GAN AHDS tf classification.ipynb

{"cells":[{"cell_type":"code","source":["from google.colab import drive\n","drive.mount('/content/drive')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"R87TiToDDoYz","executionInfo":{"status":"ok","timestamp":1670402848239,"user_tz":-330,"elapsed":28164,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"outputId":"108680bb-125c-4be0-e050-775cec9e20c9"},"execution_count":1,"outputs":[{"output_type":"stream","name":"stdout","text":["Mounted at /content/drive\n"]}]},{"cell_type":"code","execution_count":2,"metadata":{"id":"icQNarJ2L3CI","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1670402855574,"user_tz":-330,"elapsed":7345,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"outputId":"f4b41f6d-c8b1-4fa3-e7c1-e013c029f26d"},"outputs":[{"output_type":"stream","name":"stdout","text":["Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n","Collecting emnist\n","  Downloading emnist-0.0-py3-none-any.whl (7.3 kB)\n","Requirement already satisfied: requests in /usr/local/lib/python3.8/dist-packages (from emnist) (2.23.0)\n","Requirement already satisfied: numpy in /usr/local/lib/python3.8/dist-packages (from emnist) (1.21.6)\n","Requirement already satisfied: tqdm in /usr/local/lib/python3.8/dist-packages (from emnist) (4.64.1)\n","Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.8/dist-packages (from requests->emnist) (1.24.3)\n","Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.8/dist-packages (from requests->emnist) (2.10)\n","Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.8/dist-packages (from requests->emnist) (2022.9.24)\n","Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.8/dist-packages (from requests->emnist) (3.0.4)\n","Installing collected packages: emnist\n","Successfully installed emnist-0.0\n"]}],"source":["%matplotlib inline\n","from matplotlib import pyplot as plt\n","import numpy as np\n","import tensorflow as tf\n","from tensorflow import keras\n","from tensorflow.keras import Sequential, layers\n","from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout, Activation, Flatten, BatchNormalization, Input ,concatenate\n","from tensorflow.keras.utils import to_categorical\n","from tensorflow.keras.regularizers import l2\n","from tensorflow.keras.models import Model\n","from tensorflow.keras.optimizers import Adam\n","from tensorflow.keras.preprocessing.image import ImageDataGenerator\n","from tensorflow.keras.callbacks import  EarlyStopping, ModelCheckpoint\n","import glob\n","import os\n","import PIL\n","import time\n","!pip install emnist\n","\n","from IPython import display"]},{"cell_type":"code","execution_count":3,"metadata":{"id":"BOKI_3ROOytg","executionInfo":{"status":"ok","timestamp":1670402855575,"user_tz":-330,"elapsed":15,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}}},"outputs":[],"source":["def make_generator_model():\n","    model = tf.keras.Sequential()\n","    model.add(layers.Dense(7*7*256, use_bias=False, input_shape=(100,)))\n","    model.add(layers.BatchNormalization())\n","    model.add(layers.LeakyReLU())\n","\n","    model.add(layers.Reshape((7, 7, 256)))\n","    assert model.output_shape == (None, 7, 7, 256) # Note: None is the batch size\n","\n","    model.add(layers.Conv2DTranspose(128, (5, 5), strides=(1, 1), padding='same', use_bias=False))\n","    assert model.output_shape == (None, 7, 7, 128)\n","    model.add(layers.BatchNormalization())\n","    model.add(layers.LeakyReLU())\n","\n","    model.add(layers.Conv2DTranspose(64, (5, 5), strides=(2, 2), padding='same', use_bias=False))\n","    assert model.output_shape == (None, 14, 14, 64)\n","    model.add(layers.BatchNormalization())\n","    model.add(layers.LeakyReLU())\n","\n","    model.add(layers.Conv2DTranspose(1, (5, 5), strides=(2, 2), padding='same', use_bias=False, activation='tanh'))\n","    assert model.output_shape == (None, 28, 28, 1)\n","\n","    return model"]},{"cell_type":"code","execution_count":4,"metadata":{"id":"Ryv6SYPgKFvV","executionInfo":{"status":"ok","timestamp":1670402855603,"user_tz":-330,"elapsed":41,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}}},"outputs":[],"source":["def make_discriminator_model():\n","    model = tf.keras.Sequential()\n","    model.add(layers.Conv2D(64, (5, 5), strides=(2, 2), padding='same',\n","                                     input_shape=[28, 28, 1]))\n","    model.add(layers.LeakyReLU())\n","    model.add(layers.Dropout(0.3))\n","\n","    model.add(layers.Conv2D(128, (5, 5), strides=(2, 2), padding='same'))\n","    model.add(layers.LeakyReLU())\n","    model.add(layers.Dropout(0.3))\n","\n","    model.add(layers.Flatten())\n","    model.add(layers.Dense(1))\n","\n","    return model"]},{"cell_type":"code","source":["generator = make_generator_model()\n","discriminator = make_discriminator_model()\n","generator_optimizer = tf.keras.optimizers.Adam(1e-4)\n","discriminator_optimizer = tf.keras.optimizers.Adam(1e-4)"],"metadata":{"id":"XcHxjUjxn3oQ","executionInfo":{"status":"ok","timestamp":1670402858701,"user_tz":-330,"elapsed":3138,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}}},"execution_count":5,"outputs":[]},{"cell_type":"code","execution_count":6,"metadata":{"id":"dfVvU4PWMbPP","executionInfo":{"status":"ok","timestamp":1670402858701,"user_tz":-330,"elapsed":14,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}}},"outputs":[],"source":["checkpoint_dir = '/content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/training_checkpoints_ahsd'\n","checkpoint_prefix = os.path.join(checkpoint_dir, \"ckpt\")\n","checkpoint = tf.train.Checkpoint(generator_optimizer=generator_optimizer,\n","                                 discriminator_optimizer=discriminator_optimizer,\n","                                 generator=generator,\n","                                 discriminator=discriminator)"]},{"cell_type":"code","source":["checkpoint.restore(tf.train.latest_checkpoint(checkpoint_dir))"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Xb4OZi6zD9UQ","executionInfo":{"status":"ok","timestamp":1670402866548,"user_tz":-330,"elapsed":7860,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"outputId":"c851f051-1d64-4717-846e-3f350fe316af"},"execution_count":7,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<tensorflow.python.training.tracking.util.CheckpointLoadStatus at 0x7fd98114a130>"]},"metadata":{},"execution_count":7}]},{"cell_type":"code","source":["noi = tf.random.normal([1, 100])\n","sample = checkpoint.generator(noi, training=False)\n","plt.imshow(sample[0, :, :, 0], cmap='gray')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":282},"id":"dPEQYWZwEALt","executionInfo":{"status":"ok","timestamp":1670402875766,"user_tz":-330,"elapsed":9225,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"outputId":"bd7fd726-5266-4550-a351-0e2c7326c77d"},"execution_count":8,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.image.AxesImage at 0x7fd9808e96a0>"]},"metadata":{},"execution_count":8},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":["disc = checkpoint.discriminator\n","disc.summary()"],"metadata":{"id":"81pDU2CZl5TE","executionInfo":{"status":"ok","timestamp":1670402875767,"user_tz":-330,"elapsed":81,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"colab":{"base_uri":"https://localhost:8080/"},"outputId":"c0c3b21f-3411-4581-a799-a29200464f77"},"execution_count":9,"outputs":[{"output_type":"stream","name":"stdout","text":["Model: \"sequential_1\"\n","_________________________________________________________________\n"," Layer (type)                Output Shape              Param #   \n","=================================================================\n"," conv2d (Conv2D)             (None, 14, 14, 64)        1664      \n","                                                                 \n"," leaky_re_lu_3 (LeakyReLU)   (None, 14, 14, 64)        0         \n","                                                                 \n"," dropout (Dropout)           (None, 14, 14, 64)        0         \n","                                                                 \n"," conv2d_1 (Conv2D)           (None, 7, 7, 128)         204928    \n","                                                                 \n"," leaky_re_lu_4 (LeakyReLU)   (None, 7, 7, 128)         0         \n","                                                                 \n"," dropout_1 (Dropout)         (None, 7, 7, 128)         0         \n","                                                                 \n"," flatten (Flatten)           (None, 6272)              0         \n","                                                                 \n"," dense_1 (Dense)             (None, 1)                 6273      \n","                                                                 \n","=================================================================\n","Total params: 212,865\n","Trainable params: 212,865\n","Non-trainable params: 0\n","_________________________________________________________________\n"]}]},{"cell_type":"code","source":["val = disc(sample)\n","print (val)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"41f9NtMvEEv_","executionInfo":{"status":"ok","timestamp":1670402875769,"user_tz":-330,"elapsed":56,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"outputId":"fcd51baf-6e53-4d39-bcb5-c2bfa3ff83f5"},"execution_count":10,"outputs":[{"output_type":"stream","name":"stdout","text":["tf.Tensor([[-0.5118373]], shape=(1, 1), dtype=float32)\n"]}]},{"cell_type":"markdown","source":["Now, we transfer the weights of discriminator to a new model, with the only difference being that the new model's last dense layer has 10 parameters instead of 1."],"metadata":{"id":"nc63XVrwEIfG"}},{"cell_type":"code","source":["new_model = Sequential()\n","\n","for i in range(len(disc.layers) - 1):\n","    new_model.add(disc.layers[i])\n","\n","# Freeze the layers \n","for layer in new_model.layers:\n","    layer.trainable = True    \n","    \n","new_model.add(Dense(10, activation='softmax'))\n","new_model.summary()"],"metadata":{"id":"zd_ernVsl5P0","executionInfo":{"status":"ok","timestamp":1670402875770,"user_tz":-330,"elapsed":55,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"colab":{"base_uri":"https://localhost:8080/"},"outputId":"a0c98203-03e7-42bd-c053-8b56abd7f73d"},"execution_count":11,"outputs":[{"output_type":"stream","name":"stdout","text":["Model: \"sequential_2\"\n","_________________________________________________________________\n"," Layer (type)                Output Shape              Param #   \n","=================================================================\n"," conv2d (Conv2D)             (None, 14, 14, 64)        1664      \n","                                                                 \n"," leaky_re_lu_3 (LeakyReLU)   (None, 14, 14, 64)        0         \n","                                                                 \n"," dropout (Dropout)           (None, 14, 14, 64)        0         \n","                                                                 \n"," conv2d_1 (Conv2D)           (None, 7, 7, 128)         204928    \n","                                                                 \n"," leaky_re_lu_4 (LeakyReLU)   (None, 7, 7, 128)         0         \n","                                                                 \n"," dropout_1 (Dropout)         (None, 7, 7, 128)         0         \n","                                                                 \n"," flatten (Flatten)           (None, 6272)              0         \n","                                                                 \n"," dense_2 (Dense)             (None, 10)                62730     \n","                                                                 \n","=================================================================\n","Total params: 269,322\n","Trainable params: 269,322\n","Non-trainable params: 0\n","_________________________________________________________________\n"]}]},{"cell_type":"code","source":["# view the initialized weights and bias of the last dense layer; weights are uniformly randomly generated \n","# and biases are all zeroes by default\n","new_model.layers[-1].weights"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"TNMbK2kvl5NL","outputId":"4de96750-ce14-44c3-8a3f-9c6cff53735b","executionInfo":{"status":"ok","timestamp":1670402875771,"user_tz":-330,"elapsed":32,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}}},"execution_count":12,"outputs":[{"output_type":"execute_result","data":{"text/plain":["[<tf.Variable 'dense_2/kernel:0' shape=(6272, 10) dtype=float32, numpy=\n"," array([[ 0.02414932, -0.00081145,  0.0191418 , ..., -0.01569872,\n","         -0.00864824,  0.01547791],\n","        [-0.0269337 , -0.0041248 , -0.00015487, ...,  0.02164961,\n","          0.02079669,  