C2PS.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-06-14T23:48:31.514428Z",
     "start_time": "2023-06-14T23:48:31.483088Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "QxNOtXqrUpWQ",
    "outputId": "052cd3f0-1230-4f72-e59a-e489454b403d"
   },
   "outputs": [],
   "source": [
    "import torch\n",
    "from torch.utils.data import Dataset, DataLoader, SubsetRandomSampler\n",
    "from transformers import Trainer, AutoTokenizer, AutoModel, AutoConfig\n",
    "from sklearn.model_selection import KFold\n",
    "import pandas as pd\n",
    "import tqdm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-06-14T16:06:24.484857Z",
     "start_time": "2023-06-14T16:06:24.472035Z"
    },
    "id": "RxKIjUbIiEyo"
   },
   "outputs": [],
   "source": [
    "device = torch.device(\"cuda\") if torch.cuda.is_available() else torch.device(\"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-06-14T16:18:23.870902Z",
     "start_time": "2023-06-14T16:18:23.823285Z"
    },
    "id": "N9PE2XG_ipWw"
   },
   "outputs": [],
   "source": [
    "class CppDataset(Dataset):\n",
    "    def __init__(self):\n",
    "        # initializing - read text\n",
    "        self.code_df = pd.read_csv(\"Datasets/spoc-train.tsv\", sep='\\t')\n",
    "        self.code_tokenizer = AutoTokenizer.from_pretrained(\"neulab/codebert-cpp\")\n",
    "        self.text_tokenizer = AutoTokenizer.from_pretrained(\"openai-gpt\")\n",
    "        \n",
    "        self.x, self.y = read_code(self.code_df)\n",
    "        self.x = [self.tokenize_code(code) for code in self.x] #code를 tokenize한 결과를 저장\n",
    "        self.y = [self.tokenize_text(text) for text in self.y] #text를 tokenize한 결과를 저장\n",
    "\n",
    "    def __getitem__(self, index):\n",
    "        inputs = {\n",
    "            \"input_ids\" : torch.tensor(self.x[index][\"input_ids\"], dtype = torch.long, device = device),\n",
    "            \"attention_mask\" : torch.tensor(self.x[index][\"attention_mask\"], dtype = torch.long, device = device)\n",
    "        }\n",
    "        targets = torch.tensor(self.y[index], dtype=torch.long)\n",
    "        return inputs, targets\n",
    "    \n",
    "    def __len__(self):\n",
    "        return len(self.x)\n",
    "    \n",
    "    def tokenize_code(self, code):\n",
    "        return self.code_tokenizer(code, truncation=True, padding='max_length')\n",
    "    \n",
    "    def tokenize_text(self, text):\n",
    "        tokenized = self.text_tokenizer(text, truncation=True)[\"input_ids\"]\n",
    "        return tokenized + [0 for i in range(512 - len(tokenized))]\n",
    "    \n",
    "\n",
    "def read_code(code_df: pd.DataFrame):\n",
    "    x = []\n",
    "    y = []\n",
    "    \n",
    "    line = -2\n",
    "    indent = ''\n",
    "    for i in range(len(code_df)):\n",
    "        # print(i)\n",
    "        indent = '\\t' * code_df['indent'][i]\n",
    "        if (code_df['line'][i] == line + 1):\n",
    "            if (not pd.isna(code_df['code'][i])): x[len(x)-1] += indent + code_df['code'][i] + '\\n'\n",
    "            if (not pd.isna(code_df['text'][i])): y[len(y)-1] += indent + code_df['text'][i] + '\\n'\n",
    "        else:\n",
    "            line = -1\n",
    "            if (not pd.isna(code_df['code'][i])):\n",
    "                x.append(indent + code_df['code'][i] + '\\n')\n",
    "            else: x.append(\"\")\n",
    "            if (not pd.isna(code_df['text'][i])):\n",
    "                y.append(indent + code_df['text'][i] + '\\n')\n",
    "            else: y.append(\"\")\n",
    "        line += 1\n",
    "        \n",
    "    return x, y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-06-14T23:00:33.578562Z",
     "start_time": "2023-06-14T23:00:33.553568Z"
    },
    "id": "m5y83DWBkJjD"
   },
   "outputs": [],
   "source": [
    "class C2PSModel(torch.nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        self.encoder = AutoModel.from_pretrained(\"neulab/codebert-cpp\")\n",
    "        self.linear = torch.nn.Linear(768, 768, bias=True)\n",
    "        self.gpt_config = AutoConfig.from_pretrained(\"openai-gpt\")\n",
    "        self.decoder = AutoModel.from_config(self.gpt_config)\n",
    "\n",
    "    def forward(self, inputs):\n",
    "        outputs = self.encoder(**inputs)\n",
    "        outputs = self.linear(outputs)\n",
    "        outputs = self.decoder(inputs_embeds=outputs)\n",
    "        return outputs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-06-14T23:00:44.576638Z",
     "start_time": "2023-06-14T23:00:34.550022Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Some weights of the model checkpoint at neulab/codebert-cpp were not used when initializing RobertaModel: ['lm_head.layer_norm.bias', 'lm_head.bias', 'lm_head.layer_norm.weight', 'lm_head.dense.bias', 'lm_head.dense.weight']\n",
      "- This IS expected if you are initializing RobertaModel from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
