ML_project_2labels.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 27,
      "metadata": {
        "id": "q9Ai4PMWLp89"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import torch\n",
        "# from sentence_transformers import SentenceTransformer\n",
        "from sklearn.model_selection import train_test_split\n",
        "import gensim\n",
        "import random\n",
        "from gensim.test.utils import common_texts\n",
        "from gensim.models import Word2Vec\n",
        "from torch.utils.data.sampler import WeightedRandomSampler\n",
        "from sklearn.metrics import confusion_matrix\n",
        "import seaborn as sn\n",
        "from torch.optim.lr_scheduler import ReduceLROnPlateau\n",
        "from sklearn.metrics import precision_score\n",
        "from sklearn.metrics import accuracy_score\n",
        "from sklearn.metrics import recall_score\n",
        "from sklearn.metrics import f1_score\n",
        "from sklearn.metrics import RocCurveDisplay\n",
        "import matplotlib.pyplot as plt"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NJDk1CEaNVOI"
      },
      "source": [
        "# Load the data"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Combine the offensive and hate languages\n",
        "data = pd.read_csv(\"/content/data_processed.csv\")\n",
        "data['class'] = data['class'].replace(0, 1)\n",
        "data['class'] = data['class']-1\n",
        "data"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 424
        },
        "id": "3RQelIId7EUm",
        "outputId": "c2832149-6d84-4511-cfe6-f9e774b7bc83"
      },
      "execution_count": 72,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       Unnamed: 0  count  hate_speech  offensive_language  neither  class  \\\n",
              "0               0      3            0                   0        3      1   \n",
              "1               1      3            0                   3        0      0   \n",
              "2               2      3            0                   3        0      0   \n",
              "3               3      3            0                   2        1      0   \n",
              "4               4      6            0                   6        0      0   \n",
              "...           ...    ...          ...                 ...      ...    ...   \n",
              "24768       25291      3            0                   2        1      0   \n",
              "24769       25292      3            0                   1        2      1   \n",
              "24770       25294      3            0                   3        0      0   \n",
              "24771       25295      6            0                   6        0      0   \n",
              "24772       25296      3            0                   0        3      1   \n",
              "\n",
              "                                                   tweet  \n",
              "0       as a woman you shouldnt complain about cleani...  \n",
              "1       boy dats coldtyga dwn bad for cuffin dat hoe ...  \n",
              "2       dawg you ever fuck a bitch and she start to c...  \n",
              "3                                 she look like a tranny  \n",
              "4       the shit you hear about me might be true or i...  \n",
              "...                                                  ...  \n",
              "24768  yous a muthafin lie right his tl is trash now ...  \n",
              "24769  youve gone and broke the wrong heart baby and ...  \n",
              "24770  young buck wanna eat dat nigguh like i aint fu...  \n",
              "24771              youu got wild bitches tellin you lies  \n",
              "24772  ruffled ntac eileen dahlia beautiful color com...  \n",
              "\n",
              "[24773 rows x 7 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-f82cb5cb-80ef-4a29-8884-804c8abd8d2a\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Unnamed: 0</th>\n",
              "      <th>count</th>\n",
              "      <th>hate_speech</th>\n",
              "      <th>offensive_language</th>\n",
              "      <th>neither</th>\n",
              "      <th>class</th>\n",
              "      <th>tweet</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>as a woman you shouldnt complain about cleani...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>boy dats coldtyga dwn bad for cuffin dat hoe ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>dawg you ever fuck a bitch and she start to c...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>she look like a tranny</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>4</td>\n",
              "      <td>6</td>\n",
              "      <td>0</td>\n",
              "      <td>6</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>the shit you hear about me might be true or i...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24768</th>\n",
              "      <td>25291</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>yous a muthafin lie right his tl is trash now ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24769</th>\n",
              "      <td>25292</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>youve gone and broke the wrong heart baby and ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24770</th>\n",
              "      <td>25294</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>young buck wanna eat dat nigguh like i aint fu...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24771</th>\n",
              "      <td>25295</td>\n",
              "      <td>6</td>\n",
              "      <td>0</td>\n",
              "      <td>6</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>youu got wild bitches tellin you lies</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24772</th>\n",
              "      <td>25296</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>ruffled ntac eileen dahlia beautiful color com...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>24773 rows × 7 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-f82cb5cb-80ef-4a29-8884-804c8abd8d2a')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-f82cb5cb-80ef-4a29-8884-804c8abd8d2a button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-f82cb5cb-80ef-4a29-8884-804c8abd8d2a');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 72
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 73,
      "metadata": {
        "id": "geG6DIsqRe8r"
      },
      "outputs": [],
      "source": [
        "# model = SentenceTransformer('all-MiniLM-L6-v2')\n",
        "sentences = data[\"tweet\"]\n",
        "# sentence_embeddings = model.encode(sentences)\n",
        "words = [list(s.lower().split()) for s in sentences]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 74,
      "metadata": {
        "id": "PwSOGtPz11Og"
      },
      "outputs": [],
      "source": [
        "w2v = Word2Vec(sentences=words, size=384, window=5, min_count=1, workers=4)\n",
        "\n",
        "# import gensim.downloader as api\n",
        "# corpus = api.load('text8')  # download the corpus and return it opened as an iterable\n",
        "# w2v = Word2Vec(corpus)\n",
        "# w2v.train(sentences=words)\n",
        "\n",
        "sentence_embeddings = [torch.as_tensor([w2v.wv[word] for word in list(s.lower().split())]) for s in sentences]"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "label = np.array(data['class'])"
      ],
      "metadata": {
        "id": "iCDqYwrw6-LD"
      },
      "execution_count": 75,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mvbeeSpdUtwt"
      },
      "source": [
        "# Train-validation-test split"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 76,
      "metadata": {
        "id": "3wZ1REB0IerP"
      },
      "outputs": [],
      "source": [
        "def create_train_and_test_set_balanced(X, y, train_ratio=0.8):\n",
        "    # TODO\n",
        "    # process the list of (x,y) pairs and split them 80-20 into train and test set\n",
        "    # train_x is a list of name embeddings each of size (num_characters_in_name, 1, n_letters), train_y is the corresponding list of language category index. Same for test_x and test_y\n",
        "\n",
        "    X, X_rem, y, y_rem = X[0:int(train_ratio * len(X))], X[int(train_ratio * len(X)) + 1:-1], y[0:int(train_ratio * len(y))], y[int(train_ratio * len(y)) + 1:-1]\n",
        "    \n",
        "    counts = np.bincount(y)\n",
        "    label_weights = 1.0 / counts\n",
        "    weights = label_weights[y]\n",
        "    index = list(WeightedRandomSampler(weights, len(weights)))\n",
        "    X_train, y_train = [X[i] for i in index], [y[i] for i in index]\n",
        "\n",
        "    return X_train, X_rem, y_train, y_rem"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 77,
      "metadata": {
        "id": "fhyriPodI4HT"
      },
      "outputs": [],
      "source": [
        "X_train, X_rem, y_train, y_rem = create_train_and_test_set_balanced(sentence_embeddings,label)\n",
        "X_valid, X_test, y_valid, y_test = create_train_and_test_set_balanced(X_rem,y_rem)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 78,
      "metadata": {
        "id": "QSbvmwrcf8wF"
      },
      "outputs": [],
      "source": [
        "from torch.utils.data import Dataset, DataLoader\n",
        "from torch.nn.utils.rnn import pad_sequence\n",
        "\n",
        "class Tweet(Dataset):\n",
        "    def __init__(self, X,y, transform=None):\n",
        "        self.X = pad_sequence(X,batch_first=True)\n",
        "        self.y = torch.as_tensor(np.array(y))\n",
        "        self.transform = transform\n",
        "\n",
        "    def __len__(self):\n",
        "        return len(self.y)\n",
        "\n",
        "    def __getitem__(self, idx):\n",
        "        sentence = self.X[idx]\n",
        "        label = self.y[idx]\n",
        "        sample = {'sentence': sentence, 'label': label}\n",
        "\n",
        "        if self.transform:\n",
        "            sample = self.transform(sample)\n",
        "\n",
        "        return sample"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "v8jbV7gZWIGB"
      },
      "source": [
        "# Define the model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 79,
      "metadata": {
        "id": "dOpX0gRfEvoz"
      },
      "outputs": [],
      "source": [
        "hidden_size = 128\n",
        "output_size = 3\n",
        "batch_size = 50\n",
        "\n",
        "valid_losses = []"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 80,
      "metadata": {
        "id": "iS5akB3pE_sO"
      },
      "outputs": [],
      "source": [
        "train_data = Tweet(X_train, y_train)\n",
        "train_loader = DataLoader(train_data, batch_size=batch_size, shuffle=True, num_workers=0,drop_last=True)\n",
        "\n",
        "valid_data = Tweet(X_valid, y_valid)\n",
        "valid_loader = DataLoader(valid_data, batch_size=batch_size, shuffle=True, num_workers=0,drop_last=True)\n",
        "\n",
        "test_data = Tweet(X_test, y_test)\n",
        "test_loader = DataLoader(test_data, batch_size=len(X_test), shuffle=True, num_workers=0,drop_last=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 86,
      "metadata": {
        "id": "Bp-5MBCQVPw5"
      },
      "outputs": [],
      "source": [
        "from sklearn.preprocessing import normalize\n",
        "import torch.nn as nn\n",
        "import torch\n",
        "from torch.autograd import Variable\n",
        "\n",
        "\n",
        "class lstm(nn.Module):\n",
        "    def __init__(self, embedding_length=384,dropout=0.3):\n",
        "        super(lstm, self).__init__()\n",
        "\n",
        "        self.dropout = dropout\n",
        "        self.hidden_size = hidden_size\n",
        "        self.output_size = output_size\n",
        "        self.batch_size = batch_size\n",
        "        self.fc1 = nn.Linear(embedding_length, 256)\n",
        "        self.lstm = nn.LSTM(embedding_length, self.hidden_size,batch_first=True,dropout=dropout, num_layers=1)\n",
        "        self.fc2 = nn.Linear(self.hidden_size, self.hidden_size//2)\n",
        "        self.fc3 = nn.Linear(self.hidden_size//2, self.hidden_size//4)\n",
        "        self.label = nn.Linear(self.hidden_size//4, self.output_size)\n",
        "        self.relu = nn.ReLU()\n",
        "        self.to(\"cuda\")\n",
        "\n",
        "\n",
        "    def forward(self,input):\n",
        "\n",
        "        input = input.to(\"cuda\")\n",
        "\n",
        "        # input = self.fc1(input)\n",
        "        # input = self.relu(input)\n",
        "\n",
        "        h_0 = Variable(torch.zeros(1, len(input), self.hidden_size).cuda())\n",
        "        c_0 = Variable(torch.zeros(1, len(input), self.hidden_size).cuda())\n",
        "\n",
        "\n",
        "        output, (final_hidden_state, final_cell_state) = self.lstm(input, (h_0, c_0))\n",
        "        x = self.fc2(final_hidden_state[-1])\n",
        "        x = self.fc3(x)\n",
        "\n",
        "        return self.label(x)\n",
        "    \n",
        "\n",
        "    def fit(self, train_loader, epochs=300,lr=1e-5,interval=100):\n",
        "        criterion = torch.nn.CrossEntropyLoss()\n",
        "        optimizer = torch.optim.Adam(self.parameters(), lr=lr)\n",
        "        scheduler = ReduceLROnPlateau(optimizer, patience=50, threshold=1e-3)\n",
        "\n",
        "\n",
        "        for epoch in range(epochs):\n",
        "            running_loss = 0.0\n",
        "            self.train()\n",
        "\n",
        "            for i_batch, sample in enumerate(train_loader):\n",
        "                sentences, labels = sample['sentence'].to(\"cuda\"), sample['label'].to(\"cuda\")\n",
        "\n",
        "                outputs = self(sentences)\n",
        "                loss = criterion(\n",