0.02357868],\n","        [-0.00663778,  0.00395533, -0.02442576, ..., -0.0149684 ,\n","          0.00136187,  0.01623679],\n","        ...,\n","        [ 0.02451091,  0.00414053,  0.03032985, ..., -0.00150685,\n","          0.00154587,  0.02238557],\n","        [-0.02899456,  0.02405351,  0.01781472, ..., -0.02032416,\n","         -0.02089991, -0.02621905],\n","        [-0.00332658,  0.01588948,  0.00463746, ...,  0.03030953,\n","          0.02239525,  0.00655846]], dtype=float32)>,\n"," <tf.Variable 'dense_2/bias:0' shape=(10,) dtype=float32, numpy=array([0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], dtype=float32)>]"]},"metadata":{},"execution_count":12}]},{"cell_type":"code","source":["import pandas as pd\n","# Load the data\n","X_train = pd.read_csv(\"/content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/ahds/csvTrainImages 60k x 784.csv\")\n","Y_train = pd.read_csv(\"/content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/ahds/csvTrainLabel 60k x 1.csv\")\n","X_test = pd.read_csv(\"/content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/ahds/csvTestImages 10k x 784.csv\")\n","Y_test = pd.read_csv(\"/content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/ahds/csvTestLabel 10k x 1.csv\")\n","Y_train = Y_train.values.reshape(Y_train.shape[0],).astype('float32')\n","Y_test = Y_test.values.reshape(Y_test.shape[0],).astype('float32')\n","X_train = X_train.values.reshape(X_train.shape[0], 28, 28, 1).astype('float32')\n","X_test = X_test.values.reshape(X_test.shape[0], 28, 28, 1).astype('float32')\n","np.random.seed(42)\n","trainsize = 30000  # number of training images we use, selected randomly\n","testsize = 9000\n","randtrain = np.random.choice(X_train.shape[0], trainsize, replace=False)\n","X_train, Y_train = X_train[randtrain], Y_train[randtrain]\n","randtest = np.random.choice(X_test.shape[0], testsize, replace=False)\n","X_test, Y_test = X_test[randtest], Y_test[randtest]\n","print(X_train.shape)\n","print(Y_train.shape)\n","print(X_test.shape)\n","print(Y_test.shape)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"smxH7tQLswXa","executionInfo":{"status":"ok","timestamp":1670402888057,"user_tz":-330,"elapsed":12315,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"outputId":"a536bc12-1d14-4c4d-c7c7-9ad646ac2922"},"execution_count":13,"outputs":[{"output_type":"stream","name":"stdout","text":["(30000, 28, 28, 1)\n","(30000,)\n","(9000, 28, 28, 1)\n","(9000,)\n"]}]},{"cell_type":"code","source":["train_images = X_train.reshape((trainsize, 28, 28, 1))\n","train_images = (X_train.astype('float32') - 127.5) / 127.5\n","\n","test_images = X_test.reshape((testsize, 28, 28, 1))\n","test_images = (test_images.astype('float32') - 127.5) / 127.5\n","\n","train_labels = to_categorical(Y_train)\n","test_labels = to_categorical(Y_test)\n","\n","new_model.compile(optimizer='sgd',\n","                  loss='categorical_crossentropy',\n","                  metrics=['accuracy'])\n","\n","filepath = \"/content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/GAN_NET_\"+\"AHDS\"+\"_best_weights.hdf5\"\n","early_stop = EarlyStopping(monitor='val_accuracy', patience=50,mode='max')\n","checkpoint = ModelCheckpoint(filepath, monitor='val_accuracy', verbose=1, save_best_only=True, mode='max')\n","callbacks_list = [checkpoint,early_stop]\n","\n","history = new_model.fit(train_images, train_labels, batch_size=100, epochs=20,\n","                        validation_data=(test_images, test_labels),callbacks = callbacks_list)\n","# Freeze the layers \n","# for layer in new_model.layers:\n","#     layer.trainable = True \n","# new_model.compile(optimizer='sgd',\n","#                   loss='categorical_crossentropy',\n","#                   metrics=['accuracy'])\n","# history1 = new_model.fit(train_images, train_labels, batch_size=100, epochs=10,\n","#                         validation_data=(test_images, test_labels),callbacks = callbacks_list)"],"metadata":{"id":"rXvhLrYjoTEC","colab":{"base_uri":"https://localhost:8080/"},"outputId":"3d7b4889-47a2-45d9-ebc2-2036ec53e558","executionInfo":{"status":"ok","timestamp":1670402930106,"user_tz":-330,"elapsed":41153,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}}},"execution_count":14,"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 1/20\n","292/300 [============================>.] - ETA: 0s - loss: 0.4317 - accuracy: 0.8873\n","Epoch 1: val_accuracy improved from -inf to 0.96122, saving model to /content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/GAN_NET_AHDS_best_weights.hdf5\n","300/300 [==============================] - 5s 13ms/step - loss: 0.4237 - accuracy: 0.8892 - val_loss: 0.1442 - val_accuracy: 0.9612\n","Epoch 2/20\n","299/300 [============================>.] - ETA: 0s - loss: 0.1209 - accuracy: 0.9664\n","Epoch 2: val_accuracy improved from 0.96122 to 0.96867, saving model to /content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/GAN_NET_AHDS_best_weights.hdf5\n","300/300 [==============================] - 2s 6ms/step - loss: 0.1207 - accuracy: 0.9665 - val_loss: 0.1081 - val_accuracy: 0.9687\n","Epoch 3/20\n","293/300 [============================>.] - ETA: 0s - loss: 0.0949 - accuracy: 0.9734\n","Epoch 3: val_accuracy improved from 0.96867 to 0.97278, saving model to /content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/GAN_NET_AHDS_best_weights.hdf5\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0953 - accuracy: 0.9734 - val_loss: 0.0927 - val_accuracy: 0.9728\n","Epoch 4/20\n","291/300 [============================>.] - ETA: 0s - loss: 0.0796 - accuracy: 0.9770\n","Epoch 4: val_accuracy improved from 0.97278 to 0.97544, saving model to /content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/GAN_NET_AHDS_best_weights.hdf5\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0802 - accuracy: 0.9768 - val_loss: 0.0821 - val_accuracy: 0.9754\n","Epoch 5/20\n","294/300 [============================>.] - ETA: 0s - loss: 0.0736 - accuracy: 0.9785\n","Epoch 5: val_accuracy improved from 0.97544 to 0.97744, saving model to /content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/GAN_NET_AHDS_best_weights.hdf5\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0739 - accuracy: 0.9784 - val_loss: 0.0764 - val_accuracy: 0.9774\n","Epoch 6/20\n","292/300 [============================>.] - ETA: 0s - loss: 0.0667 - accuracy: 0.9807\n","Epoch 6: val_accuracy improved from 0.97744 to 