVGtjB8qhxRe3D6xYL/RECM3lcrs5FWO8sDCIY5Swxt1h7TdeJCGsoHeCPIxPofKGb71yYaLNs1G3u2rPYia0190ghyB2LHn3b9M1x40JT0rV/TcM3blKXxTkClm7jgtZ5ZKOZp+Uhis6zBnF7w93XDjgfGxwzBA99E/5/WG85vz4Rrry2MIhHdsKrhqqNWnO3Dts77wcC/8kUtda/0YIT0ba26eGk8tPICe9UONsrHl2HS2sNMjPi0Lvx26heNRSdj8TicAmhImKhX0OmsNpWbmmK0PbEiqA2HX5Tj0qK/JctxxMRZ1ygeiqtGNtvQ+KGTgfrPjCjYVlBnacv6+2UCiNYlzSl2yGP6OxdZbfdJmnJ7+OEr5SU8uZssv78hN6VJAUnTbuensPbzTozbqlA/UZi0alnDSJYzQMBdIvCiSCGBo2cEoXH+QZtHxdpvIdbI1DD/znLx83E/ORJXSpoNDYtaciBGdLV5XamaORSFvoYNA6PyKik/H4/P2olIpX7zTszY+MDj+ualU2vekVkuXbdJl6n5oyf6bmPZEA23HzsnbD3WG9Iqz571Ap893ISM7D9dm9QOgCRgOX3oMhyf1QJjOvaPQUbL93c6oY8VoIOF4Imf0WryZ6xfdUY13JeZgAFx+/q0iyaYaiUSWsrRO1qPsPIxfdVrRNszfeRURVx7g2UWH0HbWTptrZvSeJ137SSA1IYhg6/l7CJ+4SfQkv+nsPdSavNkoiGWJmRsv6tVxMTzQis0C+jAjRzsbbJtZO7W9SetO3UGLmf/ppdUrQQ3NDZbYTZpYfa0sG+ri/XnM9E2URKetkS+2RiI+LQuvLDuOTefu4eLdFKjVau0saoYBWjmT6RQWF9b8/dmmi1gUcV0b8JCSlZuHn/YaD2OKTswwO9GBYX2qQzcScNKCYuNCW3vOjUD72TvRa55+IMvak7cSk90AQL1pW0Vn3RWComtPxugFEQF5F+dSExvYqudcZQOBphhmV52LSUaXL8VveJX4NrJz87V1Q61hmI2k2/FgabblOZ1advEG2YSbz93HP6fvGM0yfj8lEy8uPmI0AZdYnaJchYZWuhLdQE73ryMwRCJgKLeAu+FsyUrRLeWhpCafbNeruaVLmNgqLjUTZ6IfGj0v/NaUOq7puvPwkWgQUYzhZGbmWBJ0lCL1ls19FHJrOpoilOgw/Hxy8vK1AfYVR27rZUmr1WoM++Uozt9J1jsXGgYVAKD+9K2oN20rjtwQHw4ePnETGhtkaNpCmDU7MycPI387js5f7hYNrtw3cUMNaLKSrhZkTBl+Ngt3XzM6N4hNPGSOo+uwzdp8CQlpWdqgkKFFMoZ6S5HqXLHEQYNsRSU+HsPzjNRnbnh/ET5xE8InbsJJO05mI2bOlsvo9MVukyMI9piYZdicYRIZ8LquxaViS0HZISFrTSgnIswkfufhI9Hfu7ubStvRrtSx3HCCUXNrtWducEpmLnLz1Vh/SpMhKcxAHZ0kfg+ywUwNbinCe5BTVsCaYw8ALNx93SgTkpTFQCI51IwNMibJ0LHtwn2Tw0Cs8e2uwsLqj3Ly0PHzXTatL1LmTbGpk+ZbBkW2df0QcQ25+WqcFrkxkWvL+Xsmn5fqaT5/J1n0O1OiEL+hfLUav4gUd5ZiqtCyLc7fSTYa/iHnQksoDD1+1Rm0m70Lvx2KQoZBwe7I2FTUm2Z69tOcgvf10b8XsPzILawvKJYsdh6NScrAxoJsiRVHbosOyYtLzTI50QGgH1Cxhm7v7N3kTFyJTcPOS7HYXpAF68iMxFMSAdANZ/R/Azfj07VZLGJBWjm9zJaS+hyycvMRPnGT9kZWqdpe1vh211W9z+NcTLJiWSWLIq6jztQt2pIL1sgz+BB1j0VKTNigm0H4zp+nsXif+DHpYUYO/jx6G5PXnZO8UH1pie3Dml3d+TspVg2XV8pciRIawi6r2zFjbqZmOUxlEwo39JvP3ccAEx1/9ggkiuk5N8Km+tJKOnRdfPi2nI/ixwjrgz+6DOsmD/rhIFp/tgOAZqTF2ytPIbkgcygzJx8RVx4YdRiYCnybKpdiqzidjoq/jkUjLiVT71pigkGmp0oFdPrCsmtb3RIlX26LNKrJa00tP3vt61KdsquOx6DlpzvQfKb4RDtLZEwKpyTDdq48Gm3TBH5iDNdnaamZpxWczMZQrk6wXiBcT49YdgzHoxJFszKHi1x/yf3czsQkmww85eer0XPuXry5/CQ+WHNG7/OKTcmUNamSEHi39pQSYxCU62qQbJKvPW9pEnAmrT0nWtLBEpbOZD/OYGInqY/f2stDobPv2UWHzCwJeHsqE65S+rdHFgYS9+7di8OHrT9Rzp07FzNmzLD69VT0yZnZWJfhgcwerK2LCMCimcjEan8Jgnykqwwokfln7j2ul5hNd+3JO6LDxe1BrdbUsJT/Avu044lv9xsFK+ZI1MwSs66gF2/6PxeMMj+lJpjQtaJgePvl+6mYsu68yWWH/HwEY1acQvjETUaz7spxLS4V4RM34bkfzZ/IpeTnq5Eqsm+/+utxbcFpa6VYWABdIDbDu2GtlJG/HtPu25lKz8ZpISHDxNKh0raITckUzaAx7Bj533fyJsCRQ6oujyU8DW5m9Yb/KRDxNMzaSDGYLKV7vVDtvyeuPYcVR26jwxzbOqOKuuUmhhbb2wKJWmW6s2XGpWRi7n9XUGPyZkWCiVLkDll21L2MUJ/LFbzxx0mjx2ZsuKit3WrK7II6WYa/RWukZeUiIS0Ldx4+Ei078Pk2zbYMrwHETP/nPH4/FGX0eFxqpjYgqZQ3/ig8lx68nqCtNydl7ckYizukZhlcf+kGDq/FpeGUFZ3ZzrxvFwsIObo93+3WPz5FxqZiz5UHFp2qDLPizYmzsIyFri3n7mHHxVjFOiCeFbm2FA6TeyIf4JlFh9Dy0x16z0ttW+6hOy9fLTqRiCBbJ6C26ngMXllWeAwa8N0Bs8O/dSc/s7YGX1S86e+0dkEm9rk7yWg+4z+sPHpbewxQq9UmJ5OU8rNEp6g5QlA2PTsXT363H73mRegFf1UqFQ5ci8cnG+x3/arUhESc10N5FtVI7Nq1KypUqIA7d4wzxNq0aYOEhARcvy7dc/jll18iLi4O06dPt7ylRC7IkiLxuvWMDA1uU1XWLI3mtDVzcWkpJYY0yWVuGI6ho1GW16uRS86sbI626ng0rsamoVKILz76X0MA+kMCtlsxzPbCXeuC1A9Ss7STN/wqciNlyNqLLWuHJNSbthWb3+6E+hUCJS8cdC9K5Q4JtLfE9Gwclhgep7THCo4Vl2f20T7WYc4uvdnLBcIw1XvJj/DVNtOTHNlbvlo8UAzYZ7iP4ZB3ew3FLerSs3Lh7+2Ba3FpVk1ooTRhX3iUk6cXdIlPt62UiSlS+6WhcTplRkoySzopv9lxBRkKZBw3+mgbSvl54qFOoE93opac3HxEJ2bglWWmM9Iv3ksRLX9hjwzQHRKzrpvyx2Hxetum2re0YFIZgW72jrXlNpyZ/zP6j5PYeTkOo7vWxEttq2nrQTrS5nPGWWSJadna0jy6gc3fDkVhwc5rRsubq62pZHDkzeUntbMI20oF8REiYmUfMnM0dUZNTfDR5ONtsrdt6trW1EzmcpJddGeztjYwLTVMWIxhGaejJmp15uTlS3aWxVqYyANAUy6pYPea998VbefLCJ3jo0oFDP3lKPLy1dp7Ezmk9lqx1ivV/8eMROVZfLUn9SVER0cjLo4X2GQa+wLEKdXbYmnGpysZZWP2mpJe/VU/S+JoVCJO3HJsHRlDC3cXXvyM7FRDsYs9a7T+bId2lsfoRPPBPmecu/st2Idvnm+Gp5pXMnruZny6w4dj5uerMW+HxDBMnRsBc7ONK0H3ZvIlnVmFpQK3xwouXMVuihztrRUnJY+X7Gx2jv3X4rF4/0283aO2ZIagowmHnElrz+k9bi7L2xaGk5oY2nj2HjaedY2hxkXNNzuu4pUO1RVZ10ODbEHdiVpWn4jBajOTOZQUSpSPsde5/8e913E2Jhl7P+gmuYww6/WvB6Nw9GYijjv5Gk6QlZuPyevOGT3++ZbLSLciWJ6WZVnteXNsrSm3qKAUgSV1zBt/vA3Bvl5YMLiZ5DLWfDZiTM24bSlr929zNc9NMVXjvbaJmrLWBNH+PBqNxxtqJnbS3c90a9mroIK7m0p2Rv6JW4loWa20RddrSh1HGEZUHmskkkMxrVgcPxV5BXedadAP9qsjY6nOX+zGgWvxNn9m5iZhMUUIbJr7Sc/974pNhbNtITWz93E7ZrNKqTF5s8O3KYecmyup0gfOInXBam1BbrKNkLG667J9Jh+yVExSBj6RqMf8MEP5+r5ExVF8WhY+XHPWpixL3dEI1+Ksv94wtPHsPdxOzMB5mfWdbZmcT2m6QcRT0Zrz740HaYoFypxtzhb5pYAEOXlqxKdlYcjPxnWFlchEtpccuTMzGlhkQ91XoXySI2Tn5WsnLtIduaOb9ahSwaKSIQ9ShXOwRIewyGNK1VoVm1iUbFOiA4mrV69G165dERISAn9/fzRt2hRffPEFcnKsuzk/ceIEnn32WZQvXx4+Pj6oXr06xo4dy0xNMuuBjTNHFwfOrlNXlOTlq/HiYtsnclBiOL25wt4Ldl7FGScNFZ/73xX0EZloxkNk+K4zOWryheLuig2BcbKdEvV8ldDxc+mhgMeiXCMriSznqHrNpPHkdwdsL2+jc2rrOdf6SbakGE6GI+XhI8d3IEjN7q7rwLUE3Et+5JRh12S7JgrOxm4rc8FJa64yY5IeYdVx4wxt3VEhKohnL/9x+BbqiswjINR6FIvppWTm6NWgFCg1wSeTmZTn/EI2TjJu3DjMnz8fHh4e6N69OwICArBr1y58+OGH2LBhA7Zv3w5fX/lDB9esWYPBgwcjNzcXrVu3RvXq1XH8+HF89913WL16Nfbv349atWrZ8R0VDawvJW7lUcfVInRVhsONyL7k9uSbMmntOfx30fnDXU0RsqUyc/KRmZOHn/bekJzp1VnsOQFESZKq8BAvIiKynit0kqVn5yFdRgkWZ/n14C2nTlhFRYepjntrMkGtlaRzvyZW0uPVZce0pQWkiIX0/vftfpuGfZujUBUx0lEiMxLXr1+P+fPnIyAgAEeOHMG2bdvw999/4+rVq2jcuDH279+PadOmyV7f3bt3MWzYMOTm5uLHH3/E0aNH8ddff+HKlSt46aWXEBsbiyFDhpT4Ip/scRMX5eBabUSA/J58U1YevY34tKIzVPCTDRdcLogIuMbNFhERkZJcrSyGK1oUcZ2dYCTLF1utDxY64jJzxNKjeJiRrRdEzMzJE51BXXeCSGGiMnsGEQGWv7GHEhlInDVrFgBg4sSJaNGihfbxsmXL4vvvvwcAfPfdd0hOlpex88033yAjIwM9e/bE66+/rn3c3d0dP/zwA4KDg3Hs2DFs3+46KdDOcOSmY2YjLWq6frXH2U0gKhFcNfOXCYlERFTcsUYpkfXEZoqXS+2AqUZ2Rz7AO3+e1nus3rStqDt1q8nXzdwoXtNYaalZHPmmtBIXSLxz5w6OHdNMWz5kyBCj5zt27IgqVaogKysLmzfLK46/bt06yfUFBATgySefBACsXbvW2mYXC2J1D4iISrr/LrrGJBVERET2MvLX485uAlGRlZRufSDMUR3WEVcemF1m4e5rev9ffuS2vZqjZ/M51y7FVBRZXCMxOzsb+/btMxqmm52t6WUSe85wGWc6deoUAKB06dKoXr266DKtWrVCdHQ0Tp06hcGDB5tcX2pqKq5du6Z9ndT6fv/9d+22S6odl1gfkYiIiIiopLmdaN+hi0TFmS0Zif84cLZnc77cFunsJpBCLA4kJiUloWvXrpLPm3pOrVY7fcacmzc1s75VrVpVcpkqVaroLWtKVFSU9t9S65SzvqysLGRlFc7cm5LiGrMfEhEREREREVHRk55tXKeQyFYWD21Wq9VW/3EFqamaGTz9/f0llwkICAAgL5gnrM/UOuWsb/bs2QgODtb+EYKPRERERERERRnnOiAiMSduJTm7CWQFizISd+/eba92lHiTJk3C+PHjtf9PSUlhMJGIiIiIiIo8FRhJJCJjo5efcHYTyAoWBRK7dOlir3Y4TGBgIAAgPT1dcpm0tDQAQFBQkOz1CesMDg62an3e3t7w9vY2uz0iIiIiIiIioqIuNiXL/ELkckrcrM3h4eEAgOjoaMllhOeEZU2pVq2a9t+3b4vPOmTJ+oiIiIiIiIiIiFxRiQskNm/eHACQkJAgOfnJ8ePHAQAtWrQwu76goCDUqlVL73W2rI+IiIiIiKg4YY1EIqLio8QFEitXrozWrVsDAFasWGH0/P79+xEdHQ1vb2/069dP1joHDhwoub60tDRs2LABAPD0009b2+xioW75QPMLERERERE5mKc7I132xE+XiKj4KHGBRACYPHkyAGDOnDk4efKk9vGEhASMHj0aADBmzBi9eofr1q1DvXr10KNHD6P1jRs3Dn5+ftixYwd+/vln7eN5eXkYPXo0Hj58iNatW6NXr172ektFgrdnidzdiIiIiMjFHZ3c09lNoGLuzPSSfS9IRMVHiYzsPPXUU3j77beRlpaGtm3bom/fvnjmmWdQq1YtnDt3Dh06dMDMmTP1XpOcnIzIyEhcv37daH0VK1bEsmXL4O7ujtdffx1t27bFCy+8gDp16uD3339H+fLlsWLFCqiY009EREREJcyAZhWd3QSzQvy9nN2EYo33QYCqRN55Ezlf0yqlnN2EYqfEHs7mz5+Pv/76C+3atcPBgwexefNmVK5cGXPmzMGuXbvg6+tr0fqeffZZHDlyBE8//TRu3LiBdevWIS8vD2+99RbOnDmjraNIRETO5eVRYk99RER206dhmORz7WuWkb2eqDn9lWiOpGHtqqFLnXKImtMflUr5ol/jML1tujk53lW9rL9zG2Andx4+cnYTnM7NhYKpy0a01vv/h33qOaklRPZ3Jvqhs5tQ7JTou6nnnnsOERERSE5ORkZGBs6dO4cPP/wQXl7GPZLDhw+HWq1GVFSU5PpatmyJv//+G3FxccjKykJUVBS+++47lC9f3o7voujwci/Ru5tFmlYphW51yzm7GS6rfoUgZzcBC4e0QFiQj7Obobiqpf3wfu+6spe/+llfRH7ax44tMu2x6qX1/h/5aR/UMHMT5uHsu0Q7qVbGz+TzrcNDHNSSomvuc02x5Z1Ozm4GUZG06OWWoo/XKR+A8i50vvxkQCP8+kobAMCBid3x/YuF7b7yaV9c/UxTI93L3Q3XZxXWS7f3uU44Nfl6utt1O+R4T7eohKuf9XWZOpELh7RA17qheo+VYUYuEVmAkR1ymAWDmxs9tntCV8c3pAj46/W2mPpEA2c3w2W5woVY/yYV4CNR99PRAcaRHasrtq6qpf0s6jH3dHeDt4c7XuukXBss4W4QFPT2cMe/YzsCAB5vUB7bxnXWPnf+k97YM6Er/n6zvdM7NsoHeSu+TnPf27znm9m0fnMB2qJuUt96eLpFZeTlq53dFJdWXLOl7Klp5WC83Laas5thkWHtlGnv/BeaYfWo9maHtRpmMzprP/PycNOeV1Qq/XOMt4d9A3wfP9kQANCkcrCZJZ2re71Q8ws50Tc2nussIXZeNOzg3P5uZ3w+qAk83d1cJiPRz6toBquLaT8wWcHfwn347zfb2aklJRcDieQwFUvpDxfvUKsMb0gk+Hi6u+RNe9Sc/jg0qTuOT3VeQfKx3WvBzUWOXL1FhnF9MagJVrz2GACgSx3HZJUqGXSe/0IziF3n+nu5o1LBb1jsQm5K/wY4PMl4Mip7MwwkAkCAtwf2vt8NC4e0QN2wQAR6e2gfDy/rj/oVgvB65xqObioAoGyAN/a+3w0vPaZ8UEHs+vr1zjUQ6KN5//5eHiZff3hSD8wY0FDy+RrlXO+YpCRhv89XiwcShc+xJNv/YTdM/5/+8aZyiC8+faqRk1pU6K/X2zq7CZIaVQrGE00qWPXaH15sgVc6aDpq/n6zvcllf5LICLTGJwMaYcs7nYyGO7aqZllm84BmlRDs52k2AGDYMfdbQcYgAJyc9rhF21SK2LlQeOzoFOXPd7VDAxE1p7/VQ0xXvPaYxd+PNTrVLqvIega3qarIegD9kSpPNa+k2HrNqRSif2/TtkZpo/2mTvlAeBZ0XrpIHBE+IlmvYtdTrmbT2xwxQJr70a06iQJytKxW2vxCZBEXuR2nkqRf4zCcnv44lg5vY35hOxIuzAUhfp7af5cLVD5byFCHWqbrBblqUeoKwb4oG2D/z0eKCkC7GvJrLSnts4GNtLWU3u5RW++5sd1r4bnWVVC2YP/RDUc816qyTds1F1ge2bE6lo1ojUBvD6Oi9gHe8gMgZQK8RQNSuu/lxmzx+lVhwT5YOrw1/n6zncMyFKV+J1XL+GlrIcr5KT3dohK+eb4ZIj/tgyqlLauRa4njU3uiahk/uBlcsF/7rK/sdZz/pLf4EyLvs0udcmgTLu/iSaUCyvhL/7ZHdHDMdyrXiA7hdllv5RDxIeLPtqxi9NgPL7ZQfPtPN6+E0iJDzI5M7mFxD7zSdGOs43rWRtSc/tj/YXfUDQt0XqMKPKZzXhAbGvrz0Fbaf//zVgeHtAkAfny5Jab0r4821aV/hytfa4tBLcTPET5e7nijaw082bQiGleSzlR7o0tN9GoYZnRtAxhnkK2X+f7rVwjCG10KO12ebVkZnw1sLOu1hppUKgUA6CGRzWYYvq9S2g/Hp/bE5H719K7PdFkygcs/b3WwuNi+kCW5bnR7zH2uqd5zoYHiIw9eaG18nJBLOPdYO+lL+5plEewr/llZY2z3wvruut+bUlenSg4M2PJOJ/z6ShuTnWGCL55poth2De8Xnm1ZRe+ayzDg7CoZiW1r6B+PRnWugSeaWtfZYS+6+5/AFUobUdGj5MgtKsRAIjlU1Jz+WDikBUr5eYlOeKB7b21J8MPQhjEdTT7fuFIwmlUtpffYp08VXhwfm2J7xl0Zfy+sfE06Q8LXU/z9ffVsU9HHrSE2nNycP159DDNtyC6pUz7A6DFLZ2s09907qiD0jyLZHUN0etD9vT30CrSbujx0tzGN8rdX24gOh61aWhP0mPpEA3StG4pzn/Q2yoS0dKim2HWuVJaWoW71QtGyWmlM6W+/ofm9G5bXBuKFY8b8F5pJBgdWv9Ee08xkbfasXx5PNa8Ebw93PN9KcyNo+Dlam1Ek5iWDYY6WZAJI/T4suUHZMb6z0Q29WPDKnvo1Fp+YQe6xX6kb5s4F37NfQcZmaX8vo8kedozvgg/6GNcOrV1eP4A25+nGNmfnVSvjj7IBxt9F+SAfzHXgkD0pbcJLo0310nihdeGx0FGJLOFl/DCpb+Hx3zAzSggWPtVc/5xTOzQAjzcorFctFlAa3j5cuYYWOP9Jb/RuGAY/Lw+oVCptVrehdjXL4NWCG53+jQ2OM2pN0GrB4Obw8nDDnKfFA3nvFHRsiR0Gvja4rmhmJqC2/8Nu2n8bdtYYBo3f6FJTcj26v6NgP09EzemPJcNbi7bRUySqVDbAG693rqltQ2l/L9yc3U8bBPl8UBNtcH3FyMdMvqemVUrhr9fbYtd7XUwuJzj3cS/t9VjzqiF4WiTQK2S7T+6n2Sc71S6LOYMsC1L1LzivBPp4SHZiWEKJOFXfRppjc4MKQdrj7JLhhRNzKNXRXamUn+hx1VLCMPwudcphaLtwAJqODqn9vIGCwShhXfUKfhcqFTD76cJ9wDDgLPejs3cGvOF3OKlffcWH7htesxqWlOlsZtSOqc4XR7s4Q6IDF0ALg/tJSzR18TIGrsKa61Pda/hJ/eor2RwqIPvOtkaNGor8qVlT+oKDSgZTFyC68Q7djL1/x3TQXqgZMjy4R83pj8ZmDswbxnaEl7t+O4QLJ1v0aRiGf97qgD0TuuK/8V3QrmYZyUyy3Px80ccNa6vYoqXIEBfDuij9DYIjbm7Ay22raT5HE9kPhhcAjSsFY2z3Wlg3uoPRSdVwWPvEvvoZBrqfUWl/L2wcayIQrFLBw90N819ohlEiw1MrhyiXTSY2dFls/xWG50jt2290qQm1zECcmK+fbYrKIX7YM6EbvnymCQ5N6q597pfhrY2WNwwofS8za0oYEvW/phWNAsLv965ntyE5lgTRfnulDb4d3EIbiH+/d10MbVcNTzSpKJltUjcsUHuTLlAb5L+I7euGNc2UzCQwzCyTe2NmKltabA2BPh6Y0r8+BjSriGBfT/z2ShtsGNMRtUID0aFWYRCmX+Mw0Rt5XTbswkZWvPaY3gQHuno1kDdBmVKlDJcNb43ZTzc2mUlUKzRA7yYotOB7MPzavDzcjILE1pD6rMWOSZaY2t+2i+mwYB/4e3tg1ah2CAvWvUGW/9vY9Lbpjj4p4WX8sOf9bhjVpSbmPS/e4fZ4g/K4PLOPXscgYJztJuax6qVNZo5bU27EMCju4S79OTWoGISoOf2NfuOGx6oX2lQVPab7FhxTDLcwY0BDhPh7GQXQ1o2WHiZtLqD175jCTpuX2uoPTx3cRvM7mmLixu2PV42DfuaOP5Gf9sHRyT2gUqnweAPN78DDTaU9dravJT3cVrgm8fF0R41yxp2dYgJ9POEh0ibdz/ejgpqG5YN8cHFGb73h2JYyfP9yj4MAsHhoK72MW1NGdqwueu2ke22oe/zZOq6TUW0xseuZzwZa3oGihhqjuxpnncn1dvdaWDWqHSaLHNfG9awj2rG9bnR7NBI5388c0NCq+tZD24Vj9tON8VxBB6SbSmVy0hKp6wjD0jAf2LHDXE7gS4lhzp882Uiv86RVwYRvHm4qfPVsU5PHCMB4+LVQ+1K3M8lR/EyUhln5elurJml78bGqeLkg8F0UXJAaDWMlS+7X6lcwP+phqc790OiuNbF4WOExsSgM2y+KZAcSo6KiFPtDJMeoLjW1BaebVC6F8DLiF/GmDu5iVr+huSASTuZd6pRDk8rBigRJPNxVaFqlFMLL+pvsPQny8cB7j4v3wuq+zpZeLgCoVMoXUXP6Y1TnGtohuQ0rFvbEtqoWYjSDrUrnMvmHl1qIZsYAwC/DWuFPnZpU/47pgPd61YW/t4fRhZLhRevIjtVxanov7f//0bkp6dMoDOEmbtiENQ9oVkm0h2mohQXidT8Pa03uVw/1KwTh2YKbUKGNarUaUXP6Y2LfelYHYY5O7oFBLTXr9fVyx7OtCoMdfRuFoVao8YWy4b6syRI0Xzfpt1c1N0EVgn2x/V39m07DQJy1TonUurKk6HfnOuXg5eGmfY/lAr0xY0Ajmy4Sto7rhCqlC2+chRtTw6/MFUYkmWqCYft+Gd4KTSqXQo1yAZj/QnO4uanQuU45bUeLs66r9n/YDe1rSt/0y/4uFYhstqhaCm5uKgxuU1U0aABoslUBwM1Nhag5/RE1pz861RbPpFBqHzHVsXV8ak/RG2RzGlYMEq2LJcemtzvi5ux+ksEeOV9ZjbL+WDKsFRpWFO+gMpfxqxs4blNd09EoVofWx9Md7m4qfPJk4RBH4TsUc3N2P/zwYgv0aRSGT59qjN8LjoP1wgKxZFgrfCtk9ivw3QodP1LDewHjfSjA23zmrW7HnuHrhSH5cgNopgjrblK5lNFzQqfkvqvxAICjUYmS6xE7H5k7x3h7uGt/o690CMeVT/vCw93N5NdSWN5CuYOd7rqealYR055ogCeaVNRmnep6q5v85AnD4Nyil1rqXWMJDLOlAaBng/I6Gbem3+vUJxqIXjtdnNFHmx2lG7yuEOwrq7bYi1bU/RXe8uWZfazuzG9TvbRkJp1K5LNoXlX8WujlduGiNRvNZc15ebhhcJuqqF1wTK5Wxs/keUDqWGn4mppl/Y2+6xs6M4hba9oTDbB2tHRpg63jOhllMFurT6Mw7NTpwBDOPytfb4tnWlaGv7f0+ej93nXhY/C9CkkJrRVKuBjVuYbF9wxivD3cUb9CkOhv05TPBjaGt8joPFdl6tpM7FhlqGppP21JmqUjWmOPBROuiv2Wxe51BB/0qWe2c4psJzsCs3TpUnu2g8hIi6oh6NuoAs7GJAOQl31iWMNG8M3zzbD0YBTORD9EzYKLaeGA2KpaCMYa1Lqzltz72rMfi/fqdKtbDv46GQzznm+GLl/usblduheNwsHYx9MNb3atif8uxuot6+VReGSuHOKH41MfR50pW5Cdp59B6eHuhrY6Nal0L6AN67+V8vPCmem90HTGds3zBkf/QB9PzH+hGd758zRKmRmu2KWu6Yu61zvXxLD24ag7davRc6O61MCPETf0HpvQqy7WnIjBpnP3RNf33uN18PV/V0xuM9DHU6830s/LA3XLB2Jcz8L9yjCrRK5QkR5yc/uZ2E2T0NP+Sofq+OXATTSsGIQLd1P0ljEXlO/bKAw/77tppsWmeVp40dS0SimciX4IQJPJKlBp/7b9BtFwHcPahyMpPVsvSPHe43Vw/UGa3nLPtqyMuNQsRFx5AEBzI73prPh+ZMjD3Q0f9qmHz7detqitIX7SHRSlCp7rUqccOtcph+71TGe0iO0nug9VL+uPm/HpFrVPzGudqqNMgDfmbNG8V3PZTuVlZoUMaF4JC3Zds6ltpvb5Md1qITsvH5NFbrqlbhSV2B/VUOPdx+tgRIfqaD7zP6PnywZ4IzTQB1di00ReLa5yiC82vd0Jq45HW9WmisG+JoMxcgI1f41qZzKj1txkQLrHUKGTbN/VB5LLD2sfjo/+vQAA+KC3puNu1ah2RqUeVCoV+hYMJ/byUGmHIjarUgo9CgKQY1eeAqDJpB3y8xGT7TSlZrkAXJ/VDyoAvb7Zi2txxt+h7vlxybBWosP7TM06b/hdWFJV48DE7kjLzJX/AhFCx5CpLHzDgHaDCkGinWJSVCqV3rWKYNWodnjux0Pa/wuXItXK2D5kWFAvLFB77vRwdxMNgA55rCqS0rPxfu96WLj7usn1aTseDR53c1PhseqlMbV/fXy66ZICLYfk0HpBm+qlcSYmGf9rWhHXH6SjXU159ahDbawr7uPpLtmRY4vtF+/LWk74ybzYtirO3UnGt4Obo/50zTXkL8NaodaULWbX0al2ORyZ3MPs+UvqWGlUqkNkMcNrayle7m5G1+xSdK+xAKBeWBDqhQXhw7/Pynq9Obpv94M+ddG8Sim0lAjmGqpt0GGm7UC2sk67u5tKe/wf2q4aPuxTD25uKvx26JZV61OCK3RQy+WmUqF+hSBcupdi9FwpiTq21cr44VZCBgBNss34x+vAy90NnWqVteg3L3xOP7zYAgEFw/5ddS6BkkR2IHHYsGH2bAcRAE1h3W93XdNmwb3RpYZ20gZhGJWvpzse5eShe71Q7LocB0BzEdq/SQXRGjaAZla1IOHAU/BYtYIMx0Yy61N88mRD7U2JlF4NRW7cCzaoewKTq5pEFqYSzn3cG57ubmgVXhq1QgO0F6piF0Fy6+MJxK51gnVOMmLH/gHNNLPsGQ7dG9W5Bsb2qA3fgiwTOdwlTi4T+9QzCiRWDvHFwhdbYNPETaKvGdujNsb2qI1wiedFt++mwrZ39WcTEz7C51tVwV9W3swbkjqH6tb/0Q7VK1j2hTZVcORmAn58uSXe+OMEzt9JwcaxHVEh2HzwZlLf+hhfkEnbs3553EowHWSqWz4QkbGp2v9veruj7Pp3Ad4e2klFhM9ebDICawO0PeqX197gGe5WAd4eRjNh+4pkTVYv549ZTzdGbRk3GWLe7FoTl+6lmPwcS/l5YtbAxhi9/CRC/Dzxq4mhcz6e7hb1iOvuP2IBsJ71Q/UCx9Z+1lP6N8Av++UFoHdP6Ip1J2NkLVtTgQwrUzVpJ/Q2X7tLrdYElz5YcxYxSY8UuSlwU2mGa4b4e+HAxO7oMGeXNjvfWsKN/sDmlZCUno3ZW+QHsOXsU3IOzeY+G3PPiw0J7GAisxXQzDR+40G69gZcTs0tqXYMbFZJNJM2xM8TSRk5ZtcrEM5j/47pgAbTtxlvX+ffPSQyKbvXC8WMAQ3xvyYVjYLNhs23pCSDuUCTHI9VL4MrsWkY3l5+FrtNucU6b69N9dJoWS0EJ24lAdC89zVvtJPMgrXG8pGPISbpkcllZpmYkKZjrbLYfy3e6HGxyyyVSoWRnWqYDCSGGwRJde/N3+5eS9vZMq5nbb1yHaO61EBenhqLRY7NYUE+2DFeXi1JAHod4JbQDTbbIyQg95Jb2HbZAG+94ZAALAp2yO0EM9S7YXmj4Lrwu32semkcuSmd3SumV8Py2GjQqfl+77r4cluk0ee8elQ75IgEHSf3q48ZGy+a3M6Lj1XF8iO3TS6je/wRgpQC051Txh0OwtK6I0gAYO/73dD5y90m22HohdZVZQdm7Unq+Cx0+rsSdzcVtrzTCQMWHtALPgPSnagR73dDxJUHGPbLUbQJL41AH0+95JYmlYO1CUOC67P6YfjSo2hbowy+3BYJANqs49L+XnoTq5FzMeeTnE43a+u9XnVx7uNe2PtBNwCak4xwEm9WpRR2jO+iHf4wpX997Q3O5nc64a1u4nVWXutUHS10er+EY3at0ACc/bgXutWVHmIEaIaWnv+kN4a2q4a/Xm8rORFJ1Jz+2mCYGDmnK0uHaVuje33N+xVOXsG+nni1Y3WMLag1I5Yt9OWzTSwqemzuxkW4ePj1lTaYrVM4fkCzSkYXDpP61UeAt4dFQ1ellhW7aJG6kJAzFNgSpQuC43JnBK5Wxk+ylz+ooOe6h0TGmXBjUS8sEHMKin5XK7jwKuPvhU1vd0LlED/tib9uWCDKmOjhfbq5Zr92c1NpA2qLh7XCf2ZuNHRrNv0yXHpIo2EtQkA8MKGbzaT9Kq28A21RNQTPFAwZF5v4SaBbU85wv1ZBBU93N+x9v+B4ZUU7Fgxujn9EJodqVElzsT1rYGNtWYd+jStoO1SUmDldyGqqUdYfHz2pCZzq3hAqWRNS+JrEJooI0ikqX72sv0O76MNkBNDFCPunWq1G+5plsbxgogclisPr7vtChrbhMc3cR1TXYBIY4Zjh6e6GUTqTY9RTaLZlJTIxzVkqVg/WTWUyO2/rO51xeWYfi7YjFtCJmtPfaORC7YIMOqmbYVNBf0D6fC9n93dzU2Fou3BtzUW97D+D10t1rCkl0Eezbwmzwwv7aoCdJ4sQGJ6rf3y5pd6M7q3CS4t2BFmrlJ+XaI09uRpWDNIc5wz0M5xkx4RBLSpr9xPDgN/MAZpr1GdaVsb4XnW19eTqltc/z0/qW9+ow0w45lt6WrV2Fxuo0zloLvtRbPIRc+00Vedbl9S5zsdT/9jSsloIriswvNjQlwUdWjXKFe4XQpv+fL2tXskjqfqwuj57yvg8axh8E3h5uIkGgl+RUc7mza7iQ/fb1yyDE1M1k1aa2jXEntOtjy9XVZkZx5bUKt82rrPkc2JD4K0ldWujRNkla60aVVgT9dS0x7XnOqGtzUXqkbu76d/P6xJqhortZytEJiV1d1Ph91cf09ZtrxUaoPfbEDg/DEwMJJLTjetZR69HNdDHU/ICu1ZoQGH9Mpnngyn9G+jdhOne8AT5mK89FBrkgwBvTe2bx2qUEQ16mGJ4oHtcpIB2hWAffPpUI6tm+xRbH2DcSy0Y1bkGLs3oo/+ZqFR4r1ddyZvggc0r651YdL33eB3tbJGC7ibqP+nqUqecoidkgSXp7mIXkJ1ql8Xfb0oXorfGuz3rYNFLLbQFuc3Z/V5Xo+LbggBvD9yc3U9bO9GQ8P493FXaQOm7j9fB32+2Fw0Ymvu0OtUxnfUjRffiztQwW92i4l88owl86n6Hz7asjHd71tF7jfA7tiWTZVLfehjXs7Z25msxuoGml9pVg69OoFtootyhg0LAUY6NYzshak5/9GtcAZULgs+6HRVKxAaeblEZ7/eui+3vdtYO5+ysOwuuwTaUmGzFz+BC8uDE7oh4vxuWjWit7aQRe2vmgjLW6Flf3nFKTJ2CQJ1Qg6daGU09qwrBmu/Kkrqfhkr7iwTMZRIC0K2r6wdXxFZTMdgHq94QP65bKrRgds55zzfFgYndMba7pmOvR71Q0c954RDjyULMvVdravx5ebhZXRdSLqlmWzuruHDs+/h/pmeaB6A9Holdl+jOIquUMd2MbxSDfT1xevrj2lpjwvZsmWDMEt+/2AK7dWptlQ3wxkf/09THdNUbTcPP5tKMPrKu/4TO86+fa4obs/rh/Ce9jTLmQoN8EDWnvzbbekCzSqgXFihZl9bPy13bWSh8YPkWjp4Z/7j++fk9g/9L0c2AfaF1FZwUqaEs2Cky27bcALG5jjex38jZj3vh+FT99vz9Znu7TNog3ItsGltYIkdok0ql0rtOHdi8Mp43cx0ZLDHMVHe9cqx47TF4mpggSupaO8TPS3utaennJcy6LacmniWeNbheNgwS6/rr9bZ6s9MbTgD5To/aFtdDlCb9pua/0EyhbVimdXgI+jQMw/u96yLE3wsT+9ZDo0pB2u97isjkRm4qFcb1rKMXXBbKAzWsGIT3e9fFuyLHBTmjlHpIXKvp1vh/W6JE2ZJhrfDdkOZmt0HWcUx3IZEZq0a1E60VJMbai1Tt4jJPRKtGtZMsBDxzQEM0qxKC/3233+x6DCduEJtZLzTQW3Kmz1+Gt0Kd8oHo+Ll42v7PQ1sZDbt9v3ddyYCVSqVStHderL7kk00rKlbTx1ql/b2QmJ6t/b9wY/9080pYe+qO9nFHjWzw8XRHn0aajAM5NY/MDbkwFSwVXqp7Iebp7maUuaENhEmsy8vDDdm5+XisuvXDCMY/XkdW1tOK1x7DX8ei8WzLyvhgzVk015lo6EuR4aeFxwGrm4YyAd4Y19P0DY/u+ltUDcGlmX2w+dw9jF5+Uvvplg/yQdMqpfBWt1pGw4l0mes1b1G1FE7efmj0eJCPp6yL1lYWZtF6ebgZZXLr7ldKZpkJvcmGs/QJxdO76mSGC01oXrUUThV8HkKP9ouPVcWV2FTtZA+73uuCeTuuYsOZuxa3afEw4ww3uZ5vXQXlAr0lJ4b67ZU2eGbRIdHnTPnhxRbopVPewZLvoHKIL+a/0Bw9vo5A+UD9TMtQg//vfb8bgnw9ZHWmyVE+yAdnpvfS3sAObx+Ob3dd05z3RH6j/ZtUQPd6fXA25iGe/+kwACC8jD+qlfHDTy+3grubCquPRyPiygNcvp9qvAId1g65lyLUepKbEaJSqXDts75GddTklItYNaodzsY81Hvsra61kJevljX7t4e7m9GxQdhnpj3RAPUrBFnUsSbl8KQeKO3vpZe97evprr1RLGWidqsctgQd/bw8UL1sEbqdMfg65F6TGe5PKpVK1k14WLAPtprIrro4ozBj19pj/hNNKmr/7emuwtgetc3WljakUqn0stN1VQj20TuGzRzQECqVSjvBnRTh2PDZwEYY9fsJ6W2LvG9zx8ZAHw+ck6h3DgCB3h6Y8VRDyefF6O4Hui36eWhL2fWXBdXK+CEpPRspBTVPrfmNta9ZFrve64pOX2juP5YMa4XI2FR8sTVS9jrMDV92JOETWDqitcmOKd02l/Y3zkAWOphHda5hc2BZ9zNoUCEIbaqXxrKDUfD2dMMTTSpi7ck72jrc/7zVAQMWHhBdz4BmFdG8Sil8vMH0cHRzSvl5QqVSYdHLLbWP9ahfXq/MhtgkJkJm+pA2VbH3ygN8/2IL7VB/lUolOWpQLrFPOTTIB