      "- This IS NOT expected if you are initializing RobertaModel from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n",
      "Some weights of RobertaModel were not initialized from the model checkpoint at neulab/codebert-cpp and are newly initialized: ['roberta.pooler.dense.bias', 'roberta.pooler.dense.weight']\n",
      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "C2PSModel(\n",
       "  (encoder): RobertaModel(\n",
       "    (embeddings): RobertaEmbeddings(\n",
       "      (word_embeddings): Embedding(50265, 768, padding_idx=1)\n",
       "      (position_embeddings): Embedding(514, 768, padding_idx=1)\n",
       "      (token_type_embeddings): Embedding(1, 768)\n",
       "      (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "      (dropout): Dropout(p=0.1, inplace=False)\n",
       "    )\n",
       "    (encoder): RobertaEncoder(\n",
       "      (layer): ModuleList(\n",
       "        (0): RobertaLayer(\n",
       "          (attention): RobertaAttention(\n",
       "            (self): RobertaSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): RobertaSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): RobertaIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): RobertaOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (1): RobertaLayer(\n",
       "          (attention): RobertaAttention(\n",
       "            (self): RobertaSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): RobertaSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): RobertaIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): RobertaOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (2): RobertaLayer(\n",
       "          (attention): RobertaAttention(\n",
       "            (self): RobertaSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): RobertaSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): RobertaIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): RobertaOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (3): RobertaLayer(\n",
       "          (attention): RobertaAttention(\n",
       "            (self): RobertaSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): RobertaSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): RobertaIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): RobertaOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (4): RobertaLayer(\n",
       "          (attention): RobertaAttention(\n",
       "            (self): RobertaSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): RobertaSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): RobertaIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): RobertaOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (5): RobertaLayer(\n",
       "          (attention): RobertaAttention(\n",
       "            (self): RobertaSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): RobertaSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): RobertaIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): RobertaOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (6): RobertaLayer(\n",
       "          (attention): RobertaAttention(\n",
       "            (self): RobertaSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): RobertaSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): RobertaIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): RobertaOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (7): RobertaLayer(\n",
       "          (attention): RobertaAttention(\n",
       "            (self): RobertaSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): RobertaSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): RobertaIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): RobertaOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (8): RobertaLayer(\n",
       "          (attention): RobertaAttention(\n",
       "            (self): RobertaSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): RobertaSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): RobertaIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): RobertaOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (9): RobertaLayer(\n",
       "          (attention): RobertaAttention(\n",
       "            (self): RobertaSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): RobertaSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): RobertaIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): RobertaOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (10): RobertaLayer(\n",
       "          (attention): RobertaAttention(\n",
       "            (self): RobertaSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): RobertaSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): RobertaIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): RobertaOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "        (11): RobertaLayer(\n",