0.97767, saving model to /content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/GAN_NET_AHDS_best_weights.hdf5\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0667 - accuracy: 0.9806 - val_loss: 0.0742 - val_accuracy: 0.9777\n","Epoch 7/20\n","290/300 [============================>.] - ETA: 0s - loss: 0.0624 - accuracy: 0.9827\n","Epoch 7: val_accuracy improved from 0.97767 to 0.97878, saving model to /content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/GAN_NET_AHDS_best_weights.hdf5\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0625 - accuracy: 0.9826 - val_loss: 0.0693 - val_accuracy: 0.9788\n","Epoch 8/20\n","293/300 [============================>.] - ETA: 0s - loss: 0.0584 - accuracy: 0.9829\n","Epoch 8: val_accuracy improved from 0.97878 to 0.98056, saving model to /content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/GAN_NET_AHDS_best_weights.hdf5\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0590 - accuracy: 0.9827 - val_loss: 0.0660 - val_accuracy: 0.9806\n","Epoch 9/20\n","294/300 [============================>.] - ETA: 0s - loss: 0.0558 - accuracy: 0.9834\n","Epoch 9: val_accuracy did not improve from 0.98056\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0555 - accuracy: 0.9836 - val_loss: 0.0641 - val_accuracy: 0.9801\n","Epoch 10/20\n","290/300 [============================>.] - ETA: 0s - loss: 0.0531 - accuracy: 0.9852\n","Epoch 10: val_accuracy improved from 0.98056 to 0.98089, saving model to /content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/GAN_NET_AHDS_best_weights.hdf5\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0533 - accuracy: 0.9851 - val_loss: 0.0641 - val_accuracy: 0.9809\n","Epoch 11/20\n","290/300 [============================>.] - ETA: 0s - loss: 0.0493 - accuracy: 0.9857\n","Epoch 11: val_accuracy improved from 0.98089 to 0.98200, saving model to /content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/GAN_NET_AHDS_best_weights.hdf5\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0496 - accuracy: 0.9856 - val_loss: 0.0608 - val_accuracy: 0.9820\n","Epoch 12/20\n","292/300 [============================>.] - ETA: 0s - loss: 0.0494 - accuracy: 0.9849\n","Epoch 12: val_accuracy did not improve from 0.98200\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0493 - accuracy: 0.9848 - val_loss: 0.0589 - val_accuracy: 0.9820\n","Epoch 13/20\n","298/300 [============================>.] - ETA: 0s - loss: 0.0480 - accuracy: 0.9851\n","Epoch 13: val_accuracy improved from 0.98200 to 0.98256, saving model to /content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/GAN_NET_AHDS_best_weights.hdf5\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0478 - accuracy: 0.9852 - val_loss: 0.0581 - val_accuracy: 0.9826\n","Epoch 14/20\n","296/300 [============================>.] - ETA: 0s - loss: 0.0467 - accuracy: 0.9864\n","Epoch 14: val_accuracy improved from 0.98256 to 0.98300, saving model to /content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/GAN_NET_AHDS_best_weights.hdf5\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0465 - accuracy: 0.9865 - val_loss: 0.0549 - val_accuracy: 0.9830\n","Epoch 15/20\n","294/300 [============================>.] - ETA: 0s - loss: 0.0429 - accuracy: 0.9873\n","Epoch 15: val_accuracy did not improve from 0.98300\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0428 - accuracy: 0.9874 - val_loss: 0.0548 - val_accuracy: 0.9827\n","Epoch 16/20\n","290/300 [============================>.] - ETA: 0s - loss: 0.0427 - accuracy: 0.9878\n","Epoch 16: val_accuracy improved from 0.98300 to 0.98411, saving model to /content/drive/MyDrive/Colab Notebooks/SIFT+GAN handwriting/GAN_NET_AHDS_best_weights.hdf5\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0430 - accuracy: 0.9877 - val_loss: 0.0533 - val_accuracy: 0.9841\n","Epoch 17/20\n","295/300 [============================>.] - ETA: 0s - loss: 0.0411 - accuracy: 0.9875\n","Epoch 17: val_accuracy did not improve from 0.98411\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0411 - accuracy: 0.9875 - val_loss: 0.0525 - val_accuracy: 0.9827\n","Epoch 18/20\n","291/300 [============================>.] - ETA: 0s - loss: 0.0411 - accuracy: 0.9878\n","Epoch 18: val_accuracy did not improve from 0.98411\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0410 - accuracy: 0.9877 - val_loss: 0.0529 - val_accuracy: 0.9829\n","Epoch 19/20\n","293/300 [============================>.] - ETA: 0s - loss: 0.0383 - accuracy: 0.9885\n","Epoch 19: val_accuracy did not improve from 0.98411\n","300/300 [==============================] - 2s 5ms/step - loss: 0.0383 - accuracy: 0.9885 - val_loss: 0.0503 - val_accuracy: 0.9840\n","Epoch 20/20\n","300/300 [==============================] - ETA: 0s - loss: 0.0383 - accuracy: 0.9884\n","Epoch 20: val_accuracy did not improve from 0.98411\n","300/300 [==============================] - 2s 6ms/step - loss: 0.0383 - accuracy: 0.9884 - val_loss: 0.0502 - val_accuracy: 0.9840\n"]}]},{"cell_type":"code","source":["accuracy = history.history['accuracy'] \n","val_accuracy = history.history['val_accuracy'] \n","plt.figure()\n","plt.plot(accuracy, label='accuracy')\n","plt.plot(val_accuracy, label = 'val_accuracy')\n","plt.xlabel('Epoch')\n","plt.ylabel('Accuracy')\n","plt.ylim([0.3, 1])\n","plt.legend(loc='lower right')\n","plt.show()"],"metadata":{"id":"vaH4emaHWahX","colab":{"base_uri":"https://localhost:8080/","height":283},"executionInfo":{"status":"ok","timestamp":1670402948545,"user_tz":-330,"elapsed":6,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"outputId":"663a95cb-e218-41af-c719-945fef225388"},"execution_count":17,"outputs":[{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":["loss = history.history['loss'] \n","val_loss = history.history['val_loss'] \n","\n","plt.figure()\n","plt.plot(loss, label='loss')\n","plt.plot(val_loss, label = 'val_loss')\n","plt.xlabel('Epoch')\n","plt.ylabel('Loss')\n","plt.legend(loc='upper