18MaoK+jcO02zckVR6ElKHImffMmTNYuHAh9u/fj5iYGKSnS9d7UqlUyM21rZAzFT+hQT6ik0qIsXFEo+zAkakhai8X9JhZwtRJ3NQNrVQm1/cvttBeWB6b0hOtP9uhfa5ltRCTRe3tTYmZsppWDrZoqI+hf8d0wPUH6ehSpxzSs3K1J+uZTzUyCCTK2yEWvdQCHpZUrTfB3gWCLV2/1OJVS/vhWlyaTfuSVC+hofY1y2ozJna918XskFNtIFHhIIL09go/JOGnLDzk6e6Gf96SngUx0McDm9/uJPm8YO3oDrJrcYodT8bYeKFmSHh/VUr76mViApohTWo1sCjC9EQCgm51Q7Hvg26Sw6t0Pd6gPL7ZcRWtqoVoA4m6Vr9RmC1co1wAOtUua1Ug0RaBPp4mS1lIdQQsG9Eaw5ceE32uamk/7aQfAksPFTXLBeDXV9qgQ80y2hv5n4e2MhouJmcoWKfaZfHdYOPMQSn6dXALGz6mey1EJ2UYTaTl6+WuV+soLNgHETpZu5P61cekfvUtqk+rhPJBPjgwsTsqShyDgn09kfxIUxPR19Md7/SsbZQVNrRdNVn10tpUL210rRHs54lpT5jPRpSizWQC9IZE2kLseHxJYsi4EtnixcnMpxph2vrzks8HmJi5VnBkcg9FO4ClFJ5XrWdL557UtYvhUHC51+BCW8yX25G1Oouc+0Q6yCiHbpuaVw2RnG1al7ubCrsKMjd3v9cVagCNPtqGRzl52tIyrWTMwK1L95zdo355dK8Xqg0keuqc51aMfAxDFmsmopJbMsNU4Noe34mwP5graSW17fcer4Mt5wsn8BGb/VwuYSSX7qZWvtYWAT4eqBkagL4FiQf9m1TQBhKbigwpFsx/QZN1Z2sgUa5fhrfCK8uOA9BkCwv3CVVK+2GTjOtdOcSOB4aHl+da6yfNdKlTDp1qWzeKiixncyDxu+++w/jx45GXl+ewYQxUsmlruFi5uykZxFn9Rjs8KzPrxFRzrQnU6AbZnBk0FBPi74UqpX0RnWi6ILkpYnXjLFE5xE9b71G3LodhjQ65xZaFbEIlOOJY6aYC3ugiXr9GIASHpH4TzhoWJmcIoxDUddRpR/ezEIKXcrM3woJ8ZAXQbKV04XBhbV8Maop2NctoL2bf6VEb7z5eRzsLs5ix3WuhS51yevuW3M+gYcVgRM3pj5/33jC/MIz300+faoSpJm7cHcEw87JT7bL4+rmmRpmBuoTawGKkvtmfXm6JMgFeGPRD4XlId7ZxQLr8hRg/L3csfLEFbjxIx1PNKpocIieHWq1G86oh2P6u+YkbnmxaUfTxjWM7IkEnu9wRTE060rhSMPZfi4dKJR1MUyqAZw1bO1tt3r4C2eLFycttq+kFEvs0DMPWgmDE54May+owtXYSD0sJ+46lE+xZQ6wEjthx7r3H62Bo+3CrtlGnfCAOXk/Q1hLV5eWuycift+OKovWA5ZAafaBPuk1SHagtq4ZoJ2k0vB6oFRqgyHBc3XO6kABSys8T7WuVRcT7XZGXr5Y9UaTYx25613PM9yR1KSVMvmjKopc0ExkKDEdHAZoRC50LztO6oxqE861YqQohOFbKzxMPLZjcy1Jyg8Dd65XH/BeaIT4tW5GJ78QIGdflArzxIC1L1mvsUQaHpNkUSDxy5AjeeecdAMDo0aPRv39/9OvXD6VLl8aqVatw//597NixAytWrEBQUBAWLFiAChWUuxmnkklICnPEhY45rcNLY8Hg5rgWKz30qn6FIFyLS8PnTzdRbLZeU+qFBcouMG1P+z7o7vBMEmt4m5hooyi7Mdv8BeOCwc2x/cJ9yecdPezEEtP/1wA1yvnLGj5oC7GjjGFGojn2uAl0xOHPsIxEuYKaR+FlzQcE21QvjVbhtk08IvfzNQyEv9S2mtMDiaGBmjplwjGwXKC3ySCitVqFl0Zpfy+M7FjdZIakOYHeHkjNykUpX090qxuKbuYnrDZJKAvyVHN5bRLqEIsxN7GFsy4FTG1XahICR3B2IE8IopqqQSbo36SCxUM2izIhkCN8N8+3Vr5GtC3kTmKmxK71i8TkSYZaVAuxut7o5H718UzLykjNNB4Jd+WzvgCAeTuumD3XTOxbT68eGwAMs2JkkkDO6AOpyfZckX9BXXu5AURTWoeHIMTPE/1EOu/tdU06/4VmaFAhCI/P2wugsAayNfo0CoOflzsysvMAwCiIWLOcvzaIKGtbBT82ocPB3sf1n0TKb0mx5ZpDjgYVg/DbK23QoVZZHItKxI8RN0QnqiLnsSmQuGDBAqjVaowbNw5z587VPu7l5YXu3bsDAIYMGYK3334bvXv3xrRp03Dy5EnbWkwlXvd65bHyaLRdTrK/DG+Fm/EZFr1GKotC8MWgJhjdtSbqVwgySsG2B1O1cIqjDWM6QqUCnvjWfL1KwYgO4VCrNZk6YkW4XW3osb2UD/IxOUTIETOxWqusjPqGShCCaLpfmbbcqszv0R6FnoN9PZGQno2BzSthnc5QfSUZDlFsUDEI28Z1Rp3y9ul9Lo6Eocwj2le36vXmdjHhacPZVwFNPaXoJHnnswqlfOCdno3xvWyMIBbw9nCXnf0y/4VmNs2CW7t8IC7dS7H69ZaS87P39rD/MFQpQkeitTOS2+r1zjVQvay/diillPOf9IaPhxs2nb1nt5tjVznXNqgQhIsO3Eet5a5Q+RYl2fINenm4oVGlYBy5kWByOXMZiYajO5SbaAN4zqDOY+3QAGTm5pnM4Jf6vYhlKnp5uOFRTp5NbQSAx0yUe5LTuShG7FMvE+CNU9N7WbU+axkGxKTq7cm1+e1O6PrVHtHnLD3UFY6A0RjdtSZmmxgNYiul6iZbwlTJFyHo2rZGGUV/d6QMmwKJBw4cgEql0mYlCgyH7TVr1gzffvstXnjhBXz55ZeYMWOGLZulEu7xBuXtdjAxNbOstXy93FG/gryi7bbo3bA8tl2Itft2XE3jypbfgAozOpI8rnEr5josHZpuyUQEuyd01ZvQQMqcQU3w3I+H7JpRK5bZpDuTob3rU/p5FaEJFCR0rRtq0/nKXDDf1P1v0yqlTNZUMtyO4QyljmJrVsOKkY/JDpjak5xJtByhT6MKODalp9NKnni6u8karqs7UUircMsmiipq1rzZDmk6WXHD24djxkbH1DKzxOudayAvP1+vdqk9fNinnuxlm+lMvGatemGmr8EdNememC+e0Z9M7r/x5stASBG7NPn7zXY4ejPJ6nUCwNXP+koGW/9+sz1qhZruXBQy3o3Y8LlfmtEH9advtX4FMnzQp67ZeopiTF7DWXjZZDgCZlSXwkDi2tHt8fT3By1un6vRnWxPzgRS5Dps+rZiY2Ph7e2NatUKx/K7ubkhMzPTaNmBAwfC09MTa9euZSCRyA4WDmmB3HznD/cmy3WrWw67Ix+gU+2ymPZEAwRKzFzoaC6SzOESxD4Ke9x8yB22IQz1Uqk0M6raI6jnU1BD01lJKs+1qozJ686ZXa4o7qb9G1dA86qlzAaeHPEbfL93XfS3YWIrZwvx90KIA2sSSpU2GNmphksEEgHXq5tsyqFJ3UVHBijBVY4Nfl4eeh0jr3Ssjlc6WpelbE8B3h54v7f8IJ+15HSWCZToUDJV63Vq//roWMQmZ5A624s9Xis0ELVCrR+qC5ieQLFlNfOdAJve7oSrcdIloOTS/T3LmXyoell/3IzXTABrzbl0dFfrJrAzta13ehrXWGxZLQQnbokHe7UjYESOZk1EMvmnPdEAM63spOghUrfU0Q5M7O7sJpAFbDo6+/n5GQ0bCAwMREpKCrKysuDtXXhh4OnpCT8/P9y6dcuWTRIVK293r4X0bNuHHACAh7sbnDiaqlh5ws431YYZbUtHsDhwURJeUAeorg11dGyl1rm8tNcQxlc7VoeHmwptq0tkp5iIXZrLAJHDcCbc4mL7u51RJcQPK47elv0a6QmRbA+VvKXwbN9ElqgQLD2xDRU97WraN5tRaSM71XDKdmuU88eNB+mKrlNsAhtXULWMH6qWMR7+bM/yOX0bhWHe881w6V4KBhZk7fl4uiEzJ99u2zTn0KTuose75SMfwyOJe0GhDrjYdZ5wjVRVZxh817rlMHOjvPY0rBiEaU80wAs/HQYAvPu4/csFSVk+8jE0rhzslKHVZD2bAomVKlXC5cuXkZubCw8Pzapq1qyJU6dO4dixY+jYsXDW1bt37yI5ORl+fvafuZJITKmCHklX6aEGYFU9KjcVUFQSD5tWDnbIsG6l9ajvmhdjjuYq9aWcqTD7qPCzaFqlFE5M7YkydsqicRU+nu4YZWbmb121QgNwLS4N/l7uDs2IcsSM2EoSiqvLyeSQ+gXyt0kkA38mDnV4Ug/ttbYppr6WizN6o8H0bQA0s8gXVxvHdkR2rnJBrYszesPXs2h9Xpaexiw579UpHwgfT3f46wyV3Ti2I87fsX+tUt0A6WudquPnfTdNLu/j6a4dAWKoa91QbBzb0aY6wlKaVC6FtjplDJx5WdGhVtHKCiYNm7r769evj7y8PJw7Vzj0qGvXrlCr1ZgxY4Z2iHN2djbefvttAEDjxo1t2SSR1WY93RjfDm6ud1Ipik5N74XT051Ty8pS/4zpiDmDmji7GWQl4ZrCnkGLr59tan4hJ9IWujb4CMwFES980tteTQLgvBlZdRkOOQsqGJLv6Ka1qV4aeyZ0teg1rlC0u1mVUoia0x/lAr3xemfxrBgGDImoqAgL9pEMiMile14JC3LOhEGO4OflYVH9ZDnrK2nni/oVgtCtrmYyjoVDWmD3hK5YOsJ4NnBBrdBAPNXcvjMN6+rdsLwiM7PbI4gIaGYj16XESBIqWWyKqPTq1QurV6/Ghg0b0Ly5ZlbKt956CwsXLsTOnTtRuXJl1K1bF1euXEFiYiJUKhXGjBmjSMOJLBXk44n/mZlh2ZWtGtUOZQO8tLXRqOhyhSCQHI64Jh3UsrL5hYogR3VYOPO+QapGZK8Gyk9aZU64zNqSrujYlJ6Sz0lmJNqnKVTMtQ4PwanbD53dDEkB3h5IE5uUwUr8ndiHo65hlDy/DWhWEf+cvqvcCp2oqFxDmiN8vfXCApGYno1MC2eX3vJOJ+2/+zfRlCSKStCviVizXABefKwq3u5hXJvQXoQOaDeVyuxENNaoFxaoLbEDQFYWsKGNYzvq3U9WDvGFuzNnHaIiyaY7nUGDBiEmJgYVKxYGZ6pXr44VK1ZgxIgRSExMxKFDhwBoJmF5//338eKLL9rWYqISqk310s5ugkv7+812SErPcXYzZLH3bLdKKWGd22Qhsb347Me9itzwKlfG32DRMrB5JcSnZTm7GZJWjWrn7CaYdGBid+TkOa+GGTmG3OOa1EzB1pj3XDN8+Yxrj4AoaYQMygrBPtgwtqOZpa3rGHB3U+GzgY4dDSmUn1Jy/9W1dVxn7b+/G9IcbcItvz8sLsFoci6bAomlSpXCRx99ZPT4wIED0aVLF2zevBnR0dEIDg5Gr169UKsWC3oTkX20rMZAq9LsWQi7qNDWSHRuM4y44kWgGlC8UPaPL7fEhTvJiq6zqBnePhwvta3m7GaQDPOeb+bsJpjk6kMflR5x4ervtzgTaubawtZh0rrc3FTwKiYZV0WlM9oSpmaGFsj5OXu6adZjyczgSsvPd9yF4xNNrBtpp7sP9agXihEdXG8WeXJ9dht7Vbp0abz00kv2Wj0RkV14uqsQyFnD9BSPS2/raAN2LnpD6pqtUk7vhmHo3TAM5xUIJnp7uCFLweL2jqBSqfDxkw2d3QwiIosocW5a9HJLBdZCrsoe1y/ta5bBlH718XI753e+2SsjUQm6ndFLhkvXlSQypWjPOkFEpJANYzoiIT0LXevaf8bmwpmAXTOzTFBMOu+dqrS/csXUdblCRoLh7uHK+zIAeLkXvUCiOa6wH5Q02gmYin0Yn8h6pmIoz7aqImsdlUr5KtQackWWnr3kHHPd3FR4TWLyMkcRRmZ0rFXGzJLOExpkesJAIjkYSCQiAtC4sn1mRTPl+yEt0NMJE1PI5sK9qUXBwYnd4edl33qBzvyKiloIS2jv3ve7ObUdSuBPk8i88rxZdhqpoE+/xmEIcNBkZMVWUTv5muEqJQieb1UFJ24n2byeYD9PXJzR2yn1osv6mz/mnf24l+JlaKhksulI3r17d4tfo1KpsHPnTls2S0RUpBWOllXJqgvjLK5xaedc2u/KitdWLGHZFK54b/P7q23w8pKjeo9VLePnpNZQcSAESJgN6tqWjmiNhhWCnN0MMnA9Lt3ZTSAS9fkzTaBWaGiFn5dxiMURWezBMmZwZhCRlGJTIHHPnj2ylhN6GtRqtcv0OhAROYvu0OaioKi00x6Ei0pX+wxcfRixK/Byd0On2uWc3Qy74n5AJK6bA8qUkOUiY1Od3YQiz9mH/UUvtUBCerZDtlUvLBCX72v2GUdchxXnOEVYkA/G9azt7GZQMWJTIFFsxmZdycnJOHLkCA4dOoQyZcrgzTffhLu749N8iYhckatfrhTj6ynZCjMS5X0Yp6c/7tDaaazTVjLxW3ee4lYj8eW21VApxFfxGZOJqHjq06iCw7a1+o12GPfnaey8HOewbRZXwzuE44U2VZ3dDCpG7BpIFOzatQtPP/00Ll68iDVr1tiySSLZaoUGcLIIcklFZUgcfz6WK+Vnn8lVXJGwfywZ1grf7b6GiX3q2W1bNcsFoEppXwxtF27xawN9PJCamavYkCWi4sTdTYU3utR0djOoCJFzJGVHpP0YnsuK86Q0gT6eqFDKx9nNsEm3uuWwO/IBfO1cM9scD94Uk8IcUu22e/fumD9/Pl555RUsXrwYI0eOdMRmqYTbMb6Ls5tAJKpwaLNrn9Tf7FoLY1eedHYzyEUJtzJuKhXWje5g1235erlj3weW12UGgH/HdMSleynwcnfDquPRCrfMOcb2qI3opEcI9OGkBUREJdmBidadG52tpHTuLRnWGrGpmQ7L+u7VoDy2X4w1etyVa7JT0eSwPer555+Hu7s7Fi9e7KhNEhG5JGEmX28P1z6pP96gPC7P7OvyAU+7cvF6li7RLldog4jXOlcHAFQv649+jSugZ4Py+GloKye3ShktqoZgx/gu8OCNgdO4xG+PiEq0p5pVdHYTbGbuUFrU441ubipUCHZc1uiAZpUk20GkJId1Zfv4+MDf3x+XLl1y1CaJiFzSS22rwdPdDZ1ql3V2U8iMwnporqWoX1jbW9Sc/s5uAhVz/A0SkbNVK+Pv7CbYTO6hlJ038gjXrc2qlMKHfeph8M+HAbjedSwVfQ7ryr5z5w6Sk5NLTBozEZEUT3c3vNS2WsnO9COb1K8QiBcfq4q3e7jADHw8rRMRETlEcTnlWnoNXFwmuLI3IdTyv6YV0a5mGe3jvOUgpTkkI/HRo0cYPXo0AKBx48aO2CQREVGx5eHuhs8GOvd8ymtSKonYH05kHjtKHaMoH47kJhcV5ffoDMLnxV8g2ZtNgcQZM2aYfD4zMxPR0dHYtm0bEhISoFKp8NZbb9mySSIiIodRu3iNRGfixT2VZDwmUEklJwDEn4f9FJfODD8vTRiiW71Qk8vxOsw6hp8XMzpJaTYFEj/++GNZPU5qtRpubm6YOnUqhgwZYssmiYiIHKawZ5cXYJL40VAJwptZIiLb+Xq5I/LTPvDipGGKMgz0f/FME3yw5qyTWkPFmU2BxM6dO5sMJHp4eCAkJARNmzbFc889h9q1XaCWExERkUzsCSciIirEYcvOpffxF/H0RG8Pd2c3odgy/JXyZ0tKsymQuGfPHoWa4VipqamYPXs2/v77b9y+fRv+/v547LHH8N5776F79+4Wr69r166IiIiQfL58+fK4f/++LU0mIiInyC+4SOeNkwlF+z6GyCJF/L6dyGacONO5St7HL1yHObkZRURhB3jBB1bi9hdyFIdMtuJK4uLi0KlTJ1y5cgUVKlTA//73P8TGxmLLli3YsmUL5s+fj7Fjx1q17t69eyMsLMzo8eDgYFubTURU7ETN6e/sJpgl3DC58QLW5e19vxsSM7Kd3QwiohLPlqDPjvFdEJeaqVxjqFhgiRnLaH+DBX+7MxJLCrN5spWAgACMHz9e1vILFizAw4cPMX36dFs2a5PXX38dV65cQY8ePfDvv//Cz88PALB582Y8+eSTGDduHLp06YImTZpYvO6JEyeia9euCreYiIicJb+gJ9edkURpLvLRVC3jh6pl/JzdDCrmtEkezPIgkmRLzKJWaABqhQYo15hipqQdenistYxayOAs+P+AZhVxMz4dA5pXdF6jqFiyqbrpxx9/jK+++kr28vPmzcMnn3xiyyZtcvHiRfzzzz9wd3fHkiVLtEFEAOjXrx+GDx+O/Px8zJ4922ltJCIi18GhzUSkq7jUTa0bFogaZf3xUttqzm4KEZEkBhItk59f8I+Ck5S3hzs+7FOP9ShJcSVqmqR169YBADp06IBq1YwvnIQZpTds2ICcnByHto2IiFyPkJHIhEQiKk78vDywa0JXZn6RxcoH+Ti7CVQCFfXOG0cR4q78uMjeHFojMTExET4+zjv5nDp1CgDQqlUr0eeFx9PT03H16lU0aNDAovWvW7cO69evx6NHj1C+fHm0b98evXr1gptbiYrXEhEVG082rYhNZ++iU+1yzm6K62K2ABFRifFEkwoIDfTG8z8ddnZTSjyefkkKA69kbw6LcK1evRqpqamoWrWqozZp5ObNmwAg2YagoCAEBQXpLWuJBQsWYP78+fjpp58wc+ZM9O3bF/Xr18exY8esbzQRETlNuUBvrB3dAcG+ns5uCpHTLFu2DCqVCiqVCvv37zd6Xq1Wo0qVKlCpVHjiiSdE1/Hw4UP4+PhApVLh0qVLossMHz5cux3DP87oiD548CA6duwIPz8/hIWF4e2330ZOZobs1y9ZsgT169eHj48PateujW+//dZomfDwcMn3XLt2bSXfDpEiVCoVHqtRxvQyzIeyG91ZswN9iv+8qYY1/8g0zqpOjmLR0Wf+/PmYP3++3mMPHjxAjRo1JF+jVqvx8OFDpKSkQKVSoX9/583SmZqaCgDw9/eXXCYgIAApKSlISUmRvd5OnTrhpZdeQufOnVG5cmUkJSXh2LFjmDp1Ki5cuICePXvi8OHDqF+/vuQ6srKykJWVpf2/JdsnIiJyGl7dlxg+Pj5YsWIFOnbsqPd4REQEYmJi4O3tLfna1atXQ6VSISwsDMuXL8enn34qupy3tzcWL15s9Li7u2PrO50+fRo9evRA/fr1MXfuXMTExOCrr75C6VrH4PXEVLOv//HHH/HGG29g0KBBGD9+PPbt24e3334bGRkZ+PDDD7XLffPNN0hLS9N77a1btzB16lT06tVL8fdF5EwhfuyUU8qEXnUwokN1ZzeDXEzh0GZenJF9WRRIfPjwIaKiovQey8vLM3pMSo8ePayesfmDDz7Av//+a/HrFi9ebHTBq7SZM2fq/d/Pzw+VKlVC37590alTJxw7dgyTJk3C+vXrJdcxe/Zsp05EQ0RERGRKv379sHr1aixYsAAeHoWXkCtWrEDLli0RHx8v+do//vgD/fr1Q7Vq1bBixQrJQKKHhwdeeuklxdtuqcmTJyMkJAR79uzRjlYJDw/Ha6+9htCGJ4GwLpKvffToEaZMmYL+/ftjzZo1AIDXXnsN+fn5mDlzJl5//XWEhIQAAJ566imj1wufzYsvvqjwuyJyDKlhlZ7uLPdkKyFQNLBF5RLxeTLBzkLFZEIwcn0WBRKfeuophIeHA9BkGr7yyisIDg7GN998I/kaNzc3BAUFoVGjRqhZs6bVDb179y4iIyMtfp1uL29gYCAATQ1Ec8sLF4228Pb2xpQpU/DUU09h69atyMnJgaeneE/cpEmTMH78eO3/U1JSUKVKFZvbQERERKSEwYMHY926dfjvv//Qt29fAEB2djbWrFmDqVOnYsGCBaKvu337Nvbt24e//voL1apVw7x583Dw4EG0b9/ekc2XLSUlBf/99x/effddvevBoUOH4s0x7yDj8n6gnXQgcffu3UhISMDo0aP1Hn/rrbewfPlybNq0yWSwdMWKFahevbrLfj5E1mpY0fb7K9IoaXEiBsYsw4+L7M2iQGLTpk3RtGlT7f9feeUV+Pr6YtiwYYo3zNAff/yBP/74w6Z1hIeH4+TJk7h9+7bo87pDmoWAqa2E4cxZWVmIj49HhQoVRJfz9vY2OSSIiIiIyJnCw8PRrl07rFy5UhtI3LJlC5KTk/HCCy9IBhJXrlwJf39/PPHEE/D19UXNmjWxfPlyyUCZWGajl5eX2U7epKQk5OXlmX0ffn5+2PVeF3hITIZ37tw55ObmGk3O5+XlhZAqtZEce93k+qUm92vZsiXc3Nxw6tQpyUDiqVOncOnSJUyZMsXs+yBypE+ebIhTt5NsWsenAxsr1BoqeRgak0NbU5IfF9mZTfnQ+fn5uHv3rlJtsbsWLVoAAI4fPy76vPC4v78/6tSpo8g2ExIStP8WMiKJiIiKupfaVkO/xmFoZ6boPhUvQ4YMwfr16/Ho0SMAwPLly9GlSxdUrFhR8jXLly/HgAED4OvrCwB4/vnnsWrVKuTm5hotm56ejnLlyhn9ee6558y2rXnz5qKvNfzzxRdfoEa5AFQt4ye6nnv37gGAaOevT3AZ5KUlmmzHvXv34O7ujtDQUL3Hvby8UKZMGZPXzsuXLwfAYc3keoa1D8c3LzSXtaxUDMPX07G1Tqno48hmy+QLQ5sZeCU7K/5TPel46qmnMHXqVBw4cAC3b982mr15xYoVAID//e9/kkOQLfXnn38C0GQmBgQEKLJOIiIiZyvt74XvX2zp7GaQgz333HMYN24cNm7ciD59+mDjxo2SmYgAcPbsWZw7dw6zZ8/WPjZ48GDMmjUL27ZtM5qEz8fHBxs2bDBaT9myZc22bfny5doApymmJgkEoF2H2EgRd08v5Odmm329l5eX6HM+Pj6SbczPz8eff/6J5s2bm5ygj6iocmNsgyzEGomWaVvQudu6emknt4SKO5sCiYcPH8bo0aPRrl07LFy40OSyI0eOxMmTJ/HTTz8ZDfVwlIYNG2LAgAH4559/8Oqrr+Lff//V9o5v2bIFy5Ytg5ubGyZNmmT02qFDh+Lo0aMYM2YMxowZo3189+7dUKlU6NKlC1Q6OcTZ2dmYO3cuvv32WwDAe++9Z+d3R0RERGRf5cqVQ8+ePbFixQpkZGQgLy8PzzzzjOTyf/zxB/z9/VGjRg1cu3YNgCaYFh4ejuXLlxsFEt3d3dGzZ0+r2tahQwerXmdIuDbMysoyei4vJxtuHl5613xir8/OFg82ZmZmatdvKCIiAnfu3MG7775rRauJXIjE74NZUgoooYE1DtWVp3pZf0TN6W9+QSIb2RRIXLFiBc6cOYMPPvjA7LJt27bFL7/8ghUrVjgtkAgAP/30Ey5evIgdO3agZs2a6NSpE+Li4hAREQG1Wo358+ejSZMmRq+7ffs2IiMjjer2nDlzBu+++y7Kly+PZs2aoUyZMnjw4AHOnj2L2NhYAMCECRPw6quvOuT9ERER2SrQp0QNWCALDRkyBK+99hru37+Pvn37olSpUqLLqdVqrFy5Eunp6WjQoIHR83FxcUhLS1NsxMaDBw9k1UgMCAgwuU1hSLMwxFlXVkoC3ANKw89LeohmhQoVkJeXh7i4OL3hzdnZ2UhISJAcBr58+XK4ublh8ODBZt8DUVGkKv6TDDsMA2tE5Ew23SlEREQAAHr16mV22YEDB+L111/H7t27bdmkzUJDQ3H8+HHMnj0bf//9