       "          (attention): RobertaAttention(\n",
       "            (self): RobertaSelfAttention(\n",
       "              (query): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (key): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (value): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "            (output): RobertaSelfOutput(\n",
       "              (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "              (dropout): Dropout(p=0.1, inplace=False)\n",
       "            )\n",
       "          )\n",
       "          (intermediate): RobertaIntermediate(\n",
       "            (dense): Linear(in_features=768, out_features=3072, bias=True)\n",
       "            (intermediate_act_fn): GELUActivation()\n",
       "          )\n",
       "          (output): RobertaOutput(\n",
       "            (dense): Linear(in_features=3072, out_features=768, bias=True)\n",
       "            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "            (dropout): Dropout(p=0.1, inplace=False)\n",
       "          )\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "    (pooler): RobertaPooler(\n",
       "      (dense): Linear(in_features=768, out_features=768, bias=True)\n",
       "      (activation): Tanh()\n",
       "    )\n",
       "  )\n",
       "  (linear): Linear(in_features=768, out_features=768, bias=True)\n",
       "  (decoder): OpenAIGPTModel(\n",
       "    (tokens_embed): Embedding(40478, 768)\n",
       "    (positions_embed): Embedding(512, 768)\n",
       "    (drop): Dropout(p=0.1, inplace=False)\n",
       "    (h): ModuleList(\n",
       "      (0): Block(\n",
       "        (attn): Attention(\n",
       "          (c_attn): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (attn_dropout): Dropout(p=0.1, inplace=False)\n",
       "          (resid_dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): MLP(\n",
       "          (c_fc): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (act): NewGELUActivation()\n",
       "          (dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "      )\n",
       "      (1): Block(\n",
       "        (attn): Attention(\n",
       "          (c_attn): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (attn_dropout): Dropout(p=0.1, inplace=False)\n",
       "          (resid_dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): MLP(\n",
       "          (c_fc): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (act): NewGELUActivation()\n",
       "          (dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "      )\n",
       "      (2): Block(\n",
       "        (attn): Attention(\n",
       "          (c_attn): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (attn_dropout): Dropout(p=0.1, inplace=False)\n",
       "          (resid_dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): MLP(\n",
       "          (c_fc): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (act): NewGELUActivation()\n",
       "          (dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "      )\n",
       "      (3): Block(\n",
       "        (attn): Attention(\n",
       "          (c_attn): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (attn_dropout): Dropout(p=0.1, inplace=False)\n",
       "          (resid_dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): MLP(\n",
       "          (c_fc): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (act): NewGELUActivation()\n",
       "          (dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "      )\n",
       "      (4): Block(\n",
       "        (attn): Attention(\n",
       "          (c_attn): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (attn_dropout): Dropout(p=0.1, inplace=False)\n",
       "          (resid_dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): MLP(\n",
       "          (c_fc): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (act): NewGELUActivation()\n",
       "          (dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "      )\n",
       "      (5): Block(\n",
       "        (attn): Attention(\n",
       "          (c_attn): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (attn_dropout): Dropout(p=0.1, inplace=False)\n",
       "          (resid_dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): MLP(\n",
       "          (c_fc): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (act): NewGELUActivation()\n",
       "          (dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "      )\n",
       "      (6): Block(\n",
       "        (attn): Attention(\n",
       "          (c_attn): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (attn_dropout): Dropout(p=0.1, inplace=False)\n",
       "          (resid_dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): MLP(\n",
       "          (c_fc): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (act): NewGELUActivation()\n",