        "                    outputs,\n",
        "                    labels,\n",
        "                )\n",
        "\n",
        "                optimizer.zero_grad()\n",
        "                loss.backward()\n",
        "                optimizer.step()\n",
        "                running_loss += loss.item()\n",
        "\n",
        "                if i_batch % interval == 0 and i_batch > 0:\n",
        "                    valid_loss = self.validate(valid_loader=valid_loader)\n",
        "                    print(f\"epoch:{epoch},batch: {i_batch}, train_loss: {running_loss/interval}, valid_loss: {valid_loss}\")\n",
        "                    running_loss = 0\n",
        "            scheduler.step(self.validate(valid_loader))\n",
        "            valid_losses.append([epoch,self.validate(valid_loader)])\n",
        "\n",
        "    def validate(self, valid_loader):\n",
        "        running_loss = 0\n",
        "        val_count = 0\n",
        "        # self.eval()\n",
        "        criterion = torch.nn.CrossEntropyLoss()\n",
        "\n",
        "        for i_batch, sample in enumerate(valid_loader):\n",
        "            sentences, labels = sample['sentence'].to(\"cuda\"), sample['label'].to(\"cuda\")\n",
        "\n",
        "            outputs = self(sentences)\n",
        "            loss = criterion(\n",
        "                outputs,\n",
        "                labels,\n",
        "            )\n",
        "\n",
        "            running_loss += loss.item()\n",
        "            val_count += 1\n",
        "\n",
        "        avg_loss = running_loss / val_count\n",
        "        return avg_loss\n",
        "\n",
        "    def accuracy(self,test_loader):\n",
        "        self.eval()\n",
        "        plt.figure(figsize=(10, 10))\n",
        "        for i_batch, sample in enumerate(test_loader):\n",
        "            sentences, labels = sample['sentence'], sample['label']\n",
        "    \n",
        "            outputs = self(sentences)\n",
        "            pred = np.array([np.argmax(outputs[i].cpu().detach().numpy()) for i in range(len(outputs))])\n",
        "            assert len(pred) == len(labels)\n",
        "            cm = pd.DataFrame(confusion_matrix(labels,pred,normalize=\"true\"),index=[\"True offensive\",\"True non-offensive\"],columns=[\"Predicted offensive\",\"Predicted non-offensive\"])\n",
        "            \n",
        "            sn.heatmap(data=cm,annot=True,xticklabels=True,yticklabels=True)\n",
        "            plt.show()\n",
        "\n",
        "            return np.sum(labels.numpy() == pred) / len(pred)\n",
        "\n",
        "\n",
        "    def predict(self,sentence):\n",
        "        vec = torch.as_tensor([w2v.wv[word] for word in list(sentence.lower().split())])\n",
        "        vec = vec.unsqueeze(0)\n",
        "        index =  np.argmax(self(vec).cpu().detach().numpy()) \n",
        "        index = [np.argmax(r) for r in index]\n",
        "\n",
        "        # label_to_senti = {0:\"hate speech\",1:\"offensive language\",2:\"neither\"}\n",
        "        label_to_senti = {0:\"offensive\",1:\"normal\"}\n",
        "        return label_to_senti[index]\n",
        "\n",
        "    def metrics(self,test_loader):\n",
        "        self.eval()\n",
        "        plt.figure(figsize=(10, 10))\n",
        "        for i_batch, sample in enumerate(test_loader):\n",
        "            sentences, labels = sample['sentence'], sample['label']\n",
        "    \n",
        "            outputs = self(sentences)\n",
        "            pred = np.array([np.argmax(outputs[i].cpu().detach().numpy()) for i in range(len(outputs))])\n",
        "            assert len(pred) == len(labels)\n",
        "            accuracy = accuracy_score(labels, pred)\n",
        "            precision = precision_score(labels, pred, pos_label=0)\n",
        "            recall = recall_score(labels, pred, pos_label=0)\n",
        "            f1 = f1_score(labels, pred, pos_label=0)\n",
        "            RocCurveDisplay.from_predictions(labels, pred, name=\"LSTM\")\n",
        "            plt.xlabel(\"False Positive Rate\")\n",
        "            plt.ylabel(\"True Positive Rate\")\n",
        "            plt.show()\n",
        "            print(f\"Accuracy={accuracy}\\nPrecision={precision}\\nRecall={recall}\\nF1={f1}\")\n",
        "\n",
        "            "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 87,
      "metadata": {
        "id": "TlNCSyJEatXa",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5cb13c7c-19bd-4417-af34-49a42978a665"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.8/dist-packages/torch/nn/modules/rnn.py:67: UserWarning: dropout option adds dropout after all but last recurrent layer, so non-zero dropout expects num_layers greater than 1, but got dropout=0.3 and num_layers=1\n",
            "  warnings.warn(\"dropout option adds dropout after all but last \"\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "epoch:0,batch: 100, train_loss: 1.1629631638526916, valid_loss: 1.1439891947975642\n",
            "epoch:0,batch: 200, train_loss: 1.1357408201694488, valid_loss: 1.128324409074421\n",
            "epoch:0,batch: 300, train_loss: 1.1189956724643708, valid_loss: 1.109301763244822\n",
            "epoch:1,batch: 100, train_loss: 1.0614467579126359, valid_loss: 0.9893659790859947\n",
            "epoch:1,batch: 200, train_loss: 0.8566471272706986, valid_loss: 0.7583079413522648\n",
            "epoch:1,batch: 300, train_loss: 0.7356837797164917, valid_loss: 0.719954369188864\n",
            "epoch:2,batch: 100, train_loss: 0.7141463506221771, valid_loss: 0.7045826904381378\n",
            "epoch:2,batch: 200, train_loss: 0.7031565886735917, valid_loss: 0.7016292418105693\n",
            "epoch:2,batch: 300, train_loss: 0.7008812820911408, valid_loss: 0.6998272545730011\n",
            "epoch:3,batch: 100, train_loss: 0.7054405528306961, valid_loss: 0.6979590692097628\n",
            "epoch:3,batch: 200, train_loss: 0.6975140279531479, valid_loss: 0.6972659615021718\n",
            "epoch:3,batch: 300, train_loss: 0.6971563363075256, valid_loss: 0.6965355593946916\n",
            "epoch:4,batch: 100, train_loss: 0.7029860574007034, valid_loss: 0.6957933533040783\n",
            "epoch:4,batch: 200, train_loss: 0.6958750015497208, valid_loss: 0.6955929941768888\n",
            "epoch:4,batch: 300, train_loss: 0.6955691808462143, valid_loss: 0.6952746457691434\n",
            "epoch:5,batch: 100, train_loss: 0.7021465325355529, valid_loss: 0.6949637019181554\n",
            "epoch:5,batch: 200, train_loss: 0.6947243297100067, valid_loss: 0.6956093326399598\n",
            "epoch:5,batch: 300, train_loss: 0.6949764621257782, valid_loss: 0.694601504863063\n",
            "epoch:6,batch: 100, train_loss: 0.7015976911783218, valid_loss: 0.6945613041708741\n",
            "epoch:6,batch: 200, train_loss: 0.6946255087852478, valid_loss: 0.6943025807791119\n",
            "epoch:6,batch: 300, train_loss: 0.6937049555778504, valid_loss: 0.6951393039920662\n",
            "epoch:7,batch: 100, train_loss: 0.7008952057361603, valid_loss: 0.6942340927787974\n",
            "epoch:7,batch: 200, train_loss: 0.6940063738822937, valid_loss: 0.6937124932868571\n",
            "epoch:7,batch: 300, train_loss: 0.6937610471248626, valid_loss: 0.6934695726708521\n",
            "epoch:8,batch: 100, train_loss: 0.6999631309509278, valid_loss: 0.6927463140668748\n",
            "epoch:8,batch: 200, train_loss: 0.6924741250276566, valid_loss: 0.6918168482901175\n",
            "epoch:8,batch: 300, train_loss: 0.6915591382980346, valid_loss: 0.69047113008137\n",
            "epoch:9,batch: 100, train_loss: 0.6928117614984513, valid_loss: 0.6813801753370068\n",
            "epoch:9,batch: 200, train_loss: 0.6784376871585845, valid_loss: 0.6715402459796471\n",
            "epoch:9,batch: 300, train_loss: 0.6669864648580551, valid_loss: 0.6564086637919462\n",
            "epoch:10,batch: 100, train_loss: 0.6392462116479873, valid_loss: 0.6141121998617921\n",
            "epoch:10,batch: 200, train_loss: 0.6051464027166367, valid_loss: 0.5823655958417095\n",
            "epoch:10,batch: 300, train_loss: 0.5786887258291245, valid_loss: 0.5525244172615341\n",
            "epoch:11,batch: 100, train_loss: 0.5344225841760636, valid_loss: 0.5208470297010639\n",
            "epoch:11,batch: 200, train_loss: 0.5088351163268089, valid_loss: 0.4865982638129705\n",
            "epoch:11,batch: 300, train_loss: 0.4939373001456261, valid_loss: 0.47138027073461797\n",
            "epoch:12,batch: 100, train_loss: 0.46987518161535263, valid_loss: 0.45222534069532083\n",
            "epoch:12,batch: 200, train_loss: 0.45020838916301725, valid_loss: 0.45286731553983084\n",
            "epoch:12,batch: 300, train_loss: 0.42900161772966383, valid_loss: 0.4188574778882763\n",
            "epoch:13,batch: 100, train_loss: 0.4041599121689796, valid_loss: 0.3956975427609456\n",
            "epoch:13,batch: 200, train_loss: 0.3948845863342285, valid_loss: 0.3894854237007189\n",
            "epoch:13,batch: 300, train_loss: 0.3937117049098015, valid_loss: 0.3827689190076876\n",
            "epoch:14,batch: 100, train_loss: 0.3909530970454216, valid_loss: 0.3749868705680099\n",
            "epoch:14,batch: 200, train_loss: 0.37208389803767206, valid_loss: 0.36799423887005334\n",
            "epoch:14,batch: 300, train_loss: 0.3761500011384487, valid_loss: 0.3651446473749378\n",
            "epoch:15,batch: 100, train_loss: 0.3679055382311344, valid_loss: 0.36695830731452267\n",
            "epoch:15,batch: 200, train_loss: 0.36166629046201704, valid_loss: 0.35521455090257187\n",
            "epoch:15,batch: 300, train_loss: 0.35715137124061586, valid_loss: 0.35246425885942917\n",
            "epoch:16,batch: 100, train_loss: 0.35817123025655745, valid_loss: 0.35105787632586083\n",
            "epoch:16,batch: 200, train_loss: 0.3702293102443218, valid_loss: 0.344157926643951\n",
            "epoch:16,batch: 300, train_loss: 0.3375930543243885, valid_loss: 0.3407428404198417\n",
            "epoch:17,batch: 100, train_loss: 0.352990537583828, valid_loss: 0.3376523867438111\n",
            "epoch:17,batch: 200, train_loss: 0.3352806855738163, valid_loss: 0.33708425677275355\n",
            "epoch:17,batch: 300, train_loss: 0.3384807248413563, valid_loss: 0.34802775217007986\n",
            "epoch:18,batch: 100, train_loss: 0.3477353735268116, valid_loss: 0.3375811127922203\n",
            "epoch:18,batch: 200, train_loss: 0.3468984453380108, valid_loss: 0.3304840245201618\n",
            "epoch:18,batch: 300, train_loss: 0.3354160064458847, valid_loss: 0.33054142190685754\n",
            "epoch:19,batch: 100, train_loss: 0.3299683019518852, valid_loss: 0.32567296620411207\n",
            "epoch:19,batch: 200, train_loss: 0.3334214806556702, valid_loss: 0.32384622568571114\n",
            "epoch:19,batch: 300, train_loss: 0.330822918266058, valid_loss: 0.3259529163943061\n",
            "epoch:20,batch: 100, train_loss: 0.32190380930900575, valid_loss: 0.3235895712918873\n",
            "epoch:20,batch: 200, train_loss: 0.3288687624037266, valid_loss: 0.3193660112498682\n",
            "epoch:20,batch: 300, train_loss: 0.3410187903046608, valid_loss: 0.3196440220256395\n",
            "epoch:21,batch: 100, train_loss: 0.32476543411612513, valid_loss: 0.3171926538778257\n",
            "epoch:21,batch: 200, train_loss: 0.32161320075392724, valid_loss: 0.31499529714825786\n",
            "epoch:21,batch: 300, train_loss: 0.3257055976986885, valid_loss: 0.3113019809692721\n",
            "epoch:22,batch: 100, train_loss: 0.3284183854609728, valid_loss: 0.31180318246913863\n",
            "epoch:22,batch: 200, train_loss: 0.309295109808445, valid_loss: 0.3222629848537566\n",
            "epoch:22,batch: 300, train_loss: 0.3111328688263893, valid_loss: 0.31782901136181024\n",
            "epoch:23,batch: 100, train_loss: 0.31223260536789893, valid_loss: 0.31458135798007625\n",
            "epoch:23,batch: 200, train_loss: 0.31191511943936345, valid_loss: 0.30482874075068705\n",
            "epoch:23,batch: 300, train_loss: 0.3307320563495159, valid_loss: 0.306469534205485\n",
            "epoch:24,batch: 100, train_loss: 0.3164670440554619, valid_loss: 0.3009683344560333\n",
            "epoch:24,batch: 200, train_loss: 0.3065220318734646, valid_loss: 0.30428536057094985\n",
            "epoch:24,batch: 300, train_loss: 0.31919395461678507, valid_loss: 0.3026643544435501\n",
            "epoch:25,batch: 100, train_loss: 0.31920703902840614, valid_loss: 0.30484087048452113\n",
            "epoch:25,batch: 200, train_loss: 0.30853569000959397, valid_loss: 0.301021427293367\n",
            "epoch:25,batch: 300, train_loss: 0.2975558286905289, valid_loss: 0.2980358649658251\n",
            "epoch:26,batch: 100, train_loss: 0.29735376991331575, valid_loss: 0.30336869479734685\n",
            "epoch:26,batch: 200, train_loss: 0.31512358352541925, valid_loss: 0.29505531206915653\n",
            "epoch:26,batch: 300, train_loss: 0.308104660063982, valid_loss: 0.30026827960074703\n",
            "epoch:27,batch: 100, train_loss: 0.30984792605042455, valid_loss: 0.2971499566979046\n",
            "epoch:27,batch: 200, train_loss: 0.3112837617099285, valid_loss: 0.3005208520195152\n",
            "epoch:27,batch: 300, train_loss: 0.3056896218657494, valid_loss: 0.2930695892700666\n",
            "epoch:28,batch: 100, train_loss: 0.3078785155713558, valid_loss: 0.29564948417717896\n",
            "epoch:28,batch: 200, train_loss: 0.2911133033782244, valid_loss: 0.2916735897335825\n",
            "epoch:28,batch: 300, train_loss: 0.29811006650328636, valid_loss: 0.29916035185886336\n",
            "epoch:29,batch: 100, train_loss: 0.30964963108301163, valid_loss: 0.3029794372335265\n",