right')\n","plt.show()"],"metadata":{"id":"u2zZQ-MlTY6X","colab":{"base_uri":"https://localhost:8080/","height":279},"executionInfo":{"status":"ok","timestamp":1670402954748,"user_tz":-330,"elapsed":4,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"outputId":"c3740ad6-ab30-42c1-bdc2-2825bbe098d9"},"execution_count":18,"outputs":[{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"iVBORw0KGgoAAAANSUhEUgAAAYgAAAEGCAYAAAB/+QKOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dfXhcZZ3/8fd3JpOZpMmkadNO+kiLlucqYEBdBZ8QgVUKKhZEBVdlQVBcXX+y4irLyrUK17r+dLsiyw9FF4QuiHYVRUUUWQUboLQUBErpQ9LH9CFpm+fJ9/fHOUmn6aRNmpxMkvm8rutc55z73Gfmm2k639z3fe5zzN0RERHpL1boAEREZGxSghARkbyUIEREJC8lCBERyUsJQkRE8iopdAAjpaamxufNm1foMERExpUnn3yyyd2n5Ts2YRLEvHnzqK+vL3QYIiLjipmtH+iYuphERCQvJQgREclLCUJERPKaMGMQIlKcurq6aGhooL29vdChjGmpVIrZs2eTSCQGfY4ShIiMaw0NDVRWVjJv3jzMrNDhjEnuzo4dO2hoaGD+/PmDPk9dTCIyrrW3tzN16lQlh0MwM6ZOnTrkVpYShIiMe0oOh3ckn1HRJ4jmti6++ZsXeWbj7kKHIiIyphR9ggD45m9eYvm6nYUOQ0TGqYqKikKHEImiTxDpVAmpRIwtzboCQkQkV9EnCDOjNp1i656OQociIuOcu/P5z3+ek046iYULF3LvvfcCsHnzZs4880xOPvlkTjrpJP7whz+QzWa5/PLL++r+27/9W4GjP5gucwWmp1NsbVELQmS8+6f/Wc1zm1pG9DVPmJnmK+85cVB1f/zjH7NixQqeeeYZmpqaOO200zjzzDO5++67ede73sX1119PNpultbWVFStW0NjYyLPPPgvA7t1jbxy06FsQABklCBEZAY899hiXXHIJ8XicTCbDW97yFpYvX85pp53G9773PW644QZWrVpFZWUlRx99NGvXruVTn/oUv/zlL0mn04UO/yBqQQC16SS/bmnH3XW5nMg4Nti/9EfbmWeeyaOPPsrPf/5zLr/8cj772c/ykY98hGeeeYaHHnqIW2+9laVLl3LHHXcUOtQDqAVB0IJo7+qhpb270KGIyDh2xhlncO+995LNZtm+fTuPPvoop59+OuvXryeTyfCJT3yCj3/84zz11FM0NTXR09PD+973Pr761a/y1FNPFTr8g6gFQTAGAbC1pZ2qssHfp0REJNeFF17In/70J1772tdiZtx8883U1tZy5513csstt5BIJKioqOAHP/gBjY2NfPSjH6WnpweAf/mXfylw9AdTggBqcxLEMZnKAkcjIuPN3r17geCqyFtuuYVbbrnlgOOXXXYZl1122UHnjcVWQ65Iu5jM7Bwze8HM1pjZdYeo9z4zczOryyn7h/C8F8zsXVHGmUknATQXQkQkR2QtCDOLA0uAdwINwHIzW+buz/WrVwlcCzyRU3YCcDFwIjAT+I2ZHePu2ShizYQtiG2aCyEi0ifKFsTpwBp3X+vuncA9wKI89f4Z+DqQ++f7IuAed+9w91eANeHrRSKViFNVltClriIiOaJMELOAjTn7DWFZHzM7FZjj7j8f6rnh+VeYWb2Z1W/fvn1YwWbSSXUxiYjkKNhlrmYWA74BfO5IX8Pdb3P3OnevmzZt2rDiyeh2GyIiB4jyKqZGYE7O/uywrFclcBLwu3ByWi2wzMzOH8S5Iy6TTvHS1qYo30JEZFyJsgWxHFhgZvPNrJRg0HlZ70F3b3b3Gnef5+7zgMeB8929Pqx3sZklzWw+sAD4c4SxUptOsX1vB9kej/JtRETGjcgShLt3A9cADwHPA0vdfbWZ3Ri2Eg517mpgKfAc8Evg6qiuYOqVSSfJ9jg79qmbSUSic6hnR6xbt46TTjppFKM5tEgnyrn7g8CD/cq+PEDdt/bbvwm4KbLg+umbTd3cwfTK1Gi9rYjImKWZ1KHc2dQLqSpwNCJyRH5xHWxZNbKvWbsQzv3agIevu+465syZw9VXXw3ADTfcQElJCY888gi7du2iq6uLr371qyxalO8q/4G1t7dz1VVXUV9fT0lJCd/4xjd429vexurVq/noRz9KZ2cnPT093H///cycOZMPfOADNDQ0kM1m+cd//EcWL148rB8blCD69E6W26K5ECIyBIsXL+Yzn/lMX4JYunQpDz30EJ/+9KdJp9M0NTXxhje8gfPPP39Id4tesmQJZsaqVav4y1/+wtlnn82LL77IrbfeyrXXXsull15KZ2cn2WyWBx98kJkzZ/LznwczBpqbm0fkZ1OCCNVUlBIz2KYEITJ+HeIv/aiccsopbNu2jU2bNrF9+3aqq6upra3l7/7u73j00UeJxWI0NjaydetWamtrB/26jz32GJ/61KcAOO644zjqqKN48cUXeeMb38hNN91EQ0MD733ve1mwYAELFy7kc5/7HF/4whd497vfzRlnnDEiP5tu9x0qiceoqUiytUWD1CIyNBdddBH33Xcf9957L4sXL+auu+5i+/btPPnkk6xYsYJMJkN7+8j88fnBD36QZcuWUVZWxnnnncdvf/tbjjnmGJ566ikWLlzIl770JW688cYReS+1IHJk0il1MYnIkC1evJhPfOITNDU18fvf/56lS5cyffp0EokEjzzyCOvXrx/ya55xxhncddddvP3tb+fFF19kw4YNHHvssaxdu5ajjz6aT3/602zYsIGVK1dy3HHHMWXKFD70oQ8xefJkbr/99hH5uZQgcmTSKRp2tRY6DBEZZ0488UT27NnDrFmzmDFjBpdeeinvec97WLhwIXV1dRx33HFDfs1PfvKTXHXVVSxcuJCSkhK+//3vk0wmWbp0KT/84Q9JJBLU1tbyxS9+keXLl/P5z3+eWCxGIpHgO9/5zoj8XOY+MSaG1dXVeX19/bBe4/oHVvHgqs08/eWzRygqEYna888/z/HHH1/oMMaFfJ+VmT3p7nX56msMIkcmnWJXaxcd3ZHOyRMRGRfUxZSjdy7EtpYO5kwpL3A0IjJRrVq1ig9/+MMHlCWTSZ544okBzigMJYgc08Mny21taVeCEBlH3H1IcwwKbeHChaxYsWJU3/NIhhPUxZSjtqp3NrUudRUZL1KpFDt27DiiL8Bi4e7s2LGDVGpotxFSCyJHpnL/7TZEZHyYPXs2DQ0NDPehYRNdKpVi9uzZQzpHCSLH5PIEpSUxJQiRcSSRSDB//vxChzEhqYsph5mRSSeVIEREUII4SKZSs6lFREAJ4iCZqhTbNEgtIqIE0V+mMqUuJhERIk4QZnaOmb1gZmvM7Lo8x680s1VmtsLMHjOzE8LyeWbWFpavMLNbo4wzVyadZF9nlj3tXaP1liIiY1JkVzGZWRxYArwTaACWm9kyd38up9rd7n5rWP984BvAOeGxl9395KjiG0juXIjKVGK0315EZMyIsgVxOrDG3de6eydwD3DAM/fcvSVndxJQ8Jku0zUXQkQEiDZBzAI25uw3hGUHMLOrzexl4Gbg0zmH5pvZ02b2ezPL+3gkM7vCzOrNrH6kJsnsb0EoQYhIcSv4ILW7L3H3VwFfAL4UFm8G5rr7KcBngbvNLJ3n3Nvcvc7d66ZNmzYi8Uyv7L0fk65kEpHiFmWCaATm5OzPDssGcg9wAYC7d7j7jnD7SeBl4JiI4jzApGQJlckStSBEpOhFmSCWAwvMbL6ZlQIXA8tyK5jZgpzdvwZeCsunhYPcmNnRwAJgbYSxHiBTpUtdRUQiu4rJ3bvN7BrgISAO3OHuq83sRqDe3ZcB15jZWUAXsAu4LDz9TOBGM+sCeoAr3X1nVLH2l0knNZtaRIpepDfrc/cHgQf7lX05Z/vaAc67H7g/ytgOJVOZ4olXRi0fiYiMSQUfpB6LMlUptu1pp6en4FfdiogUjBJEHpnKJF1ZZ2drZ6FDEREpGCWIPDQXQkRECSKv6WklCBERJYg8Mmk9m1pERAkij/2zqdWCEJHipQSRRyIeo6aiVAlCRIqaEsQAMumUuphEpKgpQQwgSBBqQYhI8VKCGEAmnVSCEJGipgQxgEw6RdPeTrqyPYUORUSkIJQgBtB7qeu2PRqHEJHipAQxgFpNlhORIqcEMYDp6WAuxDYlCBEpUkoQA+jtYtrSrAQhIsVJCWIAU8pLScSNrRqDEJEipQQxgFjMmF6ZYqtaECJSpCJNEGZ2jpm9YGZrzOy6PMevNLNVZrbCzB4zsxNyjv1DeN4LZvauKOMcyPR0kq17lCBEpDhFliDMLA4sAc4FTgAuyU0AobvdfaG7nwzcDHwjPPcE4GLgROAc4D/C1xtVtbrdhogUsShbEKcDa9x9rbt3AvcAi3IruHtLzu4koPcZn4uAe9y9w91fAdaErzeqMml1MYlI8YoyQcwCNubsN4RlBzCzq83sZYIWxKeHeO4VZlZvZvXbt28fscB7ZdIp9nR0s6+je8RfW0RkrCv4ILW7L3H3VwFfAL40xHNvc/c6d6+bNm3aiMeWSeu5ECJSvKJMEI3AnJz92WHZQO4BLjjCcyOhJ8uJSDGLMkEsBxaY2XwzKyUYdF6WW8HMFuTs/jXwUri9DLjYzJJmNh9YAPw5wljz2n8/JrUgRKT4lET1wu7ebWbXAA8BceAOd19tZjcC9e6+DLjGzM4CuoBdwGXhuavNbCnwHNANXO3u2ahiHUhvF5NmU4tIMYosQQC4+4PAg/3Kvpyzfe0hzr0JuCm66A6vMpVgUmlcXUwiUpQKPkg91mXSKU2WE5GipARxGNPTSc2FEJGipARxGLVqQYhIkVKCOIxMeLsNdz98ZRGRCUQJ4jAy6RSd3T3sbu0qdCgiIqNKCeIw+ibLqZtJRIqMEsRhaC6EiBQrJYjD6JtNrbkQIlJklCAOY3pvC0I37BORIqMEcRjJkjjV5Qnd0VVEio4SxCBk9GQ5ESlCShCDECQItSBEpLgoQQxCrRKEiBQhJYhByKSTNO3toDvbU+hQRERGjRLEIExPp+hxaNrbWehQRERGjRLEINT2PXpU3UwiUjwiTRBmdo6ZvWBma8zsujzHP2tmz5nZSjN72MyOyjmWNbMV4bKs/7mjqXeynOZCiEgxieyJcmYWB5YA7wQagOVmtszdn8up9jRQ5+6tZnYVcDOwODzW5u4nRxXfUGSqgsly25QgRKSIRNmCOB1Y4+5r3b0TuAdYlFvB3R9x99Zw93FgdoTxHLGpk5LEY6YWhIgUlSgTxCxgY85+Q1g2kI8Bv8jZT5lZvZk9bmYX5DvBzK4I69Rv3759+BEPIB4zplUkNVlORIpKZF1MQ2FmHwLqgLfkFB/l7o1mdjTwWzNb5e4v557n7rcBtwHU1dVF+kSfTJXmQohIcRlUC8LMJplZLNw+xszON7PEYU5rBObk7M8Oy/q/9lnA9cD57t73J7q7N4brtcDvgFMGE2tUMpVJJQgRKSqD7WJ6lKDLZxbwK+DDwPcPc85yYIGZzTezUuBi4ICrkczsFOC7BMlhW055tZklw+0a4E1A7uD2qKut0v2YRKS4DDZBWDiY/F7gP9z9IuDEQ53g7t3ANcBDwPPAUndfbWY3mtn5YbVbgArgv/tdzno8UG9mzwCPAF/rd/XTqMukUzS3ddHelS1kGCIio2awYxBmZm8ELiUYTAaIH+4kd38QeLBf2Zdzts8a4Lw/AgsHGduomF4ZXOq6taWdo6ZOKnA0IiLRG2wL4jPAPwAPhK2Aown+si8atVW9s6nVzSQixWFQLQh3/z3we4BwsLrJ3T8dZWBjjWZTi0ixGexVTHebWdrMJgHPAs+Z2eejDW1syVT2PptaCUJEisNgu5hOcPcW4AKCyWzzCa5kKhrpshJSiZgudRWRojHYBJEI5z1cACxz9y4g0olpY42ZkUmn2KIxCBEpEoNNEN8F1gGTgEfDu662RBXUWKVHj4pIMRlUgnD3b7n7LHc/zwPrgbdFHNuYowQhIsVksIPUVWb2jd4b45nZvxK0JopK7+023Iuqd01EitRgu5juAPYAHwiXFuB7UQU1VtVWpWjv6qGlvbvQoYiIRG6wM6lf5e7vy9n/JzNbEUVAY9n0nEePVpUd7l6FIiLj22BbEG1m9ubeHTN7E9AWTUhjl55NLSLFZLAtiCuBH5hZVbi/C7gsmpDGrkw6uB/TlmYlCBGZ+AZ7q41ngNeaWTrcbzGzzwArowxurOm93ca2PZoLISIT35AeOeruLeGMaoDPRhDPmJZKxKkqS6iLSUSKwnCeSW0jFsU4kkkn1cUkIkVhOAmiKCcDZNIptqqLSUSKwCHHIMxsD/kTgQFlkUQ0xmXSKdZsayp0GCIikTtkC8LdK909nWepdPfDDnCb2Tlm9oKZrTGz6/Ic/6yZPWdmK83s4fAeT73HLjOzl8JlzFwxlUkn2bang2xPUTagRKSIDKeL6ZDMLA4sAc4FTgAuMbMT+lV7Gqhz99cA9wE3h+dOAb4CvB44HfiKmVVHFetQ1KZTZHucHfvUzSQiE1tkCYLgi32Nu691907gHmBRbgV3f8TdW8Pdx4HZ4fa7gF+7+0533wX8GjgnwlgHrW82dbMShIhMbFEmiFnAxpz9hrBsIB8jeBjRoM81syt6byC4ffv2YYY7OBnNphaRIhFlghg0M/sQUAfcMpTz3P02d69z97pp06ZFE1w/fbfb2KMEISITW5QJohGYk7M/Oyw7gJmdBVwPnO/uHUM5txBqKkqJGWzVXAgRmeCiTBDLgQVmNt/MSoGLgWW5FczsFIKn1Z3v7ttyDj0EnG1m1eHg9NlhWcGVxGPUVCTZqkePisgEN9ib9Q2Zu3eb2TUEX+xx4A53X21mNwL17r6MoEupAvhvMwPY4O7nu/tOM/tngiQDcKO774wq1qEKnk2tFoSITGyRJQgAd38QeLBf2Zdzts86xLl3EDyoaMzJpJM07Cq6u52LSJEZE4PU400mndIdXUVkwlOCOAKZdIqd+zrp6M4WOhQRkcgoQRyB3ktdt2mgWkQmMCWIIzA9fLKcJsuJyESmBHEE9s+mVgtCRCYuJYgjUKvbbYhIEVCCOAKTyxOUlsSUIERkQlOCOAJmRiadVIIQkQlNCeIIZSo1m1pEJjYliCOUSad0mauITGhKEEcok06pi0lEJjQliCOUSSfZ15llT3tXoUMREYmEEsQR0lwIEZnolCCOUKbvdhvqZhKRiUkJ4ghlwttt6EomEZmolCCOkLqYRGSiizRBmNk5ZvaCma0xs+vyHD/TzJ4ys24ze3+/Y1kzWxEuy/qfW2iTkiVUJkt0JZOITFiRPVHOzOLAEuCdQAOw3MyWuftzOdU2AJcDf5/nJdrc/eSo4hsJ0zWbWkQmsCgfOXo6sMbd1wKY2T3AIqAvQbj7uvBYT4RxRKa2SnMhRGTiirKLaRawMWe/ISwbrJSZ1ZvZ42Z2wciGNjIylSmNQYjIhBVlC2K4jnL3RjM7Gvitma1y95dzK5jZFcAVAHPnzh31ADNVKbbtaaenx4nFbNTfX0QkSlG2IBqBOTn7s8OyQXH3xnC9FvgdcEqeOre5e527102bNm0YkT4J2e4hn5apTNKVdXa2dh75e4uIjFFRJojlwAIzm29mpcDFwKCuRjKzajNLhts1wJvIGbsYUdtfhNvfCb/8ArgP6dSMHhwkIhNYZAnC3buBa4CHgOeBpe6+2sxuNLPzAczsNDNrAC4Cvmtmq8PTjwfqzewZ4BHga/2ufho5046BN14Ny2+HPy0Z0qmZqt7Z1BqHEJGJJ9IxCHd/EHiwX9mXc7aXE3Q99T/vj8DCKGM7wFn/BLvXw6++BJPnwAmLBnVabwtCs6lFZCIay4PUoycWgwu/Cy2b4MdXQOVMmHPaYU+bXhncbkNdTCIyEelWG70SZXDJPVBZCz+6GHa+cvhT4jFqKkqVIERkQlKCyDWpBi69D3q64a6LoHXnYU+ZrrkQIjJBKUH0V7MALr47GJO490PQfegvf82mFpGJSgkin3lvgkX/Aev/F356zSEvf83ofkwiMkFpkHogr7kIdq+D334VqufB26/PW216ZYqmvZ10ZXtIxJVvRWTi0DfaoZzx93DKh+DRm+Hp/8pbpbZ3LsQejUOIyMSiBHEoZvDub8LRb4X/uRZefuSgKr1PllM3k4hMNEoQhxNPwAd+ADXHwNKPwNYDJ3Tr2dQiMlEpQQxGqgo+uDSYK3H3B2DPlr5DM6rKALjvyUbaOrOFilBEZMQpQQzW5DlBkmjdCXcvhs59AEyZVMo/nHscD/9lKxd994807m4rcKAiIiNDCWIoZp4M778DtqyE+z4GPUGL4W/f8ipu/0gd65paWfTvj/HnVw4/wU5EZKxTghiqY8+Bc2+GF38Bv7yub47EO47P8JOr/4rKVIIP/ufj3PXE+gIHKiIyPEoQR+L0T8Abr4E/3waPf6ev+NXTK/nJ1W/iTa+u4foHnuX6B1bR2T0uH7ctIqIEccTe+c9w/HvgoS/C8z/rK64qS3DH5adx5VtexV1PbODS2x+naa/mSIjI+KMEcaRiMbjwNph1Ktz/cVi5tG9MIh4zrjv3OP7vxSezsqGZ87/9GM82Nhc4YBGRoVGCGI7ScrjkXqh5Nfz4E7Dk9fDMvX3Pt1508izuv+qvAHjfd/7IT1cM+pHcIiIFF2mCMLNzzOwFM1tjZtflOX6mmT1lZt1m9v5+xy4zs5fC5bIo4xyWimlwxaNw0Z0QL4UHroAlp8OKuyHbzUmzqlj2qTfz2tmTufaeFXztF38h2zO0Z1+LiBRCZAnCzOLAEuBc4ATgEjM7oV+1DcDlwN39zp0CfAV4PXA68BUzq44q1mGLxeDEC+DKx2DxfwUti59cBf/+Onjqh9SUxfivj7+eS18/l1t//zIfu3M5zW1dhY5aROSQomxBnA6scfe17t4J3AMc8LBnd1/n7iuB/pf6vAv4tbvvdPddwK+BcyKMdWTEYsHA9d/+AS7+EaQmw7Jr4NunUrriTm56z7HcdOFJPPZSExcs+V/WbNtb6IhFRAYUZYKYBWzM2W8Iy0bsXDO7wszqzax++/btRxzoiDOD486DK34HH/xvmDQNfvYZ+NYpXBr7NT/6m1NoaeviwiX/y8PPby10tCIieY3rQWp3v83d69y9btq0aYUO52BmcMzZ8PGH4UP3Q3om/PxznPbTt/PwGS/w6qklfPwH9Sx5ZA1+iIcSiYgUQpQJohGYk7M/OyyL+tyxxwxefRZ87Ffw4Z9A9VFM/t31/Ljzk/zrnD/yrYdWsWjJ//Kr1Vvo0QC2iIwRFtVfrmZWArwIvIPgy3058EF3X52n7veBn7n7feH+FOBJ4NSwylPA69x9wJsc1dXVeX19/Yj+DJFxh3V/gN99HdY/RltyKvdl38rd++romXYiV79jAX+9cAbxmBU6UhGZ4MzsSXevy3ssyq4NMzsP+CYQB+5w95vM7Eag3t2XmdlpwANANdAObHH3E8Nz/wb4YvhSN7n79w71XuMqQeRa9xg89m/4y49gnmVDbBYPdL6eFZVv49x3vI0LT5mlR5mKSGQKliBG07hNEL32NcFzP8VXPwDrHsNwXuiZzaOlZzDtjZdwzplvJpWIFzpKEZlglCDGmz1b8ed+QvPye5nc9CQALzCfllf9NSed/VHKMq8ucIAiMlEoQYxj3tzAukd/RPfK+1nQ9TwAWytOoOq0D5B67fuDBxmJiBwhJYgJYtXqVTz/mzs5dsfDvDa2FoDuGXWUvOZ9MO/NMO1YKEkWOEoRGU+UICaYZxubufdXv6fy5Z/xnvjjHG/Bw4k8VoLVHAOZEyFzEtSeFKwrMsGltiIi/ShBTFAvbd3Df/zuZZ5e8RQn2VpOjG/gdalNHMN6Jndt21+xvCZIGrULg4SROVGtDREBlCAmvG0t7Ty9cTerGppZ2djMqobd9LTu4vjYBk6Mb+T15Zs4IbaR2o5XKOkJH14UK4GaY/YnjPQsKK+GspwlWRXcX0pEJiwliCLj7jTsamNVYzMrG5pZ1biblQ3NtLZ3MM+2sLBkI2+q3MprSjYyp3Mt5e1bBnglg7LJByaNsin99sOl+iiong8lpaP6s4rI8BwqQZSMdjASPTNjzpRy5kwp57yFM4Agaazf0drXwri/sZl/amxhb0c3afYxN7mP0zJwSo1z/ORujirvpLSzGdp27V9ad8KONcF2e54n5FkcqucFLZOaBeE63C6fMrofgogMmxJEkTAz5tVMYl7NJM5/7UwAenqcV3bsY2XDbp5cv4s/rdvF95/eg3vw2NQTZqSpm1dN3TFTqJtXTSad2v+CPdkgSbTtgtYdsGsdNL0YLi/Byw9DtnN//fKa/Ilj8lyIaQKgyFikLiY5QHNbF09v2EX9ul3Ur9/Jio27ae8KHtcxZ0oZdUdN4XVHVXPavCksmF5BbKD7RfVkYff6IFnkJo6ml6C1aX+9eBKmvipIFJUzgjveVs6A9AyonBmsU5N1FZZIRDQGIUesK9vD6k0t1K/byZPrd7F83S6a9gYD3elUCaceVU3dUdUcW5tm7pRy5k4pp6z0MC2C1p0HJ47mBtizKWiN9JcoHzh59K7LayCROvhcETkkJQgZMe7Ohp2tfS2M+nW7eKnfk/GmVSaZO6Wco8JxkLlTypk7NdifVpnEDtUa6O6APZuhZXOQMFo2h/ubDlzndl/1iichVZWzpPvt9y6TDy6LD2FwPV/8pZUQV4+tjD9KEBKp5rYu1jXtY/3OVjbubGXDjlbW79zHxp1tbGpuI/dXLJWIMae6nKOm5iSPKeXMqi5jRlUZ6VTJoRMIBLdLb90ZJpBw6R04H3DZnT+pjBSLwaTpYasmZ0nPgMraoKVTWRtc8aXuMhlDlCCkYDq6szTuamNDmDzW72hlw879S2tn9oD6k0rjzJhcxoyqFDOrypgxef96RlWKGVVlTEoe4V/qXe35E0d7M/R0H1w/7/+NPGXeA227g4S1Z8v+Vk9bnseXlKQOTBjpcF1ek6elkw5aJpqLIhHSZa5SMMmSOEdPq+DoaRUHHXN3duzrZMPOVjbtbmPz7nY2NQfrzc1t/GXLHpr2dhz0PZ1OlTAzTCIzJpcxM0wcM6pS1IZLeWmeX+1EKlgqMxH9tIcOaGMAAA4eSURBVP10tcPenITRu7RsDhLJ5hXwwi+gu+0QL2L9usrydI+lqiBRFnSxlZQG63jp/u28ZUmIJ8LyhFo1kpcShBSMmVFTkaSmIsmpc6vz1uns7mFrSzubm4OksWn3getnGprZue/grqOqskRfwphRlaI2XXbgflWKylQi2h8wkQrmhVTPG7iOe3i58M6cVk3LobvLdr4CHWGdjpYRCDScEFlRG7ZuepcZwX28KmcESbWiVhcCFBklCBnTSktifZP+BtLelWVLc5BEtrS0Beve/eZ2nm1s6bvyKldFsqQvWdRUJKkqS1BdXsrk8kS4lFJdnmByWSmTJyWoTA5ifGSorHe2+uQjO793Pkp3ezDAn+3st+6A7s5g/6CynHXbrqBVs2dLcFXZ3i35u93KqvMnktJJgIUtkXBtsZwy8pTl1Isng9dIVgTr0spgnShT66aAIk0QZnYO8H8JHjl6u7t/rd/xJPAD4HXADmCxu68zs3nA88ALYdXH3f3KKGOV8SuViPdNAhxIR3eWbS0dfS2R3ASyubmNV5r20dzaxZ6OPF+KoXjMmFyWoKo8SCTV5QmqyoL19HSSmZPLmDm5jFmTy5hWkRx4jshIisWjmaXe0xO0avZshj1bw/WWIHEckEi2Qk/XyL9/L4tBaW/SCNfJMHn0lVWEXWylYbdZ6cHbJXnKcrctFiRbz+ase/rtH6K8JAWTaqB8arAurZgQiS2yBGFmcWAJ8E6gAVhuZsvc/bmcah8Ddrn7q83sYuDrwOLw2MvufnJU8UlxSZbED9sSgWDeR3NbF7tbu9jd2snu1i52tXbS3Basg/Iudrd1sml3O89v3sPOfZ20dR042J6IG7XhQPusMHEES6pv/4gH20dDLBZ80U2qCe4CPJDeRNK5D/BwYD9c524fsO7pV68naN107oWOvcFrde7Nv9+5Lyhr2XTgfnd78GU9VpSkgmTRmzAmTQsuRJg0NVzX5Kyn9muBwUEtrwKJ8jf0dGCNu68FMLN7gEVAboJYBNwQbt8H/LuNeBteZPAS8VjfuMhQtLR3sWl3G5t2t9G4u71ve9PuNp54ZSdbWtrJ9hw42l5VlghbHCkqkiXEzDAz4jGImRGLGTELt/uWoCWTW8/MSCViTCotYVKyhEml8WCdDNe95ck4pfHYyHaT5SaSQuvJQrYr6DLLdu3vVuvd7s5XHnazQfCzWDxolR2wHkR5V1vwXPnWppz1jv37O9YE+137hvlDDpBEZtXBxx4a5msfLMoEMQvYmLPfALx+oDru3m1mzcDU8Nh8M3saaAG+5O5/6P8GZnYFcAXA3LlzRzZ6kSFIpxKkaxMcV5vOe7w728O2PR1hAgkG2XsTSMOuNtq6smR7HHfocSfb4/R4cKVX1p2ecL/HPVh6craHcKV6ScwOSCLlyRIqknHKEiWUlcYpT8QpK42TSsQpS8QpL42TKg22yxJxykpjfXX3l8WpKktQWlLgy3Fj4Zf1WB5IPyCR5CSQrjb6LqHubV0NZTs9M5Jwx2obdzMw1913mNnrgJ+Y2YnufsAlG+5+G3AbBPMgChCnyKCUxGN93Ux5LzgfBneno7uHfR3d7OvIsrejm9bO7nAd7nd0s68zG9bpZm9Htq/Ovo5udu7ror0rS1tnUN7e1UNntmfQMZhBpjIVdKFVlzMrbBnNqt4/LhP5VWPjQaIseI78OHmWfJQJohHI/RRmh2X56jSYWQlQBezwYPZeB4C7P2lmLwPHAJoJJ9JP0MUU/NU/9eDpJkesO9tDe3cPbZ1B4mjrCpfOLO1dWVp7yzq7adrbSePuNhp3tbGyYTe/fHYzXdkD/2arTJWEiaOMWdX7x2ZmVZdRkSzpaz31ruHA/aCl1Nuy2n/M3SktiZEuS1CZKqEylWBSaXzkrzgrQlEmiOXAAjObT5AILgY+2K/OMuAy4E/A+4Hfurub2TRgp7tnzexoYAGwNsJYRaSfkniMiniMiiMYTO/pcZr2dtAQJo39XWtBl9rydTtpaR/4irHhihlUpoKEkU7tTxzpsv37feuyBMmSWHDFLQbhuI9BX1ms96pcDOs9HlTFDMpLS5haUcqU8lJK4hNn5ntkCSIcU7gGeIjgMtc73H21md0I1Lv7MuD/AT80szXAToIkAnAmcKOZdQE9wJXunue+BSIyFsVixvR0iunp1ICTIPe0d/W1Olo7s31fusHVwfsH6Hu/kHO/uPu+oMMv7Y7uLHvau8Oli5b2rr79lrZgu2FXK3s2d9PS3sXeju78d1IZAdXlCaZWJJk6qZSayiQ1k0qD/YpSpk5KUlOxfz+SuTUjSPdiEpGi09Pj7OvspiVMKB1dPThBd1Xf2gm3c7qzCAp7wu3eY22dWZr2dtC0t5Md+zrYsbeTHXs7aQq3m9vyzxUpjceYWlFK+TC7xI6fkebbl5xyROfqXkwiIjliMQu7oBJAWeTv19ndw67Wzv1JZG9HXwJp2tNJe9fw5nDMqY7mZ1CCEBGJWGlJjEw6deBje8eBiTOaIiIiI0oJQkRE8lKCEBGRvJQgREQkLyUIERHJSwlCRETyUoIQEZG8lCBERCSvCXOrDTPbDqwfxkvUAE0jFE4UFN/wKL7hUXzDM5bjO8rdp+U7MGESxHCZWf1A9yMZCxTf8Ci+4VF8wzPW4xuIuphERCQvJQgREclLCWK/2wodwGEovuFRfMOj+IZnrMeXl8YgREQkL7UgREQkLyUIERHJq6gShJmdY2YvmNkaM7suz/Gkmd0bHn/CzOaNYmxzzOwRM3vOzFab2bV56rzVzJrNbEW4fHm04suJYZ2ZrQrf/6BnvFrgW+FnuNLMTh3F2I7N+WxWmFmLmX2mX51R/QzN7A4z22Zmz+aUTTGzX5vZS+E670ObzeyysM5LZnbZKMZ3i5n9Jfz3e8DMJg9w7iF/FyKM7wYza8z5NzxvgHMP+f89wvjuzYltnZmtGODcyD+/YQuevTrxFyAOvAwcDZQCzwAn9KvzSeDWcPti4N5RjG8GcGq4XQm8mCe+twI/K/DnuA6oOcTx84BfAAa8AXiigP/eWwgmARXsMwTOBE4Fns0puxm4Lty+Dvh6nvOmAGvDdXW4XT1K8Z0NlITbX88X32B+FyKM7wbg7wfx73/I/+9Rxdfv+L8CXy7U5zfcpZhaEKcDa9x9rbt3AvcAi/rVWQTcGW7fB7zDhvMk8SFw983u/lS4vQd4Hpg1Gu89whYBP/DA48BkM5tRgDjeAbzs7sOZXT9s7v4osLNfce7v2Z3ABXlOfRfwa3ff6e67gF8D54xGfO7+K3fvDncfB2aP9PsO1gCf32AM5v/7sB0qvvC74wPAj0b6fUdLMSWIWcDGnP0GDv4C7qsT/gdpBqaOSnQ5wq6tU4An8hx+o5k9Y2a/MLMTRzWwgAO/MrMnzeyKPMcH8zmPhosZ+D9moT/DjLtvDre3AJk8dcbK5/g3BC3CfA73uxCla8IusDsG6KIbC5/fGcBWd39pgOOF/PwGpZgSxLhgZhXA/cBn3L2l3+GnCLpMXgt8G/jJaMcHvNndTwXOBa42szMLEMMhmVkpcD7w33kOj4XPsI8HfQ1j8lpzM7se6AbuGqBKoX4XvgO8CjgZ2EzQjTMWXcKhWw9j/v9SMSWIRmBOzv7ssCxvHTMrAaqAHaMSXfCeCYLkcJe7/7j/cXdvcfe94faDQMLMakYrvvB9G8P1NuABgqZ8rsF8zlE7F3jK3bf2PzAWPkNga2+3W7jelqdOQT9HM7sceDdwaZjEDjKI34VIuPtWd8+6ew/wnwO8b6E/vxLgvcC9A9Up1Oc3FMWUIJYDC8xsfvgX5sXAsn51lgG9V4u8H/jtQP85RlrYX/n/gOfd/RsD1KntHRMxs9MJ/v1GM4FNMrPK3m2Cwcxn+1VbBnwkvJrpDUBzTnfKaBnwL7dCf4ah3N+zy4Cf5qnzEHC2mVWHXShnh2WRM7NzgP8DnO/urQPUGczvQlTx5Y5pXTjA+w7m/3uUzgL+4u4N+Q4W8vMbkkKPko/mQnCFzYsEVzdcH5bdSPAfASBF0C2xBvgzcPQoxvZmgq6GlcCKcDkPuBK4MqxzDbCa4IqMx4G/GuXP7+jwvZ8J4+j9DHNjNGBJ+BmvAupGOcZJBF/4VTllBfsMCRLVZqCLoB/8YwTjWg8DLwG/AaaEdeuA23PO/Zvwd3EN8NFRjG8NQf997+9h75V9M4EHD/W7MErx/TD83VpJ8KU/o3984f5B/99HI76w/Pu9v3M5dUf98xvuolttiIhIXsXUxSQiIkOgBCEiInkpQYiISF5KECIikpcShIiI5KUEITIEZpbtd8fYEbtLqJnNy70rqEihlRQ6AJFxps3dTy50ECKjQS0IkREQ3tv/5vD+/n82s1eH5fPM7LfhjeUeNrO5YXkmfNbCM+HyV+FLxc3sPy14JsivzKysYD+UFD0lCJGhKevXxbQ451izuy8E/h34Zlj2beBOd38NwU3vvhWWfwv4vQc3DTyVYDYtwAJgibufCOwG3hfxzyMyIM2kFhkCM9vr7hV5ytcBb3f3teFNF7e4+1QzayK4FURXWL7Z3WvMbDsw2907cl5jHsEzIBaE+18AEu7+1eh/MpGDqQUhMnJ8gO2h6MjZzqJxQikgJQiRkbM4Z/2ncPuPBHcSBbgU+EO4/TBwFYCZxc2sarSCFBks/XUiMjRl/R5C/0t3773UtdrMVhK0Ai4Jyz4FfM/MPg9sBz4all8L3GZmHyNoKVxFcFdQkTFDYxAiIyAcg6hz96ZCxyIyUtTFJCIieakFISIieakFISIieSlBiIhIXkoQIiKSlxKEiIjkpQQhIiJ5/X8RP/2ErTF73QAAAABJRU5ErkJggg==\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","source":["We can see that, using the tuned discriminator model, both training accuracy and testing are approximately 0.90 after 15 epochs."],"metadata":{"id":"kuIT4USFEZ7F"}},{"cell_type":"code","source":["# demonstration: predict the ith test digit\n","i = 160\n","prediction = new_model.predict(test_images[i].reshape(1,28,28,1))\n","\n","# get probability distribution and classification of the test digit\n","print(prediction)\n","print('prediction:', np.argmax(prediction))\n","\n","# draw the barplot\n","plt.figure()\n","plt.bar(np.arange(0,10).astype('str'), prediction[0,:])\n","plt.show()"],"metadata":{"id":"R1K4JKmEoVx7","colab":{"base_uri":"https://localhost:8080/","height":334},"executionInfo":{"status":"ok","timestamp":1670403123567,"user_tz":-330,"elapsed":7,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"outputId":"e666d5ce-f1b6-48e9-bf73-5112e3b2a631"},"execution_count":28,"outputs":[{"output_type":"stream","name":"stdout","text":["1/1 [==============================] - 0s 16ms/step\n","[[3.7691596e-05 1.2102574e-09 1.5884789e-04 3.0095021e-10 2.9043124e-05\n","  9.9952698e-01 2.4515398e-07 1.1675343e-05 1.0065869e-05 2.2544134e-04]]\n","prediction: 5\n"]},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":["# show the actual ith digit\n","print('actual label:', np.argmax(test_labels[i]))\n","plt.figure()\n","plt.imshow(test_images[i,:,:,0], cmap='gray')\n","plt.show()"],"metadata":{"id":"fhckSRTIoXEi","colab":{"base_uri":"https://localhost:8080/","height":282},"executionInfo":{"status":"ok","timestamp":1670403127157,"user_tz":-330,"elapsed":15,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"outputId":"80fb078d-874a-4d4c-ea3f-c583cec6a35c"},"execution_count":29,"outputs":[{"output_type":"stream","name":"stdout","text":["actual label: 5\n"]},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"iVBORw0KGgoAAAANSUhEUgAAAPsAAAD4CAYAAAAq5pAIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAOBklEQVR4nO3dXaxV9ZnH8d9PpmjAl8CYQWLNaIvENCZSJQYimThpio43Ui9qxUyOL8mpiqYkYzKmQyyJqREzdmK8IMFIyowdao0i0NRYJI3OXFA5iC+oU0QDKYgQY7RwYSrwzMVZmKOe9d+Hvdd+gef7SU723uvZa6/HHX+stdd/r/13RAjAqe+0fjcAoDcIO5AEYQeSIOxAEoQdSOJverkx25z6B7osIjze8o727Lavtf0n27ts39fJawHoLrc7zm57kqSdkr4vaa+krZJuioi3C+uwZwe6rBt79isl7YqI9yPir5J+Len6Dl4PQBd1EvbzJf15zOO91bIvsT1se8T2SAfbAtChrp+gi4hVklZJHMYD/dTJnn2fpAvGPP5mtQzAAOok7FslXWz7ItuTJf1I0oZm2gLQtLYP4yPiiO27Jb0gaZKk1RHxVmOdAWhU20NvbW2Mz+xA13XlSzUATh6EHUiCsANJEHYgCcIOJEHYgSQIO5AEYQeSIOxAEoQdSIKwA0kQdiAJwg4kQdiBJAg7kARhB5Ig7EAShB1IgrADSRB2IAnCDiTR0ymbT1X2uD/m+YVOf8H3nHPOKdZvu+222tr06dOL695www3F+uuvv16s33///cX6rl27inX0Dnt2IAnCDiRB2IEkCDuQBGEHkiDsQBKEHUiCWVx7YPLkycX6rbfeWqzfe++9xfqsWbNOuKemHD58uFh/5JFHamvLly9vuBtI9bO4dvSlGtu7JR2SdFTSkYiY28nrAeieJr5B948R8VEDrwOgi/jMDiTRadhD0u9tb7M9PN4TbA/bHrE90uG2AHSg08P4BRGxz/bfSdpk+/8i4uWxT4iIVZJWSXlP0AGDoKM9e0Tsq24PSlon6commgLQvLbDbnuq7bOO35e0UNKOphoD0Ky2x9ltf0uje3Np9OPAf0fEz1usc0oexp9xxhnF+nPPPVesX3PNNR1t/5NPPqmt7djR2b+/5513XrHeaoz/2LFjtbVW/90vvvhisY7xNT7OHhHvS7qs7Y4A9BRDb0AShB1IgrADSRB2IAnCDiTBJa4NaHWJ6urVq4v10vCUJC1durRYf/7552trnf6U85QpU4r1xx57rFgv/cz1tm3biuvOnz+/WP/888+L9azqht7YswNJEHYgCcIOJEHYgSQIO5AEYQeSIOxAEoyzT1DpMtYtW7YU173ssvLFgXfddVexvnLlymK9pNV00q20+v9j9uzZxfr27dtra63G8OfMmVOst5pOOivG2YHkCDuQBGEHkiDsQBKEHUiCsANJEHYgiSYmdkzh0ksvra21Gkd/8skni/VW4+idjJV3+j2KVtveuXNnsb5x48ba2o033lhcd+HChcU64+wnhj07kARhB5Ig7EAShB1IgrADSRB2IAnCDiTBOPsELViwoLa2du3a4rrLli3raNu9/M2Bpu3fv7/tdQ8dOtRgJ2i5Z7e92vZB2zvGLJtue5Ptd6vbad1tE0CnJnIY/0tJ135l2X2SNkfExZI2V48BDLCWYY+IlyV9/JXF10taU91fI2lRw30BaFi7n9lnRMTxD2MfSppR90Tbw5KG29wOgIZ0fIIuIqL0Q5IRsUrSKunk/sFJ4GTX7tDbAdszJam6PdhcSwC6od2wb5A0VN0fkrS+mXYAdEvLw3jbayVdLelc23sl/UzSQ5J+Y/t2SXsk/bCbTQ6CSZMm1dYefPDB4rp79uwp1ltdM97PcfZOt/3000/X1lrNOz9r1qyOto0vaxn2iLippvS9hnsB0EV8XRZIgrADSRB2IAnCDiRB2IEkmLJ5gkqXW958883FdTds2NB0OyeNM888s7a2devW4rqlabIl6aKLLmqrp1MdUzYDyRF2IAnCDiRB2IEkCDuQBGEHkiDsQBKMs0/Q448/XlubPHlycd2hoaFifZAvce2mF154oVifP39+sX722Wc32c4pg3F2IDnCDiRB2IEkCDuQBGEHkiDsQBKEHUiCKZsnaNeuXbW1O+64o7hu6ZpuSTp8+HBbPZ3sWr0va9asKdZxYtizA0kQdiAJwg4kQdiBJAg7kARhB5Ig7EASjLNP0IoVK2prs2fPLq67bt26Yn3JkiXF+s6dO4v1QVZ6b+bNm1dcd2RkpOl2Umu5Z7e92vZB2zvGLFtue5/t16q/67rbJoBOTeQw/peSrh1n+X9ExJzq73fNtgWgaS3DHhEvS/q4B70A6KJOTtDdbfuN6jB/Wt2TbA/bHrHNBzCgj9oN+0pJ35Y0R9J+SY/UPTEiVkXE3IiY2+a2ADSgrbBHxIGIOBoRxyQ9LunKZtsC0LS2wm575piHP5C0o+65AAZDy3F222slXS3pXNt7Jf1M0tW250gKSbsl/biLPQ684eHhYn3mzJnF+qefftpkOwNlypQptbXTTivva44dO9Z0O6m1DHtE3DTO4ie60AuALuLrskAShB1IgrADSRB2IAnCDiTBJa4NOHr0aLG+d+/eYn369OnF+gMPPFCsr1y5srb2wQcfFNfttqlTp7a97qRJkxrsBOzZgSQIO5AEYQeSIOxAEoQdSIKwA0kQdiAJxtkHwFlnnVWsL1u2rFh/5ZVXamv9HmdfvHhx2+seOHCgwU7Anh1IgrADSRB2IAnCDiRB2IEkCDuQBGEHkmCcfQDs2bOnWH/00UeL9VtuuaW2tnHjxnZa+oLtYj0iivXTTz+97W1v2bKl7XXxdezZgSQIO5AEYQeSIOxAEoQdSIKwA0kQdiAJxtlPAkuXLi3Wt2/fXlu78847i+uWfnNeaj2O3sqRI0faXvfyyy8v1jdv3tz2a2fUcs9u+wLbf7D9tu23bP+kWj7d9ibb71a307rfLoB2TeQw/oikf4mI70iaJ2mJ7e9Iuk/S5oi4WNLm6jGAAdUy7BGxPyJere4fkvSOpPMlXS9pTfW0NZIWdatJAJ07oc/sti+U9F1Jf5Q0IyL2V6UPJc2oWWdY0nD7LQJowoTPxts+U9IzkpZGxF/G1mL0LM64Z3IiYlVEzI2IuR11CqAjEwq77W9oNOi/iohnq8UHbM+s6jMlHexOiwCa0PIw3qPXOD4h6Z2I+MWY0gZJQ5Ieqm7Xd6VDtPTUU0/V1h5++OHiukNDQ8X6ihUrivVNmzYV6wsXLizWSzq9PBdfNpHP7FdJ+mdJb9p+rVr2U42G/De2b5e0R9IPu9MigCa0DHtE/K+kul8w+F6z7QDoFr4uCyRB2IEkCDuQBGEHkiDsQBLu9BLGE9qY3buNJTJ16tTa2ksvvVRc94orrijWP/vss2J99+7dxfoll1xSW3vvvfeK61511VXFOlM6jy8ixh09Y88OJEHYgSQIO5AEYQeSIOxAEoQdSIKwA0kwzn4S6GTa5Hnz5hXXveeee4r1xYsXF+udWLSo/LOF69fzEwntYJwdSI6wA0kQdiAJwg4kQdiBJAg7kARhB5JgnB04xTDODiRH2IEkCDuQBGEHkiDsQBKEHUiCsANJtAy77Qts/8H227bfsv2Tavly2/tsv1b9Xdf9dgG0q+WXamzPlDQzIl61fZakbZIWaXQ+9sMR8e8T3hhfqgG6ru5LNROZn32/pP3V/UO235F0frPtAei2E/rMbvtCSd+V9Mdq0d2237C92va0mnWGbY/YHumoUwAdmfB3422fKeklST+PiGdtz5D0kaSQ9IBGD/Vva/EaHMYDXVZ3GD+hsNv+hqTfSnohIn4xTv1CSb+NiEtbvA5hB7qs7QthPPrTpk9Iemds0KsTd8f9QNKOTpsE0D0TORu/QNL/SHpT0rFq8U8l3SRpjkYP43dL+nF1Mq/0WuzZgS7r6DC+KYQd6D6uZweSI+xAEoQdSIKwA0kQdiAJwg4kQdiBJAg7kARhB5Ig7EAShB1IgrADSRB2IAnCDiTR8gcnG/aRpD1jHp9bLRtEg9rboPYl0Vu7muzt7+sKPb2e/Wsbt0ciYm7fGigY1N4GtS+J3trVq944jAeSIOxAEv0O+6o+b79kUHsb1L4kemtXT3rr62d2AL3T7z07gB4h7EASfQm77Wtt/8n2Ltv39aOHOrZ3236zmoa6r/PTVXPoHbS9Y8yy6bY32X63uh13jr0+9TYQ03gXphnv63vX7+nPe/6Z3fYkSTslfV/SXklbJd0UEW/3tJEatndLmhsRff8Chu1/kHRY0n8en1rL9sOSPo6Ih6p/KKdFxL8OSG/LdYLTeHept7ppxm9RH9+7Jqc/b0c/9uxXStoVEe9HxF8l/VrS9X3oY+BFxMuSPv7K4uslranur9Ho/yw9V9PbQIiI/RHxanX/kKTj04z39b0r9NUT/Qj7+ZL+PObxXg3WfO8h6fe2t9ke7ncz45gxZpqtDyXN6Gcz42g5jXcvfWWa8YF579qZ/rxTnKD7ugURcbmkf5K0pDpcHUgx+hlskMZOV0r6tkbnANwv6ZF+NlNNM/6MpKUR8ZextX6+d+P01ZP3rR9h3yfpgjGPv1ktGwgRsa+6PShpnUY/dgySA8dn0K1uD/a5ny9ExIGIOBoRxyQ9rj6+d9U0489I+lVEPFst7vt7N15fvXrf+hH2rZIutn2R7cmSfiRpQx/6+BrbU6sTJ7I9VdJCDd5U1BskDVX3hySt72MvXzIo03jXTTOuPr93fZ/+PCJ6/ifpOo2ekX9P0r/1o4eavr4l6fXq761+9yZprUYP6z7X6LmN2yX9raTNkt6V9KKk6QPU239pdGrvNzQarJl96m2BRg/R35D0WvV3Xb/fu0JfPXnf+LoskAQn6IAkCDuQBGEHkiDsQBKEHUiCsANJEHYgif8HxG51FCh9LLoAAAAASUVORK5CYII=\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":["pred_labels = new_model.predict(test_images).argmax(axis=1)\n","true_labels = test_labels.argmax(axis=1)"],"metadata":{"id":"IG-VgKNqoYuj","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1670402961186,"user_tz":-330,"elapsed":1811,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"outputId":"6db131ea-2461-4331-f15b-689851bf8369"},"execution_count":21,"outputs":[{"output_type":"stream","name":"stdout","text":["282/282 [==============================] - 1s 2ms/step\n"]}]},{"cell_type":"code","source":["from sklearn.metrics import confusion_matrix\n","confusion_mat = confusion_matrix(true_labels, pred_labels)\n","confusion_mat"],"metadata":{"id":"3-TwwCGyoaxj","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1670403002551,"user_tz":-330,"elapsed":786,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"outputId":"dd300e35-a5a5-4b4e-a9ae-36b417d486de"},"execution_count":22,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[895,   3,   0,   3,   1,   7,   1,   1,   2,   2],\n","       [  6, 885,   0,   0,   0,   0,   0,   2,   2,   0],\n","       [  1,   2, 886,  10,   2,   4,   0,   0,   1,   0],\n","       [  1,   0,  17, 880,   1,   0,   2,   0,   0,   2],\n","       [  1,   5,   4,   3, 873,   2,   0,   0,   0,   2],\n","       [  7,   0,  10,   0,   1, 880,   0,   2,   1,   1],\n","       [  0,   2,   0,   0,   2,   0, 893,   0,   0,   6],\n","       [  0,   1,   0,   0,   0,   3,   1, 885,   0,   0],\n","       [  2,   0,   2,   0,   0,   0,   0,   0, 876,   3],\n","       [  0,   0,   3,   1,   1,   1,   4,   0,   0, 903]])"]},"metadata":{},"execution_count":22}]},{"cell_type":"code","source":["import pandas as pd\n","import seaborn as sns\n","\n","confusion_mat_norm = np.around(confusion_mat.astype('float') / np.atleast_2d(confusion_mat.sum(axis=1)).T, decimals=2)\n","\n","classes = np.arange(0,10).astype('str')\n","confusion_mat_df = pd.DataFrame(confusion_mat_norm,\n","                                index = classes, \n","                                columns = classes)\n","\n","figure = plt.figure()\n","sns.heatmap(confusion_mat_df, annot=True, cmap=plt.cm.Blues)\n","# plt.tight_layout()\n","plt.ylabel('True label')\n","plt.xlabel('Predicted label')\n","plt.show()"],"metadata":{"id":"XKXDeR3nocJB","colab":{"base_uri":"https://localhost:8080/","height":279},"executionInfo":{"status":"ok","timestamp":1670403073256,"user_tz":-330,"elapsed":1654,"user":{"displayName":"ROHAN DUBEY","userId":"11200230061670192260"}},"outputId":"20cd555f-6f59-4f58-de1a-da165ac8dbaa"},"execution_count":24,"outputs":[{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 2 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":[],"metadata":{"id":"5FqM32UOoucd"},"execution_count":null,"outputs":[]}],"metadata":{"colab":{"provenance":[],"authorship_tag":"ABX9TyOUHdNKEzXltRy6XrM212K7"},"gpuClass":"standard","kernelspec":{"display_name":"Python 3","name":"python3"},"language_info":{"name":"python"},"accelerator":"GPU"},"nbformat":4,"nbformat_minor":0}