       "          (dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "      )\n",
       "      (7): Block(\n",
       "        (attn): Attention(\n",
       "          (c_attn): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (attn_dropout): Dropout(p=0.1, inplace=False)\n",
       "          (resid_dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): MLP(\n",
       "          (c_fc): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (act): NewGELUActivation()\n",
       "          (dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "      )\n",
       "      (8): Block(\n",
       "        (attn): Attention(\n",
       "          (c_attn): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (attn_dropout): Dropout(p=0.1, inplace=False)\n",
       "          (resid_dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): MLP(\n",
       "          (c_fc): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (act): NewGELUActivation()\n",
       "          (dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "      )\n",
       "      (9): Block(\n",
       "        (attn): Attention(\n",
       "          (c_attn): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (attn_dropout): Dropout(p=0.1, inplace=False)\n",
       "          (resid_dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): MLP(\n",
       "          (c_fc): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (act): NewGELUActivation()\n",
       "          (dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "      )\n",
       "      (10): Block(\n",
       "        (attn): Attention(\n",
       "          (c_attn): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (attn_dropout): Dropout(p=0.1, inplace=False)\n",
       "          (resid_dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): MLP(\n",
       "          (c_fc): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (act): NewGELUActivation()\n",
       "          (dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "      )\n",
       "      (11): Block(\n",
       "        (attn): Attention(\n",
       "          (c_attn): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (attn_dropout): Dropout(p=0.1, inplace=False)\n",
       "          (resid_dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "        (mlp): MLP(\n",
       "          (c_fc): Conv1D()\n",
       "          (c_proj): Conv1D()\n",
       "          (act): NewGELUActivation()\n",
       "          (dropout): Dropout(p=0.1, inplace=False)\n",
       "        )\n",
       "        (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)\n",
       "      )\n",
       "    )\n",
       "  )\n",
       ")"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "C2PSModel()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train(batch_size, epochs, lr=1e-05):\n",
    "    dataset = CppDataset()\n",
    "    splits = KFold(n_splits=5, shuffle=True)\n",
    "    \n",
    "    for fold, (train_index, valid_index) in enumerate(splits.split(dataset)):\n",
    "        \n",
    "        model = C2PSModel().to(device) # fold 별로 모델을 새로 쓰기 때문\n",
    "        \n",
    "        # kfold로 구분된 index를 이용해 데이터로더 생성\n",
    "        train_sampler = SubsetRandomSampler(train_index)\n",
    "        valid_sampler = SubsetRandomSampler(valid_index)\n",
    "        \n",
    "        dataloader = DataLoader(dataset, batch_size=batch_size, sampler=train_sampler)\n",
    "        val_dataloader = DataLoader(dataset, batch_size=batch_size, sampler=valid_sampler)\n",
    "        \n",
    "        loss_fn = ## loss function 정의\n",
    "        optimizer = ## 최적화 방식 정의\n",
    "        \n",
    "        # train\n",
    "        for epoch in range(epochs):\n",
    "            progress = tqdm.tqdm(dataloader, total=len(dataloader)) # 학습 시 나오는 progress bar\n",
    "            \n",
    "            for i, (inputs, targets) in enumerate(progress):\n",
    "                ## 여기부터는 해보십쇼\n",
    "            \n",
    "            \n",
    "            # validation check: no train\n",
    "            with torch.no_grad():\n",
    "                ## 여기도 해보십쇼\n",
    "                ## Accuracy 출력\n",
    "                \n",
    "            ## epoch 별 loss 를 출력해도 됩니다 안해도 됩니다\n",
    "        \n",
    "            ## early stopping을 적용해도 됩니다\n",
    "            \n",
    "        ## fold 별 loss를 출력\n",
    "        \n",
    "        ## fold 별 모델을 저장\n",
    "        torch.save(model.state_dict(), f'ckpt_fold{fold}.pth')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# main\n",
    "\n",
    "BATCH_SIZE = ##\n",
    "EPOCHS = ##\n",
    "lr = ## learning rate decay를 적용해도 됨 아니면 그냥 상수로\n",
    "\n",
    "train(BATCH_SIZE, EPOCHS, lr=lr)"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "provenance": []
  },
  "kernelspec": {
   "display_name": "aicoding",
   "language": "python",
   "name": "aicoding"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}