            "epoch:29,batch: 200, train_loss: 0.2996507316827774, valid_loss: 0.2924656960405881\n",
            "epoch:29,batch: 300, train_loss: 0.298705727905035, valid_loss: 0.28827894224396233\n",
            "epoch:30,batch: 100, train_loss: 0.2905564449727535, valid_loss: 0.28675943301825585\n",
            "epoch:30,batch: 200, train_loss: 0.30723360925912857, valid_loss: 0.28637995486018025\n",
            "epoch:30,batch: 300, train_loss: 0.2897202952206135, valid_loss: 0.28658486600918104\n",
            "epoch:31,batch: 100, train_loss: 0.2798691216111183, valid_loss: 0.2898141803997981\n",
            "epoch:31,batch: 200, train_loss: 0.2931248039007187, valid_loss: 0.28513020050676563\n",
            "epoch:31,batch: 300, train_loss: 0.3139359620213509, valid_loss: 0.28665431560594823\n",
            "epoch:32,batch: 100, train_loss: 0.2986028198897839, valid_loss: 0.28948229787093177\n",
            "epoch:32,batch: 200, train_loss: 0.30684916287660596, valid_loss: 0.3013349352758142\n",
            "epoch:32,batch: 300, train_loss: 0.2959419453144074, valid_loss: 0.2849822627215446\n",
            "epoch:33,batch: 100, train_loss: 0.29118397280573843, valid_loss: 0.28349046446854553\n",
            "epoch:33,batch: 200, train_loss: 0.31095268204808235, valid_loss: 0.28690550014188015\n",
            "epoch:33,batch: 300, train_loss: 0.2944688141345978, valid_loss: 0.29885873236233673\n",
            "epoch:34,batch: 100, train_loss: 0.2890495929121971, valid_loss: 0.2811545245820963\n",
            "epoch:34,batch: 200, train_loss: 0.301045638024807, valid_loss: 0.28162447476311575\n",
            "epoch:34,batch: 300, train_loss: 0.2970046772062778, valid_loss: 0.28228029193757453\n",
            "epoch:35,batch: 100, train_loss: 0.3024431349337101, valid_loss: 0.2808524177798742\n",
            "epoch:35,batch: 200, train_loss: 0.28102919705212115, valid_loss: 0.28173581501351125\n",
            "epoch:35,batch: 300, train_loss: 0.2860124581307173, valid_loss: 0.28529374603229235\n",
            "epoch:36,batch: 100, train_loss: 0.28813683569431303, valid_loss: 0.28090246389561063\n",
            "epoch:36,batch: 200, train_loss: 0.28691537424921987, valid_loss: 0.2840643287459506\n",
            "epoch:36,batch: 300, train_loss: 0.2923371957242489, valid_loss: 0.2807789443980289\n",
            "epoch:37,batch: 100, train_loss: 0.28335890635848043, valid_loss: 0.2982518150459362\n",
            "epoch:37,batch: 200, train_loss: 0.2877160285413265, valid_loss: 0.2812268963154358\n",
            "epoch:37,batch: 300, train_loss: 0.29130949452519417, valid_loss: 0.3015351481264151\n",
            "epoch:38,batch: 100, train_loss: 0.28911452159285544, valid_loss: 0.2801753308954118\n",
            "epoch:38,batch: 200, train_loss: 0.28524477407336235, valid_loss: 0.27855110960670665\n",
            "epoch:38,batch: 300, train_loss: 0.2903010602295399, valid_loss: 0.285974540099313\n",
            "epoch:39,batch: 100, train_loss: 0.2918662967532873, valid_loss: 0.2790073753346371\n",
            "epoch:39,batch: 200, train_loss: 0.29399509146809577, valid_loss: 0.2846218749314924\n",
            "epoch:39,batch: 300, train_loss: 0.2854367892444134, valid_loss: 0.2825125276665144\n",
            "epoch:40,batch: 100, train_loss: 0.2817286778986454, valid_loss: 0.2823602871238431\n",
            "epoch:40,batch: 200, train_loss: 0.28776159361004827, valid_loss: 0.27820598484971854\n",
            "epoch:40,batch: 300, train_loss: 0.2857617295533419, valid_loss: 0.27677498880443696\n",
            "epoch:41,batch: 100, train_loss: 0.2744467607140541, valid_loss: 0.2797544596693184\n",
            "epoch:41,batch: 200, train_loss: 0.2923953345417976, valid_loss: 0.2770714130771311\n",
            "epoch:41,batch: 300, train_loss: 0.2888533420860767, valid_loss: 0.27637253333873385\n",
            "epoch:42,batch: 100, train_loss: 0.28466403439640997, valid_loss: 0.2937374446965471\n",
            "epoch:42,batch: 200, train_loss: 0.2790864238142967, valid_loss: 0.2740467180368267\n",
            "epoch:42,batch: 300, train_loss: 0.2870718169957399, valid_loss: 0.28358210783593263\n",
            "epoch:43,batch: 100, train_loss: 0.28924557074904444, valid_loss: 0.2779774965741966\n",
            "epoch:43,batch: 200, train_loss: 0.2840913557261229, valid_loss: 0.27669566643388965\n",
            "epoch:43,batch: 300, train_loss: 0.27286974966526034, valid_loss: 0.27530298664977276\n",
            "epoch:44,batch: 100, train_loss: 0.2923350078612566, valid_loss: 0.27908287961271744\n",
            "epoch:44,batch: 200, train_loss: 0.27440531730651857, valid_loss: 0.29546444570716424\n",
            "epoch:44,batch: 300, train_loss: 0.27660811454057693, valid_loss: 0.27575555688972714\n",
            "epoch:45,batch: 100, train_loss: 0.27910714656114577, valid_loss: 0.2756131178216089\n",
            "epoch:45,batch: 200, train_loss: 0.28001046262681484, valid_loss: 0.28380771943285493\n",
            "epoch:45,batch: 300, train_loss: 0.28500522978603837, valid_loss: 0.27815281279102155\n",
            "epoch:46,batch: 100, train_loss: 0.28889669716358185, valid_loss: 0.2727999450473846\n",
            "epoch:46,batch: 200, train_loss: 0.27838643617928027, valid_loss: 0.2741697293859494\n",
            "epoch:46,batch: 300, train_loss: 0.2698549653589726, valid_loss: 0.27258573406482045\n",
            "epoch:47,batch: 100, train_loss: 0.274682808816433, valid_loss: 0.2752004888238786\n",
            "epoch:47,batch: 200, train_loss: 0.28430312633514404, valid_loss: 0.2759183239710482\n",
            "epoch:47,batch: 300, train_loss: 0.2809872856736183, valid_loss: 0.27312423534031155\n",
            "epoch:48,batch: 100, train_loss: 0.2834623674303293, valid_loss: 0.27262321152264557\n",
            "epoch:48,batch: 200, train_loss: 0.27606408089399337, valid_loss: 0.27448115120582944\n",
            "epoch:48,batch: 300, train_loss: 0.28013441555202007, valid_loss: 0.2751024216413498\n",
            "epoch:49,batch: 100, train_loss: 0.2808133224397898, valid_loss: 0.2744373073683509\n",
            "epoch:49,batch: 200, train_loss: 0.26681930273771287, valid_loss: 0.27248765377304224\n",
            "epoch:49,batch: 300, train_loss: 0.292400186881423, valid_loss: 0.28708794392362424\n",
            "epoch:50,batch: 100, train_loss: 0.2851292488723993, valid_loss: 0.27272726029534883\n",
            "epoch:50,batch: 200, train_loss: 0.2740107947587967, valid_loss: 0.2713059358581712\n",
            "epoch:50,batch: 300, train_loss: 0.27007903039455416, valid_loss: 0.2715610113890865\n",
            "epoch:51,batch: 100, train_loss: 0.2860951314121485, valid_loss: 0.2730588445180579\n",
            "epoch:51,batch: 200, train_loss: 0.2771450965851545, valid_loss: 0.2826135699884801\n",
            "epoch:51,batch: 300, train_loss: 0.26796479694545267, valid_loss: 0.27151160790950435\n",
            "epoch:52,batch: 100, train_loss: 0.2810512505471706, valid_loss: 0.2698231898719751\n",
            "epoch:52,batch: 200, train_loss: 0.27610458590090275, valid_loss: 0.2701071455508848\n",
            "epoch:52,batch: 300, train_loss: 0.2801502542197704, valid_loss: 0.2729470982581754\n",
            "epoch:53,batch: 100, train_loss: 0.28492039658129215, valid_loss: 0.2710906131358086\n",
            "epoch:53,batch: 200, train_loss: 0.2850911305844784, valid_loss: 0.29719235116167914\n",
            "epoch:53,batch: 300, train_loss: 0.2657547354698181, valid_loss: 0.27209427905610845\n",
            "epoch:54,batch: 100, train_loss: 0.27463341914117334, valid_loss: 0.27426798792579504\n",
            "epoch:54,batch: 200, train_loss: 0.2734297677129507, valid_loss: 0.2707251249989377\n",
            "epoch:54,batch: 300, train_loss: 0.28220049805939196, valid_loss: 0.27465585458882247\n",
            "epoch:55,batch: 100, train_loss: 0.26266030825674536, valid_loss: 0.27077373107777364\n",
            "epoch:55,batch: 200, train_loss: 0.2837886933982372, valid_loss: 0.2734157234618935\n",
            "epoch:55,batch: 300, train_loss: 0.2642928169667721, valid_loss: 0.2708053418164012\n",
            "epoch:56,batch: 100, train_loss: 0.2709876166284084, valid_loss: 0.269290071122254\n",
            "epoch:56,batch: 200, train_loss: 0.2732850801944733, valid_loss: 0.27400376251603986\n",
            "epoch:56,batch: 300, train_loss: 0.27449437096714974, valid_loss: 0.2692118113553977\n",
            "epoch:57,batch: 100, train_loss: 0.2729500597715378, valid_loss: 0.2765427043166342\n",
            "epoch:57,batch: 200, train_loss: 0.2792915889620781, valid_loss: 0.27357074731512915\n",
            "epoch:57,batch: 300, train_loss: 0.26753410436213015, valid_loss: 0.27011229531674447\n",
            "epoch:58,batch: 100, train_loss: 0.2698711732029915, valid_loss: 0.2707255474185642\n",
            "epoch:58,batch: 200, train_loss: 0.26764908283948896, valid_loss: 0.2908164864476723\n",
            "epoch:58,batch: 300, train_loss: 0.28147909075021743, valid_loss: 0.2700765010299562\n",
            "epoch:59,batch: 100, train_loss: 0.2688839363306761, valid_loss: 0.27634392810773245\n",
            "epoch:59,batch: 200, train_loss: 0.26422661982476714, valid_loss: 0.2766964697762381\n",
            "epoch:59,batch: 300, train_loss: 0.273679181560874, valid_loss: 0.26967208973969087\n",
            "epoch:60,batch: 100, train_loss: 0.27180395446717737, valid_loss: 0.268434091647969\n",
            "epoch:60,batch: 200, train_loss: 0.2751497481763363, valid_loss: 0.2741453995621657\n",
            "epoch:60,batch: 300, train_loss: 0.27170357830822467, valid_loss: 0.2685329373690146\n",
            "epoch:61,batch: 100, train_loss: 0.26198854982852937, valid_loss: 0.26994041692984255\n",
            "epoch:61,batch: 200, train_loss: 0.28221297547221186, valid_loss: 0.26894101349613336\n",
            "epoch:61,batch: 300, train_loss: 0.2768214736878872, valid_loss: 0.2701715745880634\n",
            "epoch:62,batch: 100, train_loss: 0.2687609815597534, valid_loss: 0.27106941717712185\n",
            "epoch:62,batch: 200, train_loss: 0.2601571075618267, valid_loss: 0.2671209593551068\n",
            "epoch:62,batch: 300, train_loss: 0.2780706059932709, valid_loss: 0.2986103678240052\n",
            "epoch:63,batch: 100, train_loss: 0.2731857173144817, valid_loss: 0.2671121515427964\n",
            "epoch:63,batch: 200, train_loss: 0.2657611231505871, valid_loss: 0.28249306429790544\n",
            "epoch:63,batch: 300, train_loss: 0.26401575937867167, valid_loss: 0.26778381195249434\n",
            "epoch:64,batch: 100, train_loss: 0.26310363568365575, valid_loss: 0.2771099799230129\n",
            "epoch:64,batch: 200, train_loss: 0.27024000488221644, valid_loss: 0.2694893661367742\n",
            "epoch:64,batch: 300, train_loss: 0.2724116749316454, valid_loss: 0.272335077954244\n",
            "epoch:65,batch: 100, train_loss: 0.2651800486445427, valid_loss: 0.27163992529805703\n",
            "epoch:65,batch: 200, train_loss: 0.26960107520222665, valid_loss: 0.26751569177530987\n",
            "epoch:65,batch: 300, train_loss: 0.26428541615605355, valid_loss: 0.3057379545290259\n",
            "epoch:66,batch: 100, train_loss: 0.2656037952005863, valid_loss: 0.2664161607434478\n",
            "epoch:66,batch: 200, train_loss: 0.26123156435787676, valid_loss: 0.2666036084860186\n",
            "epoch:66,batch: 300, train_loss: 0.27070686385035514, valid_loss: 0.2694205765294123\n",
            "epoch:67,batch: 100, train_loss: 0.26586660370230675, valid_loss: 0.2661454759632485\n",
            "epoch:67,batch: 200, train_loss: 0.27266618877649307, valid_loss: 0.2661102485807636\n",
            "epoch:67,batch: 300, train_loss: 0.2662175203859806, valid_loss: 0.2671216259462924\n",
            "epoch:68,batch: 100, train_loss: 0.27810753621160983, valid_loss: 0.27182154300846634\n",
            "epoch:68,batch: 200, train_loss: 0.26354129903018475, valid_loss: 0.2655171992469437\n",
            "epoch:68,batch: 300, train_loss: 0.2675213625282049, valid_loss: 0.26520950271736216\n",
            "epoch:69,batch: 100, train_loss: 0.2759664200246334, valid_loss: 0.26572947788842116\n",
            "epoch:69,batch: 200, train_loss: 0.2526237268000841, valid_loss: 0.2644106424118899\n",
            "epoch:69,batch: 300, train_loss: 0.265120750144124, valid_loss: 0.26458530552402326\n",
            "epoch:70,batch: 100, train_loss: 0.27415252566337583, valid_loss: 0.26544697233770465\n",
            "epoch:70,batch: 200, train_loss: 0.2614691638946533, valid_loss: 0.26666577216945114\n",
            "epoch:70,batch: 300, train_loss: 0.26099792778491976, valid_loss: 0.26585962359286563\n",
            "epoch:71,batch: 100, train_loss: 0.2671825176477432, valid_loss: 0.2693748641994935\n",
            "epoch:71,batch: 200, train_loss: 0.2514412942528725, valid_loss: 0.2666677956151057\n",
            "epoch:71,batch: 300, train_loss: 0.26661164619028566, valid_loss: 0.2697339827501321\n",
            "epoch:72,batch: 100, train_loss: 0.2508862783014774, valid_loss: 0.26497754699821713\n",
            "epoch:72,batch: 200, train_loss: 0.25742173559963705, valid_loss: 0.2644112972141821\n",
            "epoch:72,batch: 300, train_loss: 0.2873781871795654, valid_loss: 0.2705082663252384\n",
            "epoch:73,batch: 100, train_loss: 0.275856863707304, valid_loss: 0.2656343197332153\n",
            "epoch:73,batch: 200, train_loss: 0.25846447221934793, valid_loss: 0.26540606563227087\n",
            "epoch:73,batch: 300, train_loss: 0.27545657232403753, valid_loss: 0.2684971080927909\n",
            "epoch:74,batch: 100, train_loss: 0.26054304391145705, valid_loss: 0.26268685486497756\n",
            "epoch:74,batch: 200, train_loss: 0.2702003742009401, valid_loss: 0.2652563348035269\n",
            "epoch:74,batch: 300, train_loss: 0.2544852888584137, valid_loss: 0.2850508790793298\n",
            "epoch:75,batch: 100, train_loss: 0.26932592391967775, valid_loss: 0.26370396465063095\n",
            "epoch:75,batch: 200, train_loss: 0.2626116290688515, valid_loss: 0.2643168054615395\n",
            "epoch:75,batch: 300, train_loss: 0.2601525662839413, valid_loss: 0.2823277029998695\n",
            "epoch:76,batch: 100, train_loss: 0.25008107244968414, valid_loss: 0.26346826232686826\n",
            "epoch:76,batch: 200, train_loss: 0.2656931936740875, valid_loss: 0.26489223030549064\n",
            "epoch:76,batch: 300, train_loss: 0.2748238456994295, valid_loss: 0.26467282834309563\n",
            "epoch:77,batch: 100, train_loss: 0.2762485560029745, valid_loss: 0.28569023462035986\n",
            "epoch:77,batch: 200, train_loss: 0.25833077501505614, valid_loss: 0.26499829516757895\n",
            "epoch:77,batch: 300, train_loss: 0.2615378176420927, valid_loss: 0.26332024664063997\n",
            "epoch:78,batch: 100, train_loss: 0.26280960746109483, valid_loss: 0.2776259935543507\n",
            "epoch:78,batch: 200, train_loss: 0.2587426679581404, valid_loss: 0.2619882104328916\n",
            "epoch:78,batch: 300, train_loss: 0.2548317775130272, valid_loss: 0.2632450802793986\n",
            "epoch:79,batch: 100, train_loss: 0.2664875016361475, valid_loss: 0.2664067895163464\n",
            "epoch:79,batch: 200, train_loss: 0.25450206056237223, valid_loss: 0.2927041959158982\n",
            "epoch:79,batch: 300, train_loss: 0.2709841998666525, valid_loss: 0.2632221611617487\n",
            "epoch:80,batch: 100, train_loss: 0.2704126439243555, valid_loss: 0.27424002353903615\n",
            "epoch:80,batch: 200, train_loss: 0.24374490179121494, valid_loss: 0.2639577401966988\n",
            "epoch:80,batch: 300, train_loss: 0.2671535336971283, valid_loss: 0.2694495246285879\n",
            "epoch:81,batch: 100, train_loss: 0.2664659094810486, valid_loss: 0.26564168486791323\n",
            "epoch:81,batch: 200, train_loss: 0.2584344694763422, valid_loss: 0.26151319687502295\n",
            "epoch:81,batch: 300, train_loss: 0.2537247647345066, valid_loss: 0.26457037333446215\n",
            "epoch:82,batch: 100, train_loss: 0.26256540678441526, valid_loss: 0.26475504455687127\n",
            "epoch:82,batch: 200, train_loss: 0.2596617241948843, valid_loss: 0.26500915415302106\n",
            "epoch:82,batch: 300, train_loss: 0.2758526124805212, valid_loss: 0.26102375720120685\n",
            "epoch:83,batch: 100, train_loss: 0.257234567925334, valid_loss: 0.26010564742963527\n",
            "epoch:83,batch: 200, train_loss: 0.2632383278012276, valid_loss: 0.2660173527613471\n",
            "epoch:83,batch: 300, train_loss: 0.2566791225969791, valid_loss: 0.26660076010076306\n",
            "epoch:84,batch: 100, train_loss: 0.25256454274058343, valid_loss: 0.2614223673185216\n",
            "epoch:84,batch: 200, train_loss: 0.26090112172067165, valid_loss: 0.2640045405000071\n",
            "epoch:84,batch: 300, train_loss: 0.26032806672155856, valid_loss: 0.2644865311022046\n",
            "epoch:85,batch: 100, train_loss: 0.25307619765400885, valid_loss: 0.27962212981302526\n",
            "epoch:85,batch: 200, train_loss: 0.2508614368736744, valid_loss: 0.26077270139999026\n",
            "epoch:85,batch: 300, train_loss: 0.2784526036679745, valid_loss: 0.26184950486014164\n",
            "epoch:86,batch: 100, train_loss: 0.2688386821746826, valid_loss: 0.26961591745479196\n",
            "epoch:86,batch: 200, train_loss: 0.25367458269000054, valid_loss: 0.26284490506860275\n",
            "epoch:86,batch: 300, train_loss: 0.24764174997806548, valid_loss: 0.26242567014090623\n",
            "epoch:87,batch: 100, train_loss: 0.25145921632647517, valid_loss: 0.2741206682558301\n",
            "epoch:87,batch: 200, train_loss: 0.2551930350810289, valid_loss: 0.2625296820379511\n",
            "epoch:87,batch: 300, train_loss: 0.25822904966771604, valid_loss: 0.2637370280638526\n",
            "epoch:88,batch: 100, train_loss: 0.25672350995242593, valid_loss: 0.26163860099225106\n",
            "epoch:88,batch: 200, train_loss: 0.2530389303714037, valid_loss: 0.2881878852089749\n",
            "epoch:88,batch: 300, train_loss: 0.2704031954705715, valid_loss: 0.26451751801975165\n",
            "epoch:89,batch: 100, train_loss: 0.25728965021669864, valid_loss: 0.26365518852879727\n",
            "epoch:89,batch: 200, train_loss: 0.25916822381317617, valid_loss: 0.2761710426664051\n",
            "epoch:89,batch: 300, train_loss: 0.25050831712782384, valid_loss: 0.2616724685966214\n",
            "epoch:90,batch: 100, train_loss: 0.2691524127870798, valid_loss: 0.26105711203587206\n",
            "epoch:90,batch: 200, train_loss: 0.2541551823914051, valid_loss: 0.2767250427528273\n",
            "epoch:90,batch: 300, train_loss: 0.25598338834941387, valid_loss: 0.26369218660306326\n",
            "epoch:91,batch: 100, train_loss: 0.2782429495453835, valid_loss: 0.2757101132522655\n",
            "epoch:91,batch: 200, train_loss: 0.2569382852315903, valid_loss: 0.2626537341105787\n",
            "epoch:91,batch: 300, train_loss: 0.251984986923635, valid_loss: 0.26264218467323086\n",
            "epoch:92,batch: 100, train_loss: 0.2640053269267082, valid_loss: 0.2596056004868278\n",
            "epoch:92,batch: 200, train_loss: 0.2551874239742756, valid_loss: 0.2681106755250617\n",
            "epoch:92,batch: 300, train_loss: 0.24944356702268122, valid_loss: 0.2579323179736922\n",
            "epoch:93,batch: 100, train_loss: 0.24655403286218644, valid_loss: 0.2592281336082688\n",
            "epoch:93,batch: 200, train_loss: 0.2684767212718725, valid_loss: 0.2637536865246447\n",
            "epoch:93,batch: 300, train_loss: 0.2514323791116476, valid_loss: 0.25872677932434446\n",
            "epoch:94,batch: 100, train_loss: 0.25976209200918676, valid_loss: 0.2848517210989059\n",
            "epoch:94,batch: 200, train_loss: 0.2589640513807535, valid_loss: 0.26774944195264505\n",
            "epoch:94,batch: 300, train_loss: 0.25171695709228514, valid_loss: 0.26193370104213304\n",
            "epoch:95,batch: 100, train_loss: 0.25793883584439753, valid_loss: 0.25956769844021976\n",
            "epoch:95,batch: 200, train_loss: 0.2514512147754431, valid_loss: 0.2600097473286375\n",
            "epoch:95,batch: 300, train_loss: 0.2603572688996792, valid_loss: 0.27490394647362865\n",
            "epoch:96,batch: 100, train_loss: 0.2531237179040909, valid_loss: 0.2656061014608492\n",
            "epoch:96,batch: 200, train_loss: 0.2592897154390812, valid_loss: 0.26032354362026044\n",
            "epoch:96,batch: 300, train_loss: 0.2524621937423944, valid_loss: 0.25966351517016373\n",
            "epoch:97,batch: 100, train_loss: 0.25432366751134394, valid_loss: 0.2619437098125868\n",
            "epoch:97,batch: 200, train_loss: 0.249960745126009, valid_loss: 0.25858584583937366\n",
            "epoch:97,batch: 300, train_loss: 0.2717084476351738, valid_loss: 0.2617893965938423\n",
            "epoch:98,batch: 100, train_loss: 0.2471684283018112, valid_loss: 0.25833395066895065\n",
            "epoch:98,batch: 200, train_loss: 0.25903702519834043, valid_loss: 0.261434861564938\n",
            "epoch:98,batch: 300, train_loss: 0.2490176348388195, valid_loss: 0.2720393568654604\n",
            "epoch:99,batch: 100, train_loss: 0.2500949962437153, valid_loss: 0.2587886145220527\n",
            "epoch:99,batch: 200, train_loss: 0.2613286375254393, valid_loss: 0.25900703979821144\n",
            "epoch:99,batch: 300, train_loss: 0.24854691319167613, valid_loss: 0.26375617379251914\n",
            "epoch:100,batch: 100, train_loss: 0.2469276535511017, valid_loss: 0.2604358786646324\n",
            "epoch:100,batch: 200, train_loss: 0.26725898377597335, valid_loss: 0.25825560819121857\n",
            "epoch:100,batch: 300, train_loss: 0.2511826343089342, valid_loss: 0.2788955142037778\n",
            "epoch:101,batch: 100, train_loss: 0.25413171879947183, valid_loss: 0.25850060544436493\n",
            "epoch:101,batch: 200, train_loss: 0.2610640989989042, valid_loss: 0.2849264261843283\n",
            "epoch:101,batch: 300, train_loss: 0.255228528752923, valid_loss: 0.2579052071028118\n",
            "epoch:102,batch: 100, train_loss: 0.24578158654272556, valid_loss: 0.25688366799414913\n",
            "epoch:102,batch: 200, train_loss: 0.2540553992241621, valid_loss: 0.2572074741879596\n",
            "epoch:102,batch: 300, train_loss: 0.25485794849693777, valid_loss: 0.2588430857167968\n",
            "epoch:103,batch: 100, train_loss: 0.24265712045133114, valid_loss: 0.2599118974389909\n",
            "epoch:103,batch: 200, train_loss: 0.24716040387749671, valid_loss: 0.25897717777686785\n",
            "epoch:103,batch: 300, train_loss: 0.2518637851625681, valid_loss: 0.25843641699492176\n",
            "epoch:104,batch: 100, train_loss: 0.2515244743973017, valid_loss: 0.2561969205359869\n",
            "epoch:104,batch: 200, train_loss: 0.25175637312233445, valid_loss: 0.25827095468964756\n",
            "epoch:104,batch: 300, train_loss: 0.2563513818383217, valid_loss: 0.26812819344333455\n",
            "epoch:105,batch: 100, train_loss: 0.2562185611575842, valid_loss: 0.25861293049175527\n",
            "epoch:105,batch: 200, train_loss: 0.24951467730104923, valid_loss: 0.25872734688882587\n",
            "epoch:105,batch: 300, train_loss: 0.247139340788126, valid_loss: 0.25853647508575944\n",
            "epoch:106,batch: 100, train_loss: 0.25920200601220134, valid_loss: 0.2578592764425881\n",
            "epoch:106,batch: 200, train_loss: 0.24439856499433518, valid_loss: 0.2568460490696038\n",
            "epoch:106,batch: 300, train_loss: 0.26064777694642544, valid_loss: 0.2596208113658277\n",
            "epoch:107,batch: 100, train_loss: 0.25241312354803086, valid_loss: 0.2637018812230871\n",
            "epoch:107,batch: 200, train_loss: 0.2451990069448948, valid_loss: 0.2571056720010842\n",
            "epoch:107,batch: 300, train_loss: 0.2490699055045843, valid_loss: 0.2566551533115061\n",
            "epoch:108,batch: 100, train_loss: 0.25134138360619546, valid_loss: 0.26110783370235296\n",
            "epoch:108,batch: 200, train_loss: 0.24617188908159732, valid_loss: 0.25762342982277087\n",
            "epoch:108,batch: 300, train_loss: 0.25419384218752383, valid_loss: 0.2561354101458682\n",
            "epoch:109,batch: 100, train_loss: 0.2547368983924389, valid_loss: 0.25662037013452266\n",
            "epoch:109,batch: 200, train_loss: 0.2581999059766531, valid_loss: 0.2551576784705814\n",
            "epoch:109,batch: 300, train_loss: 0.2464967180043459, valid_loss: 0.26636879853432693\n",
            "epoch:110,batch: 100, train_loss: 0.2384292170405388, valid_loss: 0.25956792308937143\n",
            "epoch:110,batch: 200, train_loss: 0.2525043913722038, valid_loss: 0.25991671192872373\n",
            "epoch:110,batch: 300, train_loss: 0.2432599849998951, valid_loss: 0.257381837296335\n",
            "epoch:111,batch: 100, train_loss: 0.24906217731535435, valid_loss: 0.25578908518522603\n",
            "epoch:111,batch: 200, train_loss: 0.2377743923664093, valid_loss: 0.2567287597663795\n",
            "epoch:111,batch: 300, train_loss: 0.24961717531085015, valid_loss: 0.25720276664706726\n",
            "epoch:112,batch: 100, train_loss: 0.2431595679372549, valid_loss: 0.2589192102321341\n",
            "epoch:112,batch: 200, train_loss: 0.25740208722651003, valid_loss: 0.25578308907113495\n",
            "epoch:112,batch: 300, train_loss: 0.25145569115877153, valid_loss: 0.25508080998176264\n",
            "epoch:113,batch: 100, train_loss: 0.2586851193010807, valid_loss: 0.25608728637423694\n",
            "epoch:113,batch: 200, train_loss: 0.2542722949385643, valid_loss: 0.26116520147534866\n",
            "epoch:113,batch: 300, train_loss: 0.25044691778719425, valid_loss: 0.2541803215877919\n",
            "epoch:114,batch: 100, train_loss: 0.25909383825957777, valid_loss: 0.2565815784508669\n",
            "epoch:114,batch: 200, train_loss: 0.25125536784529684, valid_loss: 0.26628379514323003\n",
            "epoch:114,batch: 300, train_loss: 0.2360737431049347, valid_loss: 0.2694326389250876\n",
            "epoch:115,batch: 100, train_loss: 0.2535429160296917, valid_loss: 0.25614829974461206\n",
            "epoch:115,batch: 200, train_loss: 0.2383704899996519, valid_loss: 0.2571852026106436\n",
            "epoch:115,batch: 300, train_loss: 0.250178250297904, valid_loss: 0.25776040582339976\n",
            "epoch:116,batch: 100, train_loss: 0.24084916673600673, valid_loss: 0.2549979498114767\n",
            "epoch:116,batch: 200, train_loss: 0.2549148405343294, valid_loss: 0.2567638197465788\n",
            "epoch:116,batch: 300, train_loss: 0.23816274717450142, valid_loss: 0.2601678151684471\n",
            "epoch:117,batch: 100, train_loss: 0.2434747052192688, valid_loss: 0.2628148484833633\n",
            "epoch:117,batch: 200, train_loss: 0.2445395213365555, valid_loss: 0.25663541625194913\n",
            "epoch:117,batch: 300, train_loss: 0.2502767775952816, valid_loss: 0.2570801206593272\n",
            "epoch:118,batch: 100, train_loss: 0.2588778229057789, valid_loss: 0.2578411009869998\n",
            "epoch:118,batch: 200, train_loss: 0.24512121595442296, valid_loss: 0.2547848154849644\n",
            "epoch:118,batch: 300, train_loss: 0.2288207123428583, valid_loss: 0.2551175360626812\n",
            "epoch:119,batch: 100, train_loss: 0.24914686441421507, valid_loss: 0.26086913586794575\n",
            "epoch:119,batch: 200, train_loss: 0.2488809049874544, valid_loss: 0.2782139349021489\n",
            "epoch:119,batch: 300, train_loss: 0.24093161299824714, valid_loss: 0.2560961694468426\n",
            "epoch:120,batch: 100, train_loss: 0.24827918477356434, valid_loss: 0.2564399531936344\n",
            "epoch:120,batch: 200, train_loss: 0.23870659835636615, valid_loss: 0.2547757122901422\n",
            "epoch:120,batch: 300, train_loss: 0.25300397008657455, valid_loss: 0.2564151819936837\n",
            "epoch:121,batch: 100, train_loss: 0.23621620804071428, valid_loss: 0.26007624875895585\n",
            "epoch:121,batch: 200, train_loss: 0.2508496867120266, valid_loss: 0.25378648955610733\n",
            "epoch:121,batch: 300, train_loss: 0.24074981108307839, valid_loss: 0.25502127305239064\n",
            "epoch:122,batch: 100, train_loss: 0.2547267213463783, valid_loss: 0.25948521523158763\n",
            "epoch:122,batch: 200, train_loss: 0.2541281817108393, valid_loss: 0.2537024536087543\n",
            "epoch:122,batch: 300, train_loss: 0.24791137419641018, valid_loss: 0.2554726507278937\n",
            "epoch:123,batch: 100, train_loss: 0.2535667394846678, valid_loss: 0.2818283140659332\n",
            "epoch:123,batch: 200, train_loss: 0.2389571838825941, valid_loss: 0.25330324574739116\n",
            "epoch:123,batch: 300, train_loss: 0.2508104631304741, valid_loss: 0.2563217413387721\n",
            "epoch:124,batch: 100, train_loss: 0.257050212547183, valid_loss: 0.2552550106863432\n",
            "epoch:124,batch: 200, train_loss: 0.24166868083178997, valid_loss: 0.2555868530386611\n",
            "epoch:124,batch: 300, train_loss: 0.24275126919150353, valid_loss: 0.26648351880191246\n",
            "epoch:125,batch: 100, train_loss: 0.2558195839077234, valid_loss: 0.2577629606180553\n",
            "epoch:125,batch: 200, train_loss: 0.2441442082822323, valid_loss: 0.25400799399689783\n",
            "epoch:125,batch: 300, train_loss: 0.24441666074097157, valid_loss: 0.28538264645428596\n",
            "epoch:126,batch: 100, train_loss: 0.25075558565557005, valid_loss: 0.2609571100601667\n",
            "epoch:126,batch: 200, train_loss: 0.2459857040643692, valid_loss: 0.27992069985293133\n",
            "epoch:126,batch: 300, train_loss: 0.24762159757316113, valid_loss: 0.25441407723517356\n",
            "epoch:127,batch: 100, train_loss: 0.245473268404603, valid_loss: 0.25706215636639657\n",
            "epoch:127,batch: 200, train_loss: 0.24619472935795783, valid_loss: 0.2539408420082889\n",
            "epoch:127,batch: 300, train_loss: 0.2465622380375862, valid_loss: 0.25258602560320986\n",
            "epoch:128,batch: 100, train_loss: 0.2347858752310276, valid_loss: 0.25643604861784586\n",
            "epoch:128,batch: 200, train_loss: 0.24599733851850034, valid_loss: 0.265724544943888\n",
            "epoch:128,batch: 300, train_loss: 0.24727114409208298, valid_loss: 0.2686355809056306\n",
            "epoch:129,batch: 100, train_loss: 0.24307909056544305, valid_loss: 0.2578184841365754\n",
            "epoch:129,batch: 200, train_loss: 0.23460250027477741, valid_loss: 0.252018230813968\n",
            "epoch:129,batch: 300, train_loss: 0.2462747311592102, valid_loss: 0.2538497779942766\n",
            "epoch:130,batch: 100, train_loss: 0.2558807516098022, valid_loss: 0.2520652273787728\n",
            "epoch:130,batch: 200, train_loss: 0.2398001540452242, valid_loss: 0.26549205051947244\n",
            "epoch:130,batch: 300, train_loss: 0.23388763941824436, valid_loss: 0.2537913319619396\n",
            "epoch:131,batch: 100, train_loss: 0.25112806566059587, valid_loss: 0.258892978105364\n",
            "epoch:131,batch: 200, train_loss: 0.2479933064058423, valid_loss: 0.25109298517809636\n",
            "epoch:131,batch: 300, train_loss: 0.23969901144504546, valid_loss: 0.25192156754717043\n",
            "epoch:132,batch: 100, train_loss: 0.24110614761710167, valid_loss: 0.2606107313893264\n",
            "epoch:132,batch: 200, train_loss: 0.2545898304879665, valid_loss: 0.2551089951509162\n",
            "epoch:132,batch: 300, train_loss: 0.23697066381573678, valid_loss: 0.25342723265101635\n",
            "epoch:133,batch: 100, train_loss: 0.24788317576050758, valid_loss: 0.26659760124321225\n",
            "epoch:133,batch: 200, train_loss: 0.2447339625656605, valid_loss: 0.25427645765527895\n",
            "epoch:133,batch: 300, train_loss: 0.2407190605998039, valid_loss: 0.25736310436755794\n",
            "epoch:134,batch: 100, train_loss: 0.24515594981610775, valid_loss: 0.26663996118910704\n",
            "epoch:134,batch: 200, train_loss: 0.23228434301912784, valid_loss: 0.25156242853101296\n",
            "epoch:134,batch: 300, train_loss: 0.23924727071076632, valid_loss: 0.25666585508026657\n",
            "epoch:135,batch: 100, train_loss: 0.23853698015213012, valid_loss: 0.2540566216729864\n",
            "epoch:135,batch: 200, train_loss: 0.25611629255115986, valid_loss: 0.2552125736882415\n",
            "epoch:135,batch: 300, train_loss: 0.24723228096961974, valid_loss: 0.25214361097616483\n",
            "epoch:136,batch: 100, train_loss: 0.24003137819468975, valid_loss: 0.2601098951093758\n",
            "epoch:136,batch: 200, train_loss: 0.23801050931215287, valid_loss: 0.2651339596302449\n",
            "epoch:136,batch: 300, train_loss: 0.2544966033101082, valid_loss: 0.2625821574202067\n",
            "epoch:137,batch: 100, train_loss: 0.24632828406989574, valid_loss: 0.2540929080753387\n",
            "epoch:137,batch: 200, train_loss: 0.24363659270107746, valid_loss: 0.2652901827911787\n",
            "epoch:137,batch: 300, train_loss: 0.24206328816711903, valid_loss: 0.2531525377608553\n",
            "epoch:138,batch: 100, train_loss: 0.24445794865489007, valid_loss: 0.25281355369694625\n",
            "epoch:138,batch: 200, train_loss: 0.24584401365369557, valid_loss: 0.2521326681481132\n",
            "epoch:138,batch: 300, train_loss: 0.23902294486761094, valid_loss: 0.25304640112798427\n",
            "epoch:139,batch: 100, train_loss: 0.23604040190577508, valid_loss: 0.25275095259841485\n",
            "epoch:139,batch: 200, train_loss: 0.2517669900506735, valid_loss: 0.2517420973015737\n",
            "epoch:139,batch: 300, train_loss: 0.24125361122190953, valid_loss: 0.256690997488891\n",
            "epoch:140,batch: 100, train_loss: 0.24448617126792668, valid_loss: 0.26558053045521807\n",
            "epoch:140,batch: 200, train_loss: 0.23973822444677353, valid_loss: 0.2557844140484363\n",
            "epoch:140,batch: 300, train_loss: 0.2389578879624605, valid_loss: 0.2593312499266637\n",
            "epoch:141,batch: 100, train_loss: 0.2372744468599558, valid_loss: 0.25642293418132805\n",
            "epoch:141,batch: 200, train_loss: 0.2517890341579914, valid_loss: 0.26109196773812743\n",
            "epoch:141,batch: 300, train_loss: 0.23148600779473782, valid_loss: 0.25299058750837666\n",
            "epoch:142,batch: 100, train_loss: 0.231159885302186, valid_loss: 0.25924428373197966\n",
            "epoch:142,batch: 200, train_loss: 0.24086053408682345, valid_loss: 0.2617216953371145\n",
            "epoch:142,batch: 300, train_loss: 0.24885161146521567, valid_loss: 0.25931008659963367\n",
            "epoch:143,batch: 100, train_loss: 0.2587397242337465, valid_loss: 0.2593566959228697\n",
            "epoch:143,batch: 200, train_loss: 0.24739946648478509, valid_loss: 0.2528344889796233\n",
            "epoch:143,batch: 300, train_loss: 0.22217278108000754, valid_loss: 0.2514312109803852\n",
            "epoch:144,batch: 100, train_loss: 0.2363992953300476, valid_loss: 0.25330681570723085\n",
            "epoch:144,batch: 200, train_loss: 0.24675836943089963, valid_loss: 0.25583609061527857\n",
            "epoch:144,batch: 300, train_loss: 0.23374857842922211, valid_loss: 0.25381562170348587\n",
            "epoch:145,batch: 100, train_loss: 0.24653466165065765, valid_loss: 0.273213493484485\n",
            "epoch:145,batch: 200, train_loss: 0.2456764957308769, valid_loss: 0.25069821465619\n",
            "epoch:145,batch: 300, train_loss: 0.22987234961241484, valid_loss: 0.2511488964097409\n",
            "epoch:146,batch: 100, train_loss: 0.23875093773007394, valid_loss: 0.251869274091117\n",
            "epoch:146,batch: 200, train_loss: 0.2333118249475956, valid_loss: 0.26077776491830623\n",
            "epoch:146,batch: 300, train_loss: 0.2400324036180973, valid_loss: 0.2602638916312894\n",
            "epoch:147,batch: 100, train_loss: 0.2409508618712425, valid_loss: 0.25816096827576435\n",
            "epoch:147,batch: 200, train_loss: 0.2488105806708336, valid_loss: 0.25366766137790075\n",
            "epoch:147,batch: 300, train_loss: 0.23922600626945495, valid_loss: 0.2625400806906857\n",
            "epoch:148,batch: 100, train_loss: 0.2425126550346613, valid_loss: 0.2504337812903561\n",
            "epoch:148,batch: 200, train_loss: 0.24195149675011635, valid_loss: 0.2522891496744337\n",
            "epoch:148,batch: 300, train_loss: 0.23569803684949875, valid_loss: 0.25325620532790316\n",
            "epoch:149,batch: 100, train_loss: 0.2513247431814671, valid_loss: 0.24944615420661395\n",
            "epoch:149,batch: 200, train_loss: 0.2299086022377014, valid_loss: 0.25395531914656677\n",
            "epoch:149,batch: 300, train_loss: 0.24172024197876454, valid_loss: 0.2518641892113263\n",
            "epoch:150,batch: 100, train_loss: 0.2525803276151419, valid_loss: 0.2552681233309492\n",
            "epoch:150,batch: 200, train_loss: 0.23590179808437825, valid_loss: 0.2578421968825256\n",
            "epoch:150,batch: 300, train_loss: 0.23485667057335377, valid_loss: 0.2630911407214177\n",
            "epoch:151,batch: 100, train_loss: 0.23639374062418939, valid_loss: 0.250567204495774\n",
            "epoch:151,batch: 200, train_loss: 0.23472185395658016, valid_loss: 0.25085602983643734\n",
            "epoch:151,batch: 300, train_loss: 0.24388528659939765, valid_loss: 0.2497017423186121\n",
            "epoch:152,batch: 100, train_loss: 0.2343587190657854, valid_loss: 0.25376345453005805\n",
            "epoch:152,batch: 200, train_loss: 0.2404658918082714, valid_loss: 0.2531634763449053\n",
            "epoch:152,batch: 300, train_loss: 0.2359897831082344, valid_loss: 0.24983917271034628\n",
            "epoch:153,batch: 100, train_loss: 0.23350602455437183, valid_loss: 0.25139673195685014\n",
            "epoch:153,batch: 200, train_loss: 0.24295779600739478, valid_loss: 0.2497502438063863\n",
            "epoch:153,batch: 300, train_loss: 0.23726583793759345, valid_loss: 0.24935758151585544\n",
            "epoch:154,batch: 100, train_loss: 0.25615221984684466, valid_loss: 0.26405773189248916\n",
            "epoch:154,batch: 200, train_loss: 0.23282402724027634, valid_loss: 0.2520621088863928\n",
            "epoch:154,batch: 300, train_loss: 0.2340680157393217, valid_loss: 0.24923974567953544\n",
            "epoch:155,batch: 100, train_loss: 0.23494081012904644, valid_loss: 0.25007400557964665\n",
            "epoch:155,batch: 200, train_loss: 0.2343435077369213, valid_loss: 0.2504431014951271\n",
            "epoch:155,batch: 300, train_loss: 0.24274700477719308, valid_loss: 0.25146239986525304\n",
            "epoch:156,batch: 100, train_loss: 0.2338088681176305, valid_loss: 0.25068727041347116\n",
            "epoch:156,batch: 200, train_loss: 0.2457331870496273, valid_loss: 0.24934974067573307\n",
            "epoch:156,batch: 300, train_loss: 0.23565215788781643, valid_loss: 0.24832830832728855\n",
            "epoch:157,batch: 100, train_loss: 0.22902209855616093, valid_loss: 0.2522672408177883\n",
            "epoch:157,batch: 200, train_loss: 0.2285109055787325, valid_loss: 0.251841413399464\n",
            "epoch:157,batch: 300, train_loss: 0.24040193177759647, valid_loss: 0.24969816915219342\n",
            "epoch:158,batch: 100, train_loss: 0.24968084067106247, valid_loss: 0.24989620718774916\n",
            "epoch:158,batch: 200, train_loss: 0.22889141831547022, valid_loss: 0.2613690558679496\n",
            "epoch:158,batch: 300, train_loss: 0.2437782721966505, valid_loss: 0.2555777651222446\n",
            "epoch:159,batch: 100, train_loss: 0.24718878675252198, valid_loss: 0.25057347081130066\n",
            "epoch:159,batch: 200, train_loss: 0.227455178424716, valid_loss: 0.2500979558010645\n",
            "epoch:159,batch: 300, train_loss: 0.23166537888348102, valid_loss: 0.24999761326780803\n",
            "epoch:160,batch: 100, train_loss: 0.245095367282629, valid_loss: 0.25056997696055644\n",
            "epoch:160,batch: 200, train_loss: 0.229756303653121, valid_loss: 0.2483869301556032\n",
            "epoch:160,batch: 300, train_loss: 0.23956212885677813, valid_loss: 0.24910603435356407\n",
            "epoch:161,batch: 100, train_loss: 0.22867917016148567, valid_loss: 0.25080922019632557\n",
            "epoch:161,batch: 200, train_loss: 0.23178683795034885, valid_loss: 0.250012971932375\n",
            "epoch:161,batch: 300, train_loss: 0.2386124459654093, valid_loss: 0.26787307424635826\n",
            "epoch:162,batch: 100, train_loss: 0.24103585615754128, valid_loss: 0.25114253361391115\n",
            "epoch:162,batch: 200, train_loss: 0.2382163845002651, valid_loss: 0.2513779516838774\n",
            "epoch:162,batch: 300, train_loss: 0.22199247196316718, valid_loss: 0.2489311439515669\n",
            "epoch:163,batch: 100, train_loss: 0.237021301984787, valid_loss: 0.2674906180252003\n",
            "epoch:163,batch: 200, train_loss: 0.2241077256202698, valid_loss: 0.24968069697482675\n",
            "epoch:163,batch: 300, train_loss: 0.23919438295066356, valid_loss: 0.24871935176698468\n",
            "epoch:164,batch: 100, train_loss: 0.23148936845362186, valid_loss: 0.25038294201787514\n",
            "epoch:164,batch: 200, train_loss: 0.23602003663778304, valid_loss: 0.2553160040627552\n",
            "epoch:164,batch: 300, train_loss: 0.22234587632119657, valid_loss: 0.25094596466308905\n",
            "epoch:165,batch: 100, train_loss: 0.24089906603097916, valid_loss: 0.24898994129292573\n",
            "epoch:165,batch: 200, train_loss: 0.2370767156034708, valid_loss: 0.25356529747383505\n",
            "epoch:165,batch: 300, train_loss: 0.2357171481847763, valid_loss: 0.25274699643442905\n",
            "epoch:166,batch: 100, train_loss: 0.23191987790167332, valid_loss: 0.255633316462553\n",
            "epoch:166,batch: 200, train_loss: 0.23360997036099435, valid_loss: 0.25848494540852834\n",
            "epoch:166,batch: 300, train_loss: 0.24733723916113376, valid_loss: 0.2642088976653316\n",
            "epoch:167,batch: 100, train_loss: 0.25301790572702887, valid_loss: 0.24723974577611005\n",
            "epoch:167,batch: 200, train_loss: 0.2336363884061575, valid_loss: 0.25160980130298227\n",
            "epoch:167,batch: 300, train_loss: 0.21754929944872856, valid_loss: 0.2518297249003302\n",
            "epoch:168,batch: 100, train_loss: 0.22611105516552926, valid_loss: 0.24979365673623508\n",
            "epoch:168,batch: 200, train_loss: 0.2425122334808111, valid_loss: 0.24815036790280404\n",
            "epoch:168,batch: 300, train_loss: 0.23448751620948316, valid_loss: 0.25047117141605935\n",
            "epoch:169,batch: 100, train_loss: 0.23260596252977847, valid_loss: 0.24677381070354318\n",
            "epoch:169,batch: 200, train_loss: 0.2351034802943468, valid_loss: 0.247997246116777\n",
            "epoch:169,batch: 300, train_loss: 0.2287434583157301, valid_loss: 0.24893326038801217\n",
            "epoch:170,batch: 100, train_loss: 0.23648187585175037, valid_loss: 0.24997478798974918\n",
            "epoch:170,batch: 200, train_loss: 0.23545198857784272, valid_loss: 0.25233226948523824\n",
            "epoch:170,batch: 300, train_loss: 0.22806626506149769, valid_loss: 0.25089524959838844\n",
            "epoch:171,batch: 100, train_loss: 0.22090240627527236, valid_loss: 0.2510375623838811\n",
            "epoch:171,batch: 200, train_loss: 0.23551858820021151, valid_loss: 0.2964221873803984\n",
            "epoch:171,batch: 300, train_loss: 0.2364800713211298, valid_loss: 0.25693109574951706\n",
            "epoch:172,batch: 100, train_loss: 0.23346362702548504, valid_loss: 0.25704506520606296\n",
            "epoch:172,batch: 200, train_loss: 0.24068620163947343, valid_loss: 0.24892669982170756\n",
            "epoch:172,batch: 300, train_loss: 0.22732786811888217, valid_loss: 0.24776777049785928\n",
            "epoch:173,batch: 100, train_loss: 0.22290719754993915, valid_loss: 0.25449274725551846\n",
            "epoch:173,batch: 200, train_loss: 0.23368096590042114, valid_loss: 0.24685426707131952\n",
            "epoch:173,batch: 300, train_loss: 0.23150312565267087, valid_loss: 0.2675895804091345\n",
            "epoch:174,batch: 100, train_loss: 0.23920373290777205, valid_loss: 0.24900540813237806\n",
            "epoch:174,batch: 200, train_loss: 0.23037941619753838, valid_loss: 0.2546231724038909\n",
            "epoch:174,batch: 300, train_loss: 0.22613946571946145, valid_loss: 0.2565170730216594\n",
            "epoch:175,batch: 100, train_loss: 0.22963461458683013, valid_loss: 0.24932152476114564\n",
            "epoch:175,batch: 200, train_loss: 0.23020148552954198, valid_loss: 0.254432908058921\n",
            "epoch:175,batch: 300, train_loss: 0.24121166691184043, valid_loss: 0.2467451410580285\n",
            "epoch:176,batch: 100, train_loss: 0.23946692042052745, valid_loss: 0.2559282252494293\n",
            "epoch:176,batch: 200, train_loss: 0.2413764598220587, valid_loss: 0.2546719284374503\n",
            "epoch:176,batch: 300, train_loss: 0.2297589433938265, valid_loss: 0.2602550983428955\n",
            "epoch:177,batch: 100, train_loss: 0.22674407117068768, valid_loss: 0.2525901728415791\n",
            "epoch:177,batch: 200, train_loss: 0.21448082230985166, valid_loss: 0.25022463340170775\n",
            "epoch:177,batch: 300, train_loss: 0.2395929519087076, valid_loss: 0.24615942818831793\n",
            "epoch:178,batch: 100, train_loss: 0.2312558661773801, valid_loss: 0.2516571388025827\n",
            "epoch:178,batch: 200, train_loss: 0.2216941261291504, valid_loss: 0.24834501837627798\n",
            "epoch:178,batch: 300, train_loss: 0.23069947876036168, valid_loss: 0.2504481500651263\n",
            "epoch:179,batch: 100, train_loss: 0.23946169309318066, valid_loss: 0.255272051201591\n",
            "epoch:179,batch: 200, train_loss: 0.23069016516208649, valid_loss: 0.24950947206986102\n",
            "epoch:179,batch: 300, train_loss: 0.23041286077350379, valid_loss: 0.24934579677219632\n",
            "epoch:180,batch: 100, train_loss: 0.22382725354284047, valid_loss: 0.24694567749017401\n",
            "epoch:180,batch: 200, train_loss: 0.22515695117413997, valid_loss: 0.2616793223762814\n",
            "epoch:180,batch: 300, train_loss: 0.2387437929958105, valid_loss: 0.24769545101289508\n",
            "epoch:181,batch: 100, train_loss: 0.2399456053227186, valid_loss: 0.2514857698844958\n",
            "epoch:181,batch: 200, train_loss: 0.22167377412319184, valid_loss: 0.2507175054354004\n",
            "epoch:181,batch: 300, train_loss: 0.2340298243612051, valid_loss: 0.24628090688699408\n",
            "epoch:182,batch: 100, train_loss: 0.22701246090233326, valid_loss: 0.2553810764905773\n",
            "epoch:182,batch: 200, train_loss: 0.2397048420459032, valid_loss: 0.25187967840252046\n",
            "epoch:182,batch: 300, train_loss: 0.21749949663877488, valid_loss: 0.2525261633003814\n",
            "epoch:183,batch: 100, train_loss: 0.2413123571127653, valid_loss: 0.2559395462651796\n",
            "epoch:183,batch: 200, train_loss: 0.21850623652338982, valid_loss: 0.24855392265923415\n",
            "epoch:183,batch: 300, train_loss: 0.22407299123704433, valid_loss: 0.2477045352511768\n",
            "epoch:184,batch: 100, train_loss: 0.23264863818883896, valid_loss: 0.24933711363922192\n",
            "epoch:184,batch: 200, train_loss: 0.233166236653924, valid_loss: 0.24944797290276877\n",
            "epoch:184,batch: 300, train_loss: 0.2294316717982292, valid_loss: 0.24823782244060613\n",
            "epoch:185,batch: 100, train_loss: 0.24419644065201282, valid_loss: 0.25656602918347227\n",
            "epoch:185,batch: 200, train_loss: 0.2280330952256918, valid_loss: 0.2522623014600971\n",
            "epoch:185,batch: 300, train_loss: 0.21716266863048075, valid_loss: 0.2483328875578657\n",
            "epoch:186,batch: 100, train_loss: 0.21612934686243535, valid_loss: 0.257829261165631\n",
            "epoch:186,batch: 200, train_loss: 0.23391945123672486, valid_loss: 0.24518892404776585\n",
            "epoch:186,batch: 300, train_loss: 0.2303949723392725, valid_loss: 0.2519671142667155\n",
            "epoch:187,batch: 100, train_loss: 0.23724505975842475, valid_loss: 0.2629638121852392\n",
            "epoch:187,batch: 200, train_loss: 0.23270212590694428, valid_loss: 0.26320452280814133\n",
            "epoch:187,batch: 300, train_loss: 0.22407408222556113, valid_loss: 0.26001164222819895\n",
            "epoch:188,batch: 100, train_loss: 0.22774611562490463, valid_loss: 0.2539830727479126\n",
            "epoch:188,batch: 200, train_loss: 0.22844032101333142, valid_loss: 0.2541855586857735\n",
            "epoch:188,batch: 300, train_loss: 0.251956270262599, valid_loss: 0.26551193960859804\n",
            "epoch:189,batch: 100, train_loss: 0.2283695913851261, valid_loss: 0.2491693054384823\n",
            "epoch:189,batch: 200, train_loss: 0.23067917600274085, valid_loss: 0.24542105084733118\n",
            "epoch:189,batch: 300, train_loss: 0.21754721604287625, valid_loss: 0.26415131005305276\n",
            "epoch:190,batch: 100, train_loss: 0.2446728761866689, valid_loss: 0.24781656001187577\n",
            "epoch:190,batch: 200, train_loss: 0.21030356764793395, valid_loss: 0.247487499555455\n",
            "epoch:190,batch: 300, train_loss: 0.22572975553572178, valid_loss: 0.2478372455586361\n",
            "epoch:191,batch: 100, train_loss: 0.2307035606354475, valid_loss: 0.2781893599636947\n",
            "epoch:191,batch: 200, train_loss: 0.21468321062624454, valid_loss: 0.2528847362610358\n",
            "epoch:191,batch: 300, train_loss: 0.23219849221408367, valid_loss: 0.2505807110780402\n",
            "epoch:192,batch: 100, train_loss: 0.22604101657867431, valid_loss: 0.2529282485948333\n",
            "epoch:192,batch: 200, train_loss: 0.22424247935414315, valid_loss: 0.25209361258186874\n",
            "epoch:192,batch: 300, train_loss: 0.23062284462153912, valid_loss: 0.24553382226937934\n",
            "epoch:193,batch: 100, train_loss: 0.23821561470627783, valid_loss: 0.24614606269552738\n",
            "epoch:193,batch: 200, train_loss: 0.22055795207619666, valid_loss: 0.24659435837706434\n",
            "epoch:193,batch: 300, train_loss: 0.23912554308772088, valid_loss: 0.24563437228715873\n",
            "epoch:194,batch: 100, train_loss: 0.2336019517481327, valid_loss: 0.2458844591356531\n",
            "epoch:194,batch: 200, train_loss: 0.23269211143255233, valid_loss: 0.2519203038721145\n",
            "epoch:194,batch: 300, train_loss: 0.22951493289321662, valid_loss: 0.25119877045479\n",
            "epoch:195,batch: 100, train_loss: 0.23000921465456486, valid_loss: 0.2451853061778636\n",
            "epoch:195,batch: 200, train_loss: 0.241102102920413, valid_loss: 0.2467359688746024\n",
            "epoch:195,batch: 300, train_loss: 0.20287232674658298, valid_loss: 0.24549806674447242\n",
            "epoch:196,batch: 100, train_loss: 0.2276240549236536, valid_loss: 0.2509629101127009\n",
            "epoch:196,batch: 200, train_loss: 0.23407119754701852, valid_loss: 0.2502662315021587\n",
            "epoch:196,batch: 300, train_loss: 0.22603107132017614, valid_loss: 0.2467135035350353\n",
            "epoch:197,batch: 100, train_loss: 0.24292475409805775, valid_loss: 0.24749426656885992\n",
            "epoch:197,batch: 200, train_loss: 0.22044940654188394, valid_loss: 0.24627565545371816\n",
            "epoch:197,batch: 300, train_loss: 0.21995990172028543, valid_loss: 0.24650619612841668\n",
            "epoch:198,batch: 100, train_loss: 0.22686050660908222, valid_loss: 0.257482917720004\n",
            "epoch:198,batch: 200, train_loss: 0.22955944441258908, valid_loss: 0.24447096521152725\n",
            "epoch:198,batch: 300, train_loss: 0.22485557459294797, valid_loss: 0.25116958761517005\n",
            "epoch:199,batch: 100, train_loss: 0.2374892960488796, valid_loss: 0.25101802847053434\n",
            "epoch:199,batch: 200, train_loss: 0.2340973598510027, valid_loss: 0.2455034152239184\n",
            "epoch:199,batch: 300, train_loss: 0.21520721599459647, valid_loss: 0.2579610840995101\n",
            "epoch:200,batch: 100, train_loss: 0.2280830642580986, valid_loss: 0.24643550341642356\n",
            "epoch:200,batch: 200, train_loss: 0.22097596883773804, valid_loss: 0.24465592724235752\n",
            "epoch:200,batch: 300, train_loss: 0.23144926600158214, valid_loss: 0.24342163256074809\n",
            "epoch:201,batch: 100, train_loss: 0.23520616330206395, valid_loss: 0.24643303746286827\n",
            "epoch:201,batch: 200, train_loss: 0.23564317792654038, valid_loss: 0.24555727358483062\n",
            "epoch:201,batch: 300, train_loss: 0.23445254057645798, valid_loss: 0.25541187360694134\n",
            "epoch:202,batch: 100, train_loss: 0.2295190766453743, valid_loss: 0.25797538949718957\n",
            "epoch:202,batch: 200, train_loss: 0.23819239486008884, valid_loss: 0.25201990540269054\n",
            "epoch:202,batch: 300, train_loss: 0.23150101847946644, valid_loss: 0.2463785571670985\n",
            "epoch:203,batch: 100, train_loss: 0.22602056711912155, valid_loss: 0.2492616622696949\n",
            "epoch:203,batch: 200, train_loss: 0.21935804709792137, valid_loss: 0.24551826772056048\n",
            "epoch:203,batch: 300, train_loss: 0.2261806207150221, valid_loss: 0.24687965374581422\n",
            "epoch:204,batch: 100, train_loss: 0.22855733759701252, valid_loss: 0.24773144467344768\n",
            "epoch:204,batch: 200, train_loss: 0.23279980652034282, valid_loss: 0.24557860827521433\n",
            "epoch:204,batch: 300, train_loss: 0.23000740505754946, valid_loss: 0.24599507479350777\n",
            "epoch:205,batch: 100, train_loss: 0.23653834648430347, valid_loss: 0.24588489843697486\n",
            "epoch:205,batch: 200, train_loss: 0.22441185951232911, valid_loss: 0.24794606572087807\n",
            "epoch:205,batch: 300, train_loss: 0.2182623053714633, valid_loss: 0.24744494306513026\n",
            "epoch:206,batch: 100, train_loss: 0.22136063545942306, valid_loss: 0.2488704624621174\n",
            "epoch:206,batch: 200, train_loss: 0.2169732430577278, valid_loss: 0.2475443942637383\n",
            "epoch:206,batch: 300, train_loss: 0.2348534519225359, valid_loss: 0.25244276027513457\n",
            "epoch:207,batch: 100, train_loss: 0.23475581526756287, valid_loss: 0.25050967993049683\n",
            "epoch:207,batch: 200, train_loss: 0.22197306614369153, valid_loss: 0.2473793737118757\n",
            "epoch:207,batch: 300, train_loss: 0.21935415014624596, valid_loss: 0.2543581883552708\n",
            "epoch:208,batch: 100, train_loss: 0.2290210347622633, valid_loss: 0.24683191966784152\n",
            "epoch:208,batch: 200, train_loss: 0.2305097459256649, valid_loss: 0.2466560411679594\n",
            "epoch:208,batch: 300, train_loss: 0.2236695320904255, valid_loss: 0.24410677823839308\n",
            "epoch:209,batch: 100, train_loss: 0.2316789738088846, valid_loss: 0.24775382140769234\n",
            "epoch:209,batch: 200, train_loss: 0.24191205881536007, valid_loss: 0.2488884206248235\n",
            "epoch:209,batch: 300, train_loss: 0.22595418497920036, valid_loss: 0.24851624120639848\n",
            "epoch:210,batch: 100, train_loss: 0.22110320284962653, valid_loss: 0.24286300292875193\n",
            "epoch:210,batch: 200, train_loss: 0.23545932359993457, valid_loss: 0.25518006124073944\n",
            "epoch:210,batch: 300, train_loss: 0.23464078992605208, valid_loss: 0.25417064997968797\n",
            "epoch:211,batch: 100, train_loss: 0.21488867066800593, valid_loss: 0.24749507094863094\n",
            "epoch:211,batch: 200, train_loss: 0.22617616288363934, valid_loss: 0.24777666800007037\n",
            "epoch:211,batch: 300, train_loss: 0.22349427312612533, valid_loss: 0.24983527275580394\n",
            "epoch:212,batch: 100, train_loss: 0.21225426249206067, valid_loss: 0.25888163657693924\n",
            "epoch:212,batch: 200, train_loss: 0.23061783257871865, valid_loss: 0.24653650386424003\n",
            "epoch:212,batch: 300, train_loss: 0.22514166712760925, valid_loss: 0.2517163452656963\n",
            "epoch:213,batch: 100, train_loss: 0.21474417813122273, valid_loss: 0.2489215144062344\n",
            "epoch:213,batch: 200, train_loss: 0.2287324795126915, valid_loss: 0.2558929027069973\n",
            "epoch:213,batch: 300, train_loss: 0.22781762212514878, valid_loss: 0.2669348303653017\n",
            "epoch:214,batch: 100, train_loss: 0.22535127110779285, valid_loss: 0.2534135705496691\n",
            "epoch:214,batch: 200, train_loss: 0.2222205790877342, valid_loss: 0.24897105190195615\n",
            "epoch:214,batch: 300, train_loss: 0.22513835497200488, valid_loss: 0.249618716255019\n",
            "epoch:215,batch: 100, train_loss: 0.2186056936532259, valid_loss: 0.24558776598187942\n",
            "epoch:215,batch: 200, train_loss: 0.2197045822441578, valid_loss: 0.246428903050815\n",
            "epoch:215,batch: 300, train_loss: 0.23384664952754974, valid_loss: 0.24393749840651888\n",
            "epoch:216,batch: 100, train_loss: 0.21937303058803082, valid_loss: 0.256525649017171\n",
            "epoch:216,batch: 200, train_loss: 0.22250677220523357, valid_loss: 0.25433784855317465\n",
            "epoch:216,batch: 300, train_loss: 0.2268872693181038, valid_loss: 0.24848366594767268\n",
            "epoch:217,batch: 100, train_loss: 0.23068618949502706, valid_loss: 0.2537731499988822\n",
            "epoch:217,batch: 200, train_loss: 0.2307478429377079, valid_loss: 0.2443829002448275\n",
            "epoch:217,batch: 300, train_loss: 0.21244059361517428, valid_loss: 0.24515239111607587\n",
            "epoch:218,batch: 100, train_loss: 0.245961068123579, valid_loss: 0.24682362471954733\n",
            "epoch:218,batch: 200, train_loss: 0.22054877668619155, valid_loss: 0.24531886726617813\n",
            "epoch:218,batch: 300, train_loss: 0.2226773376762867, valid_loss: 0.24620706984136678\n",
            "epoch:219,batch: 100, train_loss: 0.21307336799800397, valid_loss: 0.24993927710795705\n",
            "epoch:219,batch: 200, train_loss: 0.24288842529058458, valid_loss: 0.24604347904648963\n",
            "epoch:219,batch: 300, train_loss: 0.2103680132329464, valid_loss: 0.24586482034831109\n",
            "epoch:220,batch: 100, train_loss: 0.24000033058226108, valid_loss: 0.24493619780751724\n",
            "epoch:220,batch: 200, train_loss: 0.21924760047346353, valid_loss: 0.2473760348143457\n",
            "epoch:220,batch: 300, train_loss: 0.22967373460531235, valid_loss: 0.2421879534479938\n",
            "epoch:221,batch: 100, train_loss: 0.22564079500734807, valid_loss: 0.24512572171567362\n",
            "epoch:221,batch: 200, train_loss: 0.21625834316015244, valid_loss: 0.2440725161304957\n",
            "epoch:221,batch: 300, train_loss: 0.2267140506953001, valid_loss: 0.2440090307706519\n",
            "epoch:222,batch: 100, train_loss: 0.23019972674548625, valid_loss: 0.2534875368958787\n",
            "epoch:222,batch: 200, train_loss: 0.2262883546948433, valid_loss: 0.25395145738803887\n",
            "epoch:222,batch: 300, train_loss: 0.22138047084212303, valid_loss: 0.24585129961937288\n",
            "epoch:223,batch: 100, train_loss: 0.2191497441008687, valid_loss: 0.24502963893398455\n",
            "epoch:223,batch: 200, train_loss: 0.2227969154715538, valid_loss: 0.24452482079026067\n",
            "epoch:223,batch: 300, train_loss: 0.23596009455621242, valid_loss: 0.24792029953832867\n",
            "epoch:224,batch: 100, train_loss: 0.2258141505718231, valid_loss: 0.2427459646818004\n",
            "epoch:224,batch: 200, train_loss: 0.21332121312618255, valid_loss: 0.2468803373130062\n",
            "epoch:224,batch: 300, train_loss: 0.22137416310608388, valid_loss: 0.249098255381554\n",
            "epoch:225,batch: 100, train_loss: 0.231482093334198, valid_loss: 0.24425062971024575\n",
            "epoch:225,batch: 200, train_loss: 0.2215038288757205, valid_loss: 0.24925852793304226\n",
            "epoch:225,batch: 300, train_loss: 0.220281802713871, valid_loss: 0.24604419175582595\n",
            "epoch:226,batch: 100, train_loss: 0.21513498835265638, valid_loss: 0.25916957421393333\n",
            "epoch:226,batch: 200, train_loss: 0.22255674630403519, valid_loss: 0.24963489221998408\n",
            "epoch:226,batch: 300, train_loss: 0.23855161361396313, valid_loss: 0.2446726587942884\n",
            "epoch:227,batch: 100, train_loss: 0.21896251745522022, valid_loss: 0.24646080341897433\n",
            "epoch:227,batch: 200, train_loss: 0.21843705840408803, valid_loss: 0.24900637010607538\n",
            "epoch:227,batch: 300, train_loss: 0.2141476000100374, valid_loss: 0.24423920033098775\n",
            "epoch:228,batch: 100, train_loss: 0.22887477599084377, valid_loss: 0.24420013156118273\n",
            "epoch:228,batch: 200, train_loss: 0.22238089017570017, valid_loss: 0.24416724667896197\n",
            "epoch:228,batch: 300, train_loss: 0.2269544992595911, valid_loss: 0.27344248990846587\n",
            "epoch:229,batch: 100, train_loss: 0.23198594018816948, valid_loss: 0.24417538818302034\n",
            "epoch:229,batch: 200, train_loss: 0.21731161445379257, valid_loss: 0.24404600102312957\n",
            "epoch:229,batch: 300, train_loss: 0.2209798275306821, valid_loss: 0.24678604125599318\n",
            "epoch:230,batch: 100, train_loss: 0.22556140802800656, valid_loss: 0.2487532320090487\n",
            "epoch:230,batch: 200, train_loss: 0.2132887042313814, valid_loss: 0.25185564971422847\n",
            "epoch:230,batch: 300, train_loss: 0.21842544890940188, valid_loss: 0.24625646780374683\n",
            "epoch:231,batch: 100, train_loss: 0.2298158370703459, valid_loss: 0.2440920855991448\n",
            "epoch:231,batch: 200, train_loss: 0.22703524019569157, valid_loss: 0.25839609954553316\n",
            "epoch:231,batch: 300, train_loss: 0.21835161179304122, valid_loss: 0.24867041144944443\n",
            "epoch:232,batch: 100, train_loss: 0.21999062810093165, valid_loss: 0.24779563236840163\n",
            "epoch:232,batch: 200, train_loss: 0.2271440002322197, valid_loss: 0.24681086055462872\n",
            "epoch:232,batch: 300, train_loss: 0.2245631232857704, valid_loss: 0.2441201928672911\n",
            "epoch:233,batch: 100, train_loss: 0.22899812653660775, valid_loss: 0.24837565987924987\n",
            "epoch:233,batch: 200, train_loss: 0.21573438510298729, valid_loss: 0.24315929893828644\n",
            "epoch:233,batch: 300, train_loss: 0.22010457362979652, valid_loss: 0.24785688302562206\n",
            "epoch:234,batch: 100, train_loss: 0.21539632931351663, valid_loss: 0.243853001277658\n",
            "epoch:234,batch: 200, train_loss: 0.22577079486101867, valid_loss: 0.24531390800883499\n",
            "epoch:234,batch: 300, train_loss: 0.23129251517355442, valid_loss: 0.24350357508357567\n",
            "epoch:235,batch: 100, train_loss: 0.22521375864744186, valid_loss: 0.24820371311676653\n",
            "epoch:235,batch: 200, train_loss: 0.21717406783252954, valid_loss: 0.24488454529001744\n",
            "epoch:235,batch: 300, train_loss: 0.22832793913781643, valid_loss: 0.2739755042557475\n",
            "epoch:236,batch: 100, train_loss: 0.22756791837513446, valid_loss: 0.24479873619879347\n",
            "epoch:236,batch: 200, train_loss: 0.23945654764771462, valid_loss: 0.24368642742120766\n",
            "epoch:236,batch: 300, train_loss: 0.21943996384739875, valid_loss: 0.24504146630628199\n",
            "epoch:237,batch: 100, train_loss: 0.22899561785161496, valid_loss: 0.2507601477488687\n",
            "epoch:237,batch: 200, train_loss: 0.2218740024417639, valid_loss: 0.2500799233211747\n",
            "epoch:237,batch: 300, train_loss: 0.20805798411369325, valid_loss: 0.2499773136422604\n",
            "epoch:238,batch: 100, train_loss: 0.22469951778650285, valid_loss: 0.24660536750585219\n",
            "epoch:238,batch: 200, train_loss: 0.23225764147937297, valid_loss: 0.264116861397707\n",
            "epoch:238,batch: 300, train_loss: 0.21540896467864512, valid_loss: 0.26565447110164014\n",
            "epoch:239,batch: 100, train_loss: 0.21911338333040475, valid_loss: 0.25446093761468236\n",
            "epoch:239,batch: 200, train_loss: 0.22546629633754492, valid_loss: 0.2586273995192745\n",
            "epoch:239,batch: 300, train_loss: 0.22289986323565245, valid_loss: 0.243888871201986\n",
            "epoch:240,batch: 100, train_loss: 0.21307393580675124, valid_loss: 0.2414389861912667\n",
            "epoch:240,batch: 200, train_loss: 0.20707764990627767, valid_loss: 0.2439318226673935\n",
            "epoch:240,batch: 300, train_loss: 0.24365822613239288, valid_loss: 0.25013491964038415\n",
            "epoch:241,batch: 100, train_loss: 0.2438584416359663, valid_loss: 0.24391319150034385\n",
            "epoch:241,batch: 200, train_loss: 0.21228628493845464, valid_loss: 0.24549394940273672\n",
            "epoch:241,batch: 300, train_loss: 0.2127053478360176, valid_loss: 0.24781457947779306\n",
            "epoch:242,batch: 100, train_loss: 0.22283684633672238, valid_loss: 0.24266422577673877\n",
            "epoch:242,batch: 200, train_loss: 0.22453143924474717, valid_loss: 0.26111123607128484\n",
            "epoch:242,batch: 300, train_loss: 0.20765627779066562, valid_loss: 0.24628322156546992\n",
            "epoch:243,batch: 100, train_loss: 0.2229158328473568, valid_loss: 0.24979291280990915\n",
            "epoch:243,batch: 200, train_loss: 0.2142129523307085, valid_loss: 0.24291670105502575\n",
            "epoch:243,batch: 300, train_loss: 0.21518352337181568, valid_loss: 0.24270073937464365\n",
            "epoch:244,batch: 100, train_loss: 0.22606031127274037, valid_loss: 0.24317597484664072\n",
            "epoch:244,batch: 200, train_loss: 0.20766294181346892, valid_loss: 0.24946535350401192\n",
            "epoch:244,batch: 300, train_loss: 0.2376742734387517, valid_loss: 0.24612086615230463\n",
            "epoch:245,batch: 100, train_loss: 0.2139389231801033, valid_loss: 0.2506607110741772\n",
            "epoch:245,batch: 200, train_loss: 0.21082080103456974, valid_loss: 0.24286330774237838\n",
            "epoch:245,batch: 300, train_loss: 0.23488230273127556, valid_loss: 0.24536934853354586\n",
            "epoch:246,batch: 100, train_loss: 0.22067586064338685, valid_loss: 0.2542638310903235\n",
            "epoch:246,batch: 200, train_loss: 0.21717027813196182, valid_loss: 0.24497293077314955\n",
            "epoch:246,batch: 300, train_loss: 0.23059403993189334, valid_loss: 0.2543688482498821\n",
            "epoch:247,batch: 100, train_loss: 0.22471645925194025, valid_loss: 0.24282052207596694\n",
            "epoch:247,batch: 200, train_loss: 0.21184078425168992, valid_loss: 0.24327054478322405\n",
            "epoch:247,batch: 300, train_loss: 0.22474212657660245, valid_loss: 0.25194601639162134\n",
            "epoch:248,batch: 100, train_loss: 0.22108617901802063, valid_loss: 0.2465995909481109\n",
            "epoch:248,batch: 200, train_loss: 0.21565957084298135, valid_loss: 0.24235041538559937\n",
            "epoch:248,batch: 300, train_loss: 0.21702937297523023, valid_loss: 0.24707546971644026\n",
            "epoch:249,batch: 100, train_loss: 0.20850533246994019, valid_loss: 0.25033142313927037\n",
            "epoch:249,batch: 200, train_loss: 0.21346719846129417, valid_loss: 0.24212364079077034\n",
            "epoch:249,batch: 300, train_loss: 0.22680447563529016, valid_loss: 0.2482786703901955\n",
            "epoch:250,batch: 100, train_loss: 0.23179276317358016, valid_loss: 0.25132288689477533\n",
            "epoch:250,batch: 200, train_loss: 0.21994929410517217, valid_loss: 0.24281136672708054\n",
            "epoch:250,batch: 300, train_loss: 0.2142713624984026, valid_loss: 0.2443138399644743\n",
            "epoch:251,batch: 100, train_loss: 0.2300917062163353, valid_loss: 0.24529487950892387\n",
            "epoch:251,batch: 200, train_loss: 0.20626872219145298, valid_loss: 0.25449236793608604\n",
            "epoch:251,batch: 300, train_loss: 0.22906304128468036, valid_loss: 0.24742137197452255\n",
            "epoch:252,batch: 100, train_loss: 0.2324389760941267, valid_loss: 0.24571967530476896\n",
            "epoch:252,batch: 200, train_loss: 0.21938841059803962, valid_loss: 0.2439381855575344\n",
            "epoch:252,batch: 300, train_loss: 0.2004613886028528, valid_loss: 0.24510110433720336\n",
            "epoch:253,batch: 100, train_loss: 0.2315181229636073, valid_loss: 0.2709703507679927\n",
            "epoch:253,batch: 200, train_loss: 0.21323315910995005, valid_loss: 0.24282884626071663\n",
            "epoch:253,batch: 300, train_loss: 0.21640010394155978, valid_loss: 0.24778292629914947\n",
            "epoch:254,batch: 100, train_loss: 0.21892611660063266, valid_loss: 0.2472139764812928\n",
            "epoch:254,batch: 200, train_loss: 0.21753057919442653, valid_loss: 0.24944061069171639\n",
            "epoch:254,batch: 300, train_loss: 0.22176216274499894, valid_loss: 0.24150235045559798\n",
            "epoch:255,batch: 100, train_loss: 0.23490209463983774, valid_loss: 0.24368395109342622\n",
            "epoch:255,batch: 200, train_loss: 0.2154395519569516, valid_loss: 0.24646162911306455\n",
            "epoch:255,batch: 300, train_loss: 0.22028662107884883, valid_loss: 0.24381224490419218\n",
            "epoch:256,batch: 100, train_loss: 0.21172261841595172, valid_loss: 0.24410066088732285\n",
            "epoch:256,batch: 200, train_loss: 0.23318394139409065, valid_loss: 0.24147137377081038\n",
            "epoch:256,batch: 300, train_loss: 0.21210162941366434, valid_loss: 0.24395302560510515\n",
            "epoch:257,batch: 100, train_loss: 0.21991647601127626, valid_loss: 0.24366360440661636\n",
            "epoch:257,batch: 200, train_loss: 0.2177918790653348, valid_loss: 0.24667490518923046\n",
            "epoch:257,batch: 300, train_loss: 0.22024544186890124, valid_loss: 0.25351331854545617\n",
            "epoch:258,batch: 100, train_loss: 0.21217249616980552, valid_loss: 0.2413658775106261\n",
            "epoch:258,batch: 200, train_loss: 0.21921847641468048, valid_loss: 0.28888542007041884\n",
            "epoch:258,batch: 300, train_loss: 0.21880302503705024, valid_loss: 0.24451396887815452\n",
            "epoch:259,batch: 100, train_loss: 0.20226788863539696, valid_loss: 0.24403370218940929\n",
            "epoch:259,batch: 200, train_loss: 0.2354559399187565, valid_loss: 0.25372922816608523\n",
            "epoch:259,batch: 300, train_loss: 0.22361382387578488, valid_loss: 0.25178928531800643\n",
            "epoch:260,batch: 100, train_loss: 0.21955481626093387, valid_loss: 0.2463561934006365\n",
            "epoch:260,batch: 200, train_loss: 0.21532442003488542, valid_loss: 0.24608578117965144\n",
            "epoch:260,batch: 300, train_loss: 0.21900059506297112, valid_loss: 0.24321412210223042\n",
            "epoch:261,batch: 100, train_loss: 0.228560486510396, valid_loss: 0.25111942719432373\n",
            "epoch:261,batch: 200, train_loss: 0.20611190743744373, valid_loss: 0.25016840990585615\n",
            "epoch:261,batch: 300, train_loss: 0.22881929703056814, valid_loss: 0.2464962908172909\n",
            "epoch:262,batch: 100, train_loss: 0.22166887696832419, valid_loss: 0.2433777328910707\n",
            "epoch:262,batch: 200, train_loss: 0.19269633688032628, valid_loss: 0.24882195540998556\n",
            "epoch:262,batch: 300, train_loss: 0.22630756065249444, valid_loss: 0.2465343751673457\n",
            "epoch:263,batch: 100, train_loss: 0.21346394311636685, valid_loss: 0.2466447445787961\n",
            "epoch:263,batch: 200, train_loss: 0.22294144943356514, valid_loss: 0.24630606900664825\n",
            "epoch:263,batch: 300, train_loss: 0.21174416013062, valid_loss: 0.2506332773762413\n",
            "epoch:264,batch: 100, train_loss: 0.21829700715839861, valid_loss: 0.2449986279199395\n",
            "epoch:264,batch: 200, train_loss: 0.2215824305266142, valid_loss: 0.25166007584031624\n",
            "epoch:264,batch: 300, train_loss: 0.21479339346289636, valid_loss: 0.2688710184036931\n",
            "epoch:265,batch: 100, train_loss: 0.21663234755396843, valid_loss: 0.242368330495267\n",
            "epoch:265,batch: 200, train_loss: 0.2230913907289505, valid_loss: 0.24329301205616963\n",
            "epoch:265,batch: 300, train_loss: 0.2134451949968934, valid_loss: 0.242188461124897\n",
            "epoch:266,batch: 100, train_loss: 0.2205060275644064, valid_loss: 0.24343480537586573\n",
            "epoch:266,batch: 200, train_loss: 0.21418246656656265, valid_loss: 0.24824896120006526\n",
            "epoch:266,batch: 300, train_loss: 0.21617226846516133, valid_loss: 0.24469452275882794\n",
            "epoch:267,batch: 100, train_loss: 0.2192522217705846, valid_loss: 0.24447022076649003\n",
            "epoch:267,batch: 200, train_loss: 0.1972340427339077, valid_loss: 0.2435779674803909\n",
            "epoch:267,batch: 300, train_loss: 0.22530452154576777, valid_loss: 0.26328418988593016\n",
            "epoch:268,batch: 100, train_loss: 0.21510607428848744, valid_loss: 0.2828710775209379\n",
            "epoch:268,batch: 200, train_loss: 0.22382329687476157, valid_loss: 0.24256798247747782\n",
            "epoch:268,batch: 300, train_loss: 0.213732454366982, valid_loss: 0.2437773548726794\n",
            "epoch:269,batch: 100, train_loss: 0.21689937710762025, valid_loss: 0.24172273108476325\n",
            "epoch:269,batch: 200, train_loss: 0.2174770388752222, valid_loss: 0.24252319882942153\n",
            "epoch:269,batch: 300, train_loss: 0.20423564597964286, valid_loss: 0.25023330106765407\n",
            "epoch:270,batch: 100, train_loss: 0.20876212246716022, valid_loss: 0.2809490635047985\n",
            "epoch:270,batch: 200, train_loss: 0.2181242496520281, valid_loss: 0.2441362153408648\n",
            "epoch:270,batch: 300, train_loss: 0.23165357306599618, valid_loss: 0.2470181207113628\n",
            "epoch:271,batch: 100, train_loss: 0.2169801203161478, valid_loss: 0.24393792063752306\n",
            "epoch:271,batch: 200, train_loss: 0.20976403404027225, valid_loss: 0.24505232444292382\n",
            "epoch:271,batch: 300, train_loss: 0.20861789632588626, valid_loss: 0.25416952220699457\n",
            "epoch:272,batch: 100, train_loss: 0.2069330457597971, valid_loss: 0.24379930371725106\n",
            "epoch:272,batch: 200, train_loss: 0.20070820324122907, valid_loss: 0.24394532377961314\n",
            "epoch:272,batch: 300, train_loss: 0.233434848934412, valid_loss: 0.24797925437930263\n",
            "epoch:273,batch: 100, train_loss: 0.22986406877636908, valid_loss: 0.2438272120831888\n",
            "epoch:273,batch: 200, train_loss: 0.21937994942069053, valid_loss: 0.26508337127257\n",
            "epoch:273,batch: 300, train_loss: 0.21775537967681885, valid_loss: 0.25019064639942556\n",
            "epoch:274,batch: 100, train_loss: 0.21943803645670415, valid_loss: 0.24722926818494556\n",
            "epoch:274,batch: 200, train_loss: 0.2030224995315075, valid_loss: 0.2470762381165088\n",
            "epoch:274,batch: 300, train_loss: 0.2220312147587538, valid_loss: 0.2426341000614287\n",
            "epoch:275,batch: 100, train_loss: 0.21919220518320798, valid_loss: 0.24653835057080548\n",
            "epoch:275,batch: 200, train_loss: 0.2194593533873558, valid_loss: 0.2443091458157648\n",
            "epoch:275,batch: 300, train_loss: 0.21392176132649182, valid_loss: 0.24233024397605582\n",
            "epoch:276,batch: 100, train_loss: 0.21654970020055772, valid_loss: 0.2419821940645387\n",
            "epoch:276,batch: 200, train_loss: 0.22662684455513954, valid_loss: 0.24709139990655682\n",
            "epoch:276,batch: 300, train_loss: 0.20362818956375123, valid_loss: 0.24150003172174284\n",
            "epoch:277,batch: 100, train_loss: 0.20734248585999013, valid_loss: 0.25756438180238383\n",
            "epoch:277,batch: 200, train_loss: 0.21493628449738025, valid_loss: 0.2444520757734021\n",
            "epoch:277,batch: 300, train_loss: 0.22277532696723937, valid_loss: 0.2487960750166374\n",
            "epoch:278,batch: 100, train_loss: 0.2128687649220228, valid_loss: 0.24622944591543341\n",
            "epoch:278,batch: 200, train_loss: 0.21121978893876076, valid_loss: 0.2438797076480298\n",
            "epoch:278,batch: 300, train_loss: 0.20511536657810212, valid_loss: 0.2453956180546857\n",
            "epoch:279,batch: 100, train_loss: 0.22708293229341506, valid_loss: 0.25305081253187567\n",
            "epoch:279,batch: 200, train_loss: 0.2040185585618019, valid_loss: 0.24223515391349792\n",
            "epoch:279,batch: 300, train_loss: 0.21619117133319377, valid_loss: 0.2433732060503356\n",
            "epoch:280,batch: 100, train_loss: 0.20270923927426338, valid_loss: 0.24595257111742527\n",
            "epoch:280,batch: 200, train_loss: 0.20690238520503043, valid_loss: 0.24339884340385848\n",
            "epoch:280,batch: 300, train_loss: 0.23243436343967916, valid_loss: 0.25422191629304164\n",
            "epoch:281,batch: 100, train_loss: 0.2205492402240634, valid_loss: 0.2439032412028011\n",
            "epoch:281,batch: 200, train_loss: 0.21608341500163078, valid_loss: 0.24545659907633746\n",
            "epoch:281,batch: 300, train_loss: 0.21171515606343747, valid_loss: 0.26112121371906016\n",
            "epoch:282,batch: 100, train_loss: 0.2379704661667347, valid_loss: 0.24942078809194926\n",
            "epoch:282,batch: 200, train_loss: 0.21842080824077129, valid_loss: 0.24527294990382617\n",
            "epoch:282,batch: 300, train_loss: 0.20335448421537877, valid_loss: 0.24465134654995763\n",
            "epoch:283,batch: 100, train_loss: 0.21327809851616622, valid_loss: 0.24278185007315647\n",
            "epoch:283,batch: 200, train_loss: 0.21653912335634232, valid_loss: 0.24380600433560867\n",
            "epoch:283,batch: 300, train_loss: 0.21057701002806425, valid_loss: 0.24489826303494128\n",
            "epoch:284,batch: 100, train_loss: 0.20916488707065584, valid_loss: 0.2551845001834857\n",
            "epoch:284,batch: 200, train_loss: 0.21556528612971307, valid_loss: 0.2430715663712236\n",
            "epoch:284,batch: 300, train_loss: 0.21850480653345586, valid_loss: 0.24510555978440032\n",
            "epoch:285,batch: 100, train_loss: 0.220932155624032, valid_loss: 0.24944859299855895\n",
            "epoch:285,batch: 200, train_loss: 0.2140861766040325, valid_loss: 0.24984766302418104\n",
            "epoch:285,batch: 300, train_loss: 0.21916919227689505, valid_loss: 0.2489626221830332\n",
            "epoch:286,batch: 100, train_loss: 0.21875041238963605, valid_loss: 0.2426492221936395\n",
            "epoch:286,batch: 200, train_loss: 0.21272089190781115, valid_loss: 0.2528308415903321\n",
            "epoch:286,batch: 300, train_loss: 0.22110671795904635, valid_loss: 0.24470962716054312\n",
            "epoch:287,batch: 100, train_loss: 0.22435442075133324, valid_loss: 0.24435358107844485\n",
            "epoch:287,batch: 200, train_loss: 0.2102994740009308, valid_loss: 0.24220756997790518\n",
            "epoch:287,batch: 300, train_loss: 0.21373471844941377, valid_loss: 0.2476114038425156\n",
            "epoch:288,batch: 100, train_loss: 0.22657042413949965, valid_loss: 0.257575623030904\n",
            "epoch:288,batch: 200, train_loss: 0.21488511271774768, valid_loss: 0.24301338068480732\n",
            "epoch:288,batch: 300, train_loss: 0.2127378036081791, valid_loss: 0.2456248122491414\n",
            "epoch:289,batch: 100, train_loss: 0.20573365591466428, valid_loss: 0.2501860535408877\n",
            "epoch:289,batch: 200, train_loss: 0.22970818161964415, valid_loss: 0.24994224111867858\n",
            "epoch:289,batch: 300, train_loss: 0.2202028124779463, valid_loss: 0.24760858893771714\n",
            "epoch:290,batch: 100, train_loss: 0.21392856240272523, valid_loss: 0.24320745147481748\n",
            "epoch:290,batch: 200, train_loss: 0.2087538992986083, valid_loss: 0.24604592882568324\n",
            "epoch:290,batch: 300, train_loss: 0.2123691364377737, valid_loss: 0.2451642657382579\n",
            "epoch:291,batch: 100, train_loss: 0.2096135676279664, valid_loss: 0.24129485112579563\n",
            "epoch:291,batch: 200, train_loss: 0.21619081251323224, valid_loss: 0.24078211556129817\n",
            "epoch:291,batch: 300, train_loss: 0.22421436555683613, valid_loss: 0.2754309629149075\n",
            "epoch:292,batch: 100, train_loss: 0.22682035088539124, valid_loss: 0.24292706320934657\n",
            "epoch:292,batch: 200, train_loss: 0.21568262323737145, valid_loss: 0.27249385303334345\n",
            "epoch:292,batch: 300, train_loss: 0.20419671583920718, valid_loss: 0.252631947586808\n",
            "epoch:293,batch: 100, train_loss: 0.21579795133322477, valid_loss: 0.24435126847481425\n",
            "epoch:293,batch: 200, train_loss: 0.21724520791321994, valid_loss: 0.2419927551210681\n",
            "epoch:293,batch: 300, train_loss: 0.22220465943217277, valid_loss: 0.24447650726460204\n",
            "epoch:294,batch: 100, train_loss: 0.2155258559435606, valid_loss: 0.24471267416507383\n",
            "epoch:294,batch: 200, train_loss: 0.2080603801086545, valid_loss: 0.24139068756676926\n",
            "epoch:294,batch: 300, train_loss: 0.20739113252609967, valid_loss: 0.24061168683103368\n",
            "epoch:295,batch: 100, train_loss: 0.21901634778827428, valid_loss: 0.24632588924863671\n",
            "epoch:295,batch: 200, train_loss: 0.21604917600750922, valid_loss: 0.2509832212442084\n",
            "epoch:295,batch: 300, train_loss: 0.2218959740176797, valid_loss: 0.25087674763760986\n",
            "epoch:296,batch: 100, train_loss: 0.21207540143281223, valid_loss: 0.24278267774778076\n",
            "epoch:296,batch: 200, train_loss: 0.22254689574241637, valid_loss: 0.24431004460099376\n",
            "epoch:296,batch: 300, train_loss: 0.2142691758647561, valid_loss: 0.2439631260459936\n",
            "epoch:297,batch: 100, train_loss: 0.2210762605816126, valid_loss: 0.25856360566766956\n",
            "epoch:297,batch: 200, train_loss: 0.21466986402869226, valid_loss: 0.24213298259279395\n",
            "epoch:297,batch: 300, train_loss: 0.21072635538876056, valid_loss: 0.24249144011660467\n",
            "epoch:298,batch: 100, train_loss: 0.2126842188835144, valid_loss: 0.241296837224236\n",
            "epoch:298,batch: 200, train_loss: 0.21243647836148738, valid_loss: 0.24245956370347663\n",
            "epoch:298,batch: 300, train_loss: 0.21560948353260756, valid_loss: 0.2464844100648844\n",
            "epoch:299,batch: 100, train_loss: 0.20872038885951041, valid_loss: 0.2567400501310071\n",
            "epoch:299,batch: 200, train_loss: 0.22256496481597424, valid_loss: 0.2502772043211551\n",
            "epoch:299,batch: 300, train_loss: 0.20987906217575072, valid_loss: 0.24197259526464004\n"
          ]
        }
      ],
      "source": [
        "lstm_model = lstm()\n",
        "valid_losses = []\n",
        "lstm_model.fit(train_loader=train_loader)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 88,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 296
        },
        "id": "c4pMer0t_ZcG",
        "outputId": "1b1d7b02-b220-4355-b1d4-475c09eeafe0"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'valid loss')"
            ]
          },
          "metadata": {},
          "execution_count": 88
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEGCAYAAABo25JHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dd3hc1Z3/8fd3Rr1bltw7yMQGF2wH