N/755x/4+/ujd+/emDBhAnr06GHR+rp06YI333wTJ06cwNmzZ5GYmAg3NzdUrFgRvXv3xqhRozjrHhERFRkXPukNN96hkAkDBw7EqFGjcPjwYfz111+Sy0VERCAmJgYzZswwGqqblJSE119/HevXrzc5g7ElWrdujVu3bpld7qOPPsLHH38s+XyjRo3g4eGB48eP69VmzM7ORnLMNXjWao/XOlWXfH2zZs0AaGpv9+vXT/v48ePHkZ+fr31eV1ZWFv7++2907drVZL1JoqKM5xbbqUtYSqJ28hAnt4OI9NkUSIyJiUFwcDBKlzY/Br9MmTIIDg7GnTt3bNmkIoKCgjB79my9ej3m7NmzR/Tx5s2b4/vvv1eoZUTFX8VgH2c3gYhM8PdmNiKZFhAQgB9++AFRUVH43//+J7mcMKz5/fffh4+P8bH/yy+/xPLlyxULJCpVIzE4OBj/b+/O46uo7v+Pvy8J2TdD2AJhD2DZihpBocoOCkQFAhTZapSKWhCpiiiKIn4rv4oPoVSwsolQRQQtX1BZxCqLyCaKFRuEgKwFQ0gIEEjy+f3B995yvUluVrLwej4e9/EIc86ZOXPncM7M556Z6datm95++21NmjTJ9bK8RYsWKfN8hsKbd1RVn8tTq86dO6dDhw4pKirK9RzHLl26KDIyUq+//rpbIPH1119XUFCQx+3ckrR69WqlpqbykhVUCnkFfQgGlZxr7Tbx/B4nAeDqK9bVQn4Pk86NmSk9Pb04mwRQgX3+eGeFBRKkAICKbsSIEfmmO2fYde/ePdcgoiTFx8frtddec7sFOCsrS2+//Xau+e+55x4FBwfnuc2SekaiJE2dOlW33nqrbr/9do0aNUqHDx/WK6+8ouY3ddT5Rv99ydBXX32lzp07u81yDAwM1JQpU/Twww8rISFBPXv21BdffKG3335bU6dOzfUH+MWLF8vf31/9+/cvsX0AStO6x25XRqbnm9fzw4xEFNq1NQETqDCKdUVfo0YN/fTTTzp69KjX2zCOHDmitLQ01alTpzibBFCB1asWVNZVAABcBatWrVJqamq+Mxb79u2rV155Re+8847GjBkj6XIActiwYbnmP3DgQL6BxJJ0ww03aN26dXryySc1btw4hYaGKjExUQ16JeovG496fT7ZQw89pKpVq+qVV17RP/7xD8XExOjVV1/V2LFjPfKmpaVp1apV6t27t8LDw0tpj4CS1aRG3s8ZjQzOfaIJcUQAqByKFUhs3769fvrpJ82aNUtTp07NN6/zrc7t2rUrziYBAABwFY0cOVIjR470mi85Odn1d79+/WSW/1SS22+/3S3PggULtGDBgiLWsuR17NhRmzZtclv2l0+T3P7dqVOnPPfzgQce0AMPPOB1O2FhYQW6JRuoKF5JaKON+07pD3/f5bacQGLxeelWKy2aDlC+FOvdWYmJiTIzTZs2TW+88Uae+ebMmaNp06bJ4XDw9mIAAABUaNfqxTxQENcF+6lvG8+71bi1ueRcK18lXS1QPhVrRmL37t01YMAALVu2TKNHj9asWbPUp08f1a9fX5J08OBBrVy5Ut99953MTP3799cdd9xRIhUHAAAAriYe+A8UHYHE4rvWfsRwzvim6QDlS7HferBw4UI5HA699957+vbbb7Vnzx63dOd//sGDB2vu3LnF3RwAAABQJrzdrg0gb8WJBW2d2FUXs3JKrC4VHXE1AGWpWLc2S5ffTPfuu+9q3bp1GjJkiOrXry9/f38FBASoQYMGuvfee/Xpp59qyZIlCgwMLIk6AwAAAGWG2TFA4RXn/03NsADFRPLSPgAoD4o9I9GpS5cu6tKlS0mtDgAAAABQSfBogOIb36OpMi5m5flm7MrGOQfcwRxMoFwp9ozEgsrJydHKlSt19913X61NAgAAAABQKcTWDNWixHby9blql/Flyvk0CWLQQPlSYjMS85KUlKS5c+fqrbfe0okTJ0p7cwAAAAAAAABKQakEEs+dO6elS5dq7ty52rx5s6T/Ppz6+uuvL41NAgAAAKWqUfUQSVKN0IAyrgkAVH683goon0o0kPjll19q7ty5Wrp0qc6ePSvpcgCxefPmSkhIUEJCglq2bFmSmwQAAACuijtb1daGP3ZSw6jgsq4KAFwzuLUZKF+KHUg8efKk3nrrLc2bN0979+6V9N/Zhw6HQ9u2bdONN95Y3M0AAAAAZY4gIgAAuJYVKZBoZlq9erXmzZun//3f/1VWVpbMTIGBgbr77rs1YsQI9erVSxK3MgMAAAAAgMJxTlACUL4UKpD4448/at68eVq4cKGOHTsmM5PD4VDHjh01fPhwDRw4UKGhoaVVVwAAAABABXNj/evKugqogJxhRAf3NgPlSqECibGxsXI4HDIzNWzYUMOHD9fw4cPVsGHD0qofAAAAAKACGxwXU9ZVQAVGGBEoX4p0a/OYMWM0bdo0+fn5lXR9AAAAAACVCDeooihG3tpAn+39j6qFEHcAypMqhcns7+8vM9PMmTMVHR2thx9+WF9++WVp1Q0AAAAAAFyD4hpE6rsXesnf16esqwLgCoUKJB47dkwzZsxQ69atlZKSotdff10dOnRQs2bN9NJLL+nQoUOlVU8AAAAAAAAAZahQgcSIiAg98sgj2rVrl3bs2KHRo0crPDxcSUlJmjRpkho1aqQuXbpo/vz5pVVfAAAAAEAFwjPuAKDyKFQg8Upt27bVrFmzdOzYMS1atEi33367zEyfffaZ7r//fle+NWvWKCsrq0QqCwAAAAAAAKBsFDmQ6OTv7697771Xn376qfbt26enn35aderUkSSZmfr3768aNWrod7/7nVavXk1QEQCucZHBPDAbAAAAACoih5mV+Eu0zEyffPKJ3nzzTa1cuVKXLl2Sw3F5QntERIR+/vnnkt5kpZOWlqbw8HCdOXNGYWFhZV0dACgxFy5ly0wK9OPB2QAAVFYNJqxy/f3/BrRWwk0xZVgbAKgcykOsqNgzEnPjcDjUq1cvLVu2TEeOHNGf//xnXX/99TIzpaamlsYmAQAVREBVH4KIAABcQ5yTSgAAFV+pBBKvFBUVpccee0x79uzR5s2blZiYWNqbBAAAAACUoQW/i1O/tnXKuhoAgBLmezU31r59e7Vv3/5qbhIAAAAAcJV1alZDR1LPa/muI2VdFQBACSr1GYkAAAAAgGtPyT+NHwBQ1ggkAgAAAABKDU9IBIDKg0AiAAAAAAAAAK8IJAIAAAAAAADwikAiAAAAAAAAAK8IJAIAAAAAShzvWgGAyodAIgAAAACgxHVqWl2SdEvjamVcEwBASfEt6woAAAAAACqfmMggJf+pd1lXAwBQgpiRCAAAAAAAAMArAokAAAAAAAAAvCKQCAAAAAAAAMArAokAAAAAAAAAvCKQCAAAAAAAAMArAokAAAAAAAAAvCKQCAAAAAAAAMArAokAAAAAAAAAvPIt6wogd2YmSUpLSyvjmgAAAAAAAKCsOWNEzphRWSCQWE6lp6dLkmJiYsq4JgAAAAAAACgv0tPTFR4eXibbdlhZhjGRp5ycHB09elShoaFyOBxlXZ0SkZaWppiYGP30008KCwsr6+qgHKKNwBvaCAqCdgJvaCPwhjYCb2gj8IY2Am+K0kbMTOnp6YqOjlaVKmXztEJmJJZTVapUUd26dcu6GqUiLCyMjhT5oo3AG9oICoJ2Am9oI/CGNgJvaCPwhjYCbwrbRspqJqITL1sBAAAAAAAA4BWBRAAAAAAAAABeEUjEVePv76/nnntO/v7+ZV0VlFO0EXhDG0FB0E7gDW0E3tBG4A1tBN7QRuBNRW0jvGwFAAAAAAAAgFfMSAQAAAAAAADgFYFEAAAAAAAAAF4RSAQAAAAAAADgFYFEXBXvvfeeOnXqpOuuu07BwcFq06aNpk2bpkuXLpV11VBAly5d0vr16/X4448rLi5OERERqlq1qmrVqqX4+HitWrUq13KTJ0+Ww+HI97N37948t7tv3z6NHDlSdevWlb+/v+rWrauRI0dq//79+dY3PT1dEydOVLNmzRQYGKioqCj17t1bn376abG+B+Rv5MiRXo/3hQsXci27Y8cOJSQkqGbNmgoICFDDhg31hz/8Qf/5z3/y3eaJEyf0yCOPqGHDhvL391fNmjWVkJCgnTt35lvu4sWLevnll9WmTRsFBwfruuuuU6dOnbRs2bIi7z/yl5yc7LV9OD+ff/65qxz9SOXzww8/aObMmRo5cqRatWolX19fORwOvfjii17Lrlu3TnfeeaeioqIUGBio5s2b6+mnn9bZs2fzLXe120FOTo7mzJmjdu3aKTQ0VKGhoWrXrp3eeOMN8Yhy7wrbRnJycrR582Y9++yz6tixo6pVq6aqVasqKipK3bt31+LFi/P83hcsWOC1j/n444/zrCvjUNkoSj9S0cYT+pHiKUobKeh5yltvveVWjn6k4inq9a3TNX0+YkApGzt2rEkyX19f69Gjh/Xr188iIiJMknXs2NHOnTtX1lVEAaxdu9YkmSSrVauW9e7d2wYOHGgtW7Z0LR81apTl5OS4lXvuuedMkrVp08ZGjBiR6+fo0aO5bnPjxo0WFBRkkqxFixY2aNAga9GihUmy4OBg27JlS67lTpw4YU2bNjVJVrt2bUtISLDbbrvNHA6HORwOmzFjRol/P7hsxIgRJsk6dOiQ5/G+ePGiR7n33nvPfH19TZLFxcXZwIEDrVGjRibJatasaUlJSblu74cffrAaNWqYJGvUqJENHDjQ4uLiXH3O8uXLcy2XkZFht956q0myiIgI69evn/Xo0cNVh/Hjx5fo94LLTp48mWe7GDFihOvYhYaG2tmzZ13l6EcqH+e5wS8/U6ZMybfc9OnTTZI5HA677bbbLCEhwWrVqmWSrFmzZnby5Mlcy13tdpCVlWX9+vUzSRYUFGR9+/a1vn37WmBgoEmyhIQEy87OLtyXdo0pbBtJSkpy5YmMjLQePXrYoEGDXP2KJOvTp49lZmZ6lJ0/f75JssaNG+fZx3zzzTe5bpdxqOwUpR+pSOMJ/UjxFaWN5Hee0rlzZ9cYlJyc7FaOfqTiKer1rRnnIwQSUapWrFhhkiwkJMR27NjhWn7y5Elr1aoVHVwFsn79euvfv799/vnnHmnvvPOO+fj4mCRbuHChW5rzhO25554r1PYyMjIsOjraJNlTTz3llvbUU0+ZJIuJick1EH3XXXeZJOvatatlZGS4lq9atcp8fHysSpUqtnv37kLVBwXjDCTOnz+/wGWOHDniGlDnzJnjWp6VlWVDhw51BRd/OYjn5ORY27ZtTZINGzbMsrKyXGlz5sxx9T3Hjh3z2KbzxLJVq1ZuA/327dstJCTEJNnKlSsLsecoCXfccYdJsgceeMBtOf1I5fO3v/3N/vjHP9rixYvt+++/t2HDhnm9uNu5c6c5HA7z8fGx1atXu5ZnZGRY165dTZL179/fo1xZtINXX33VJFmdOnVs//79ruX79+931WXmzJn5f0nXuMK2kX379lmXLl3so48+chsPzMw+++wzCw4ONkn2/PPPe5R1BgBGjBhRqDoyDpWtovQjFWk8oR8pvqK0kfyMHj3aJFn37t090uhHKp6iXt9yPkIgEaXM+UvKiy++6JH2xRdfmCTz9/e31NTUMqgdSlJiYqKrc7tSUU/YZs2aZZKsadOmHr+SZGdnu36RmT17tlvad999Z5LMx8fH45fCK+s5ePDgQtUHBVOUQOLjjz9ukqxbt24eaenp6RYeHm6S7OOPP3ZLW7VqlesX2PT0dI+yzoF8woQJbstTUlLMz8/PJNnGjRs9yk2ZMsUkWfv27Qu8Dyi+w4cPW5UqVUySffnll25p9COVn7PvyO/iLiEhwSTZ/fff75GWnJzsaj/ff/+9W9rVbgfZ2dmuWQlvv/22R7lFixaZJIuOjmY2USEUpI3kx9m3N27c2COtqAEAxqHypSBtpKKMJ/QjpaM4/cj58+ddd9W98847Hun0I5VPXte3nI+Y8YxElJojR45o27ZtkqQhQ4Z4pHfs2FExMTHKzMzU6tWrr3b1UMLatm0rSfrpp59KZH0rVqyQJA0ePFhVqrh3VVWqVNGgQYMkScuXL8+1XIcOHVS/fn2P9Trb4sqVK3lGZznhPGa59RMhISGKj4+XlPexjo+PV0hIiEdZ5/p+WW716tW6ePGi6tWrpw4dOuRZ7ssvv9TRo0cLuzsoogULFignJ0ctWrRQu3btSmSd9COVx8WLF13PKsqtr6hfv77r/7Pz+Dld7XawZcsWHT9+XP7+/urfv79Huf79+8vPz09Hjx7V1q1b89lrlKSSPk+RGIeuJfQjeP/995WamqrIyEjdfffdJbZe+pHyK7dxg/OR/6tvoXIDhbBr1y5JUmRkpBo2bJhrnptuusktLyqupKQkSVLt2rVzTd+5c6cmTJigUaNG6fHHH9eSJUuUnp6e5/qcbcLZRn4pr7ZT0HIZGRmuOqPkbdiwQePHj9eoUaP01FNPacWKFcrMzPTIl56ern379kkqvWOdlJSkjIyMApdr1KiRIiMjJUlff/11rnlQ8hYsWCBJSkxMzDMP/ci169///rfOnTsnqfSOZ0m1A2e5Fi1aKCAgwKNcYGCgWrRokes2UXq8nadIlx+A/8wzz2jUqFF67LHHNG/ePJ06dSrP/IxDFVd5H0/oR8qfefPmSZKGDh0qf3//PPPRj1QeuY0bnI9c5luo3EAhHDhwQJJUr169PPPExMS45UXFdPz4cVcQILdfO6TLv5CsXLnSbVl4eLhmzJih4cOHuy1PT0/Xzz//LCnv9uNsOydPnlRGRoaCg4MleW93YWFhCgsLU1pamg4cOKBf/epXBdhDFNYv32QnXR6E582bp169ermWJScnu/72dqx/2U94O9bOcmam5ORk10BZkL6pbt26SklJoW+6Sv75z39q37598vPz07Bhw/LMRz9y7XIek4iICIWGhuaaJ7e+oizaQUHPf3bt2kUfc5WcO3dOM2bMkJT3eYokbdq0SZs2bXJbFhAQoMmTJ+vJJ5/0yM84VHGV9/GEfqR8SU5O1oYNGyTl/4OnRD9SWeR1fcv5yGXMSESpcf6q5/wPkBvn9O20tLSrUieUvKysLA0dOlRnzpxRq1at9Pvf/94tvXHjxnrppZe0a9cupaSkKCUlRRs3blSfPn105swZjRgxQosXL3Yrc+Uvwnm1nyun/l/Zfmh3ZatNmzZ67bXXtGfPHqWlpenEiRNas2aNbr31Vh07dkzx8fH67LPPXPkLc6x/eby8HWvaSMXh/JU/Pj5eUVFRHun0IyjqMSmLdkD7KX8eeughHThwQNHR0Zo4caJHeq1atfT0009r69atOnnypNLS0rRt2zYNHz5cmZmZmjBhgl566SWPcoxDFU9FGU9oI+XL/PnzZWa66aab1Lp161zz0I9UHvld33I+chkzEgEUy4MPPqj169erWrVqWrZsmfz8/NzSc5td1KFDB61cuVJjxozRzJkzNW7cOCUkJHiURcUzbtw4t3+Hhoaqe/fu6tatm+655x59+OGHevTRR7nFAi5paWlatmyZJOm+++7LNQ/9CICimjJlihYuXKiAgAAtXbpU1apV88jTq1cvt9ny0uVbxRYuXKg2bdpo/PjxeuGFF5SYmKiaNWteraqjFDCeoLBycnJcM9PyOk+R6EcqE2/Xt2BGIkqRc6rvlc9y+KWzZ89KujwdFxXP2LFjNXfuXF133XVau3atmjZtWqjykydPlo+Pj06ePOn2gNcrp4nn1X6cbUdybz+0u/LJ4XDo+eeflyTt3r3b9dDiwhzrXx4vb8eaNlIxvPPOOzp37pzq1q2rnj17Fro8/ci1oajHpCzaAe2n/Jg+fbqeffZZ+fv7a8WKFbm+kMCbsWPHKioqSpmZmVqzZo1bGuNQ5VKexhPaSPmxbt06HTp0SIGBgbm+XKMg6EcqDm/Xt5yPXEYgEaWmQYMGkvJ/O54zzZkXFcf48eM1Y8YMRUREaM2aNa63WhVGZGSkatSoIUk6fPiwa3loaKjr4cCHDh3Ktayz7URFRblN13a2pbzKpaWluaZu0+6uruuvv971t/N4X/nmMW/H+pfHy9uxdpZzOBxu2/FW7sr60UZKn/O25pEjR3q8wa4g6EeuDc7vOTU1Nc8XIuTWV5RFOyhIH8P5T+mbOXOmxo8fLz8/P73//vseM4UKysfHR7GxsZLc+xiJcaiyKU/jCf1I+eE8T+nfv7/Cw8OLtA76kYqhINe3nI9cRiARpcb5H+/nn3/O8+Gd27dvlyTdcMMNV61eKL4nnnhC06dPV3h4uNasWZPnG6S8yc7O1pkzZyTJ42G1zjbhbCO/lFfbKWi54ODgQs+gRPE4HzAs/fd4h4WFqUmTJpJK71jHxsa6PXPEW7n9+/crJSVFkooUIEfB/etf/9LWrVvlcDj0u9/9rkjroB+5NjRr1kxBQUGSSu94llQ7cJb77rvvdOHCBY9y58+f13fffZfrNlEyZs2apTFjxriCiL179y7W+pzjV1H7GMahiqE8jSf0I+VDSkqKPvjgA0neX7LiDf1I+VbQ61vOR/6PAaUoLi7OJNmLL77okfbFF1+YJPP397fU1NQyqB2K4sknnzRJFh4ebl999VWx1rV8+XKTZA6Hw44dO+aWNmvWLJNkTZs2tezsbLe07Oxsa9q0qUmy2bNnu6Xt2bPHJJmPj48dPHjQY5uJiYkmyQYPHlysuqPwpk+fbpIsLCzMLl686Fr++OOPmyTr1q2bR5n09HSLiIgwSfbxxx+7pa1atcokWUREhJ09e9ajbNeuXU2STZgwwW15SkqK+fn5mSTbuHGjR7kpU6aYJGvfvn1RdxUF9Nhjj5kk69KlS5HXQT9SOYwYMcIk2ZQpU/LMk5CQYJLs/vvv90hLTk42Hx8fk2Tff/+9W9rVbgfZ2dlWq1Ytk2Rvv/22R7lFixaZJIuOjvaoD/JWkDZiZvb666+bJPPz87OVK1cWe7s7duwwSSbJtm7d6pbGOFS+FLSN5KU8jSf0I6WjsG1kxowZJskaN25sOTk5Rd4u/Uj5VtjrW85HzAgkolStWLHCJFlISIjt2LHDtfzUqVPWqlUrk2Tjx48vwxqiMJ5++mnXQFeQTvbgwYO2aNEiO3/+vEfaihUrLDIy0iTZ0KFDPdIzMjIsOjraJNnEiRPd0iZOnGiSrG7dunbu3DmPsnfddZcrMHVl+urVq83Hx8eqVKliu3fvLsguoxB27dplH374oV26dMlteXZ2tr355psWEBBgkuyZZ55xSz9y5IgFBQWZJHvjjTdcy7OysmzYsGEmyeLi4jxO4HJycqxt27YmyYYPH25ZWVmutDlz5rj6nl9eDJiZjR071iRZ69at7dSpU67lO3bssJCQEJNUIhegyNvFixetRo0aJskWL16cZz76kWtDQS7uduzYYQ6Hw3x8fOyjjz5yLc/IyHBdaPXv39+jXFm0g1dffdUkWZ06dWz//v2u5fv377c6deqYJJs5c2b+XwrcFKSNvPHGG+ZwOAoVRMzIyLC//OUvlpaW5pH2z3/+0xo0aGCSrGPHjh7pjEPli7c2UtHGE/qRklfYQOKvf/1rk2RTp07NNx/9SMVV2OtbM85HzAgk4ioYM2aMSbKqVatar169rH///q4ZRh06dMj1PwrKnw8//ND1S9pNN91kI0aMyPVzZWB4165drsHvN7/5jQ0ePNjuuusui42Nda2rc+fOlp6enus2N27c6AowtWzZ0gYPHmwtW7Y0SRYcHGxbtmzJtdyJEydc26hdu7YNHDjQOnXqZA6HwyTZa6+9Virf0bXO+cPBddddZ127drUhQ4bYnXfeafXq1XMd79/+9rcegUYzs6VLl7p+vWvXrp0NGjTIGjVqZJKsZs2alpSUlOs29+7da9WrVzdJ1qhRIxs0aJDdfPPNJsl8fX1t+fLluZbLyMiwW265xVXf/v37W69evaxq1aomyR577LES/W7gyTnzIyIiIteLOif6kcppx44d1q5dO9cnKirKdQJ95fKjR4+6lXPObHY4HNapUycbOHCg1a5d2yRZs2bN7OTJk7lu72q3g6ysLLvnnntMkgUFBVl8fLzFx8e76jBgwABmEXlR2Daya9cu13Fp3rx5nucpI0aMcNvO6dOnTbp8h0z79u1t4MCB1q9fP1f7kGStWrXyaItOjENlpyhtpCKNJ/QjxVfUscbMbOfOna5ZYEeOHMl3O/QjFVNRrm+drvXzEQKJuCreffddu+222ywsLMwCAwOtZcuW9qc//ckyMzPLumoooPnz57s62vw+9evXd5U5deqUPfnkk9alSxerV6+eBQcHW9WqVa127drWp08fW7JkideOKykpyYYPH27R0dFWtWpVi46OtuHDh9u+ffvyLXfmzBmbMGGCxcbGmr+/v0VGRlqvXr1s3bp1JfF1IBf79++3Rx991Dp27Gh16tSxgIAA8/f3t3r16tmAAQNs1apV+Zbfvn279evXz6pXr25+fn5Wv359e/jhh+348eP5ljt27Jg9/PDDVr9+ffPz87Pq1atbv3793GZB5yYzM9P+53/+x1q2bGmBgYEWHh5ut912my1durTQ+47C69Onj0myhx56KN989COV04YNGwo0phw4cMCj7Nq1a61Xr14WGRlp/v7+Fhsba0899VSuM0GudLXbQXZ2ts2ePdtuuukmCw4OtuDgYIuLi7PZs2cX6xa5a0Vh20hB80vulz+ZmZk2adIku+OOO6xhw4YWGhpqvr6+Vr16devWrZvNmTPH6/kq41DZKGwbqYjjCf1I8RRnrHnkkUdMkt15551et0M/UjEV5fr2Stfy+YjDzEwAAAAAAAAAkA/e2gwAAAAAAADAKwKJAAAAAAAAALwikAgAAAAAAADAKwKJAAAAAAAAALwikAgAAAAAAADAKwKJAAAAAAAAALwikAgAAAAAAADAKwKJAAAAAAAAALwikAgAAAAAAADAKwKJAAAAlZDD4ZDD4dBnn31W1lUBAABAJUEgEQAAoJxxBgGL8lmwYEFZV7/cSU5O1uTJkzV58uSyrgoAAECF5lvWFQAAAIC7mjVr5rr87NmzysjIyDdPYGCgJKlZs2aSpKCgoFKoYcWSnJys559/XpIIJgIAABQDgUQAAIBy5vjx47kunzx5sisgllcep71795Z4vQAAAHBt49ZmAAAAAAAAAF4RSAQAAKiE8nrZSnJysistOTlZBw8e1AMPPKB69eopICBAjRs31jPPPOO6hVqS9uzZo6FDhyomJkYBAQGKjY3Viy++qEuXLuVbh+TkZD366KNq0aKFQkJCFBQUpObNm2vs2LE6dOhQkfdt7969GjVqlJo2baqgoCAFBAQoJiZG7du318SJE91mYzZo0ECdO3f2+F6cn5EjR3qsPz09XX/60590yy23KDIyUv7+/oqJidHgwYO1ZcuWPPf1yu81KSlJI0eOVN26deXv76969erpwQcf1NGjR/Pcr8OHD2vcuHFq0aKFgoOD5e/vr+joaN14440aN26ctm3bVuTvDAAAoCRwazMAAMA1aufOnUpMTFRqaqrCwsKUlZWl/fv3a+rUqfr888+1fv16rVmzRgMHDtS5c+cUHh6uixcvat++fZo0aZL27Nmjd955J9d1L168WImJicrMzJQk+fv7q0qVKvrhhx/0ww8/aP78+Vq2bJl69OhRqDqvXbtWffv2da23atWqCg4O1uHDh3X48GFt3bpVfn5+rmchVq9eXWlpaTp9+rQkz2dLhoeHu/3766+/Vt++fXX48GFJko+Pj4KCgnT48GG9++67Wrp0qaZOnaqnnnoqzzpu3bpVDzzwgNLT0xUSEiIfHx/99NNPmjNnjt577z2tXbtWN9xwg1uZ3bt3q3Pnzq56+vj4KCwsTMePH9exY8e0c+dOnT59mpfpAACAMsWMRAAAgGtUYmKibrzxRn333Xc6c+aM0tPTNWPGDPn4+OiLL77QCy+8oHvvvVd9+/ZVcnKyUlNTlZaWpqefflqS9O6772rdunUe6127dq2GDx+u7OxsPfHEEzpw4IDOnz+vjIwM7d27VwkJCUpPT1dCQkKhZyaOHj1amZmZ6tGjh7799ltdvHhRp0+f1vnz57Vnzx49//zzatCggSv/tm3btHz5cte/jx8/7vZ57bXXXGnHjh1Tz549dfjwYfXr10/bt2/X+fPnlZaWphMnTmjSpEny8fHRxIkT9cEHH+RZx9///vdq2LChtm7dqvT0dGVkZOiTTz5RvXr1lJKSonvuuUfp6eluZcaPH6/Tp0/rhhtu0JYtW3Tp0iWlpKTowoUL+ve//60///nPatGiRaG+KwAAgBJnAAAAqBCee+45k2QFOYVz5tuwYYPb8gMHDrjSWrRoYRcuXPAoO2zYMFee7t27W05Ojkee3/zmNybJEhMT3ZZnZ2dbbGysSbI5c+bkWb/4+HiTZGPHjvW6L04nTpxw1evo0aMFLrdhw4YCfW/33XefSbIhQ4bkmWf69Okmydq0aeO2/MrvtVq1anbixAmPsv/617/Mz8/PJNm0adPc0gIDA02Sbd68ucD7BQAAcLUxIxEAAOAaNW7cOPn7+3ss79mzp+vvCRMmyOFw5Jnnm2++cVv++eefKykpSVFRUbr//vvz3Pbw4cMlSZ988kmB6xsaGqoqVS6fvh47dqzA5QriwoULWrJkiSTpySefzDOfs967d+/WiRMncs3z4IMPqkaNGh7Lr7/+eg0YMECSPG4Jj4iIkFTy+wUAAFCSeEYiAADANermm2/OdfmVzxGMi4vLN4/zmX5OmzZtkiSdOXNG0dHReW774sWLkqSDBw8WuL6BgYHq2rWr1q5dq169eunBBx9U79691bZtW/n5+RV4PbnZsWOHLly4IEkFfm7jwYMHPZ65KEldunTJs0yXLl20ZMkSffPNN7p06ZKqVq0qSerTp4/+9re/acSIEdq0aZPi4+MVFxenoKCgIuwNAABA6SCQCAAAcI0KDQ3Ndbmvr2+B8/zyzc3OtxJfunQpzxl7Vzp//nyB6ur05ptvKj4+Xrt379aUKVM0ZcoU+fn5KS4uTnfddZcSExMVGRlZqHVeWW9JBaq3JJ07dy7X5XXq1MmzjDMtKytLKSkprkDktGnTtG/fPm3YsEHTp0/X9OnT5ePjo1//+tfq3bu3Ro0ale96AQAArgZubQYAAECJyc7OliS1a9dOZlagT2HUq1dPO3fu1Mcff6wxY8boxhtvVE5OjjZt2qQnnnhCTZo00aefflrkekuXg5sFqXenTp0KvZ28RERE6NNPP9UXX3yhJ554Qh06dJCvr6927NihF154QbGxsfr73/9eYtsDAAAoCgKJAAAAKDG1atWSVLhblgurSpUq6tmzp1577TVt375dKSkpWrx4serVq6fTp09ryJAhrlunC8pZb6n4dT9y5IjXNF9f31xnTnbs2FEvv/yyNm7cqNTUVH344Ydq1aqVzp8/r/vuu6/AsyUBAABKA4FEAAAAlJgOHTpIko4fP67t27dflW2GhoZqyJAhmjt3rqTLtyZ/++23rnTnC1ok5TkDMi4uzvWcxZUrVxarPhs2bPCa1rp1a9fzEfMSEBCg+Ph4LV++XNLlF8Js3LixWHUDAAAoDgKJAAAAKDGdO3dWkyZNJF1+K7S3mYEpKSkFXre3dQUGBrr+vjJ4GBYW5vo7NTU117LBwcEaMmSIJOnll1/WoUOH8t1WfvWePXu2Tp065bH8hx9+0LJlyyRJgwYNci3PyspSTk5OnuvLa78AAACuNs5EAAAAUGJ8fX01e/Zs+fr6auPGjbrtttu0fv16t5ey7N+/X7Nnz1ZcXJz++te/FnjdmzdvVuvWrfXqq6/q+++/dwXfzEybN2/W6NGjJUl169ZV69atXeWaNm3qmm345ptv5jkr8aWXXlJ0dLROnTqlW265RYsWLVJ6eror/eTJk3r//fd1zz336Le//W2e9bx06ZK6d++ubdu2ueq3bt069ezZU5mZmYqJidGDDz7oyn/48GHFxsbqxRdf1K5du5SVleVK++abbzR06FBJl4Odt99+e4G/LwAAgJJGIBEAAAAlqmvXrnrvvfcUGhqqrVu3qlu3bgoODlZUVJQCAgLUuHFjjR49Wtu3b5fD4SjUur/99ls99thj+tWvfqWAgABFRUXJz89PHTp00LfffquwsDAtWbJEPj4+rjJBQUEaNmyYJOmJJ55QSEiI6tevrwYNGuiPf/yjK1/t2rW1bt06NW3aVEePHtXw4cMVERGhatWqKSQkRDVq1NCAAQP0wQcf5DuDcM6cOfrxxx918803KzQ0VCEhIerevbsOHjyoiIgILV++3G2WpHQ5uDpp0iTdcMMNCggIULVq1eTv7682bdros88+k5+fnxYsWFCkN1IDAACUFN+yrgAAAAAqn7vvvlv79u3TX//6V3300UdKSkpSamqqgoOD1bx5c8XFxal379668847C7zOuLg4LV26VBs2bNBXX32lo0eP6tSpUwoICFCTJk3Uo0cPjR07VtHR0R5lZ82apZiYGL3//vv68ccfXbcu//IW5Ouvv17ffPONFi5cqPfff19ff/21UlJS5OfnpyZNmqht27bq3r27BgwYkGc927Vrp+3bt2vq1Klav369Tp48qTp16ujOO+/Us88+q7p167rlr1Onjv7xj39ow4YN2rJliw4fPqz//Oc/8vX1VZMmTdS5c2eNHTtWsbGxBf6uAAAASoPD8rq3AwAAAECBJCcnq2HDhpKkAwcOqEGDBmVbIQAAgFLArc0AAAAAAAAAvCKQCAAAAAAAAMArAokAAAAAAAAAvCKQCAAAAAAAAMArXrYCAAAAAAAAwCtmJAIAAAAAAADwikAiAAAAAAAAAK8IJAIAAAAAAADwikAiAAAAAAAAAK8IJAIAAAAAAADwikAiAAAAAAAAAK8IJAIAAAAAAADwikAiAAAAAAAAAK/+P1CWfG1CMH0OAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 1500x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "wss = ['WS_S1', 'TWS_S25A', 'TWS_S25B', 'TWS_S26']\n",
    "inv_yhat_reshape = inv_yhat.reshape((-1, 24, 4))\n",
    "inv_y_reshape = inv_y.reshape((-1, 24, 4))\n",
    "\n",
    "for i in range(len(wss)):\n",
    "    plt.rcParams[\"figure.figsize\"] = (15, 3)\n",
    "    error = inv_yhat_reshape - inv_y_reshape\n",
    "    plt.plot(error[:, -1, i], linewidth=1)\n",
    "    plt.xticks(fontsize=16)\n",
    "    plt.yticks(fontsize=16)\n",
    "    plt.xlabel('Time steps', fontsize=18)\n",
    "    plt.ylabel('Actual Error (ft)', fontsize=18)\n",
    "    plt.title('{}'.format(wss[i]), fontsize=18)\n",
    "    plt.text(9000, -0.6, 'MAE = {}'.format(float(\"{:.3f}\".format(sum(abs(error[:, -1, i]))/len(error)))), fontsize=12)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "d0f5ae72",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(19268,)\n",
      "11\n",
      "(19268,)\n",
      "63\n",
      "(19268,)\n",
      "34\n",
      "(19268,)\n",
      "34\n"
     ]
    }
   ],
   "source": [
    "up_thre = 0.5\n",
    "low_thre = -0.5\n",
    "\n",
    "for i in range(len(wss)):\n",
    "    plt.rcParams[\"figure.figsize\"] = (15, 3)\n",
    "    error = inv_yhat_reshape - inv_y_reshape\n",
    "    error_24h = error[:, -1, i]\n",
    "    print(error_24h.shape)\n",
    "    print(np.sum(error[:, -1, i] > up_thre) + np.sum(error[:, -1, i] < low_thre))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "88da85fe",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(19268,)\n",
      "5\n",
      "(19268,)\n",
      "37\n",
      "(19268,)\n",
      "6\n",
      "(19268,)\n",
      "28\n"
     ]
    }
   ],
   "source": [
    "for i in range(len(wss)):\n",
    "    plt.rcParams[\"figure.figsize\"] = (15, 3)\n",
    "    error = inv_yhat_reshape - inv_y_reshape\n",
    "    error_24h = error[:, -1, i]\n",
    "    print(error_24h.shape)\n",
    "    print(np.sum(error[:, -1, i] > up_thre))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "43865dcf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(19268,)\n",
      "6\n",
      "(19268,)\n",
      "26\n",
      "(19268,)\n",
      "28\n",
      "(19268,)\n",
      "6\n"
     ]
    }
   ],
   "source": [
    "for i in range(len(wss)):\n",
    "    plt.rcParams[\"figure.figsize\"] = (15, 3)\n",
    "    error = inv_yhat_reshape - inv_y_reshape\n",
    "    error_24h = error[:, -1, i]\n",
    "    print(error_24h.shape)\n",
    "    print(np.sum(error[:, -1, i] < low_thre))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "453e4fce",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6247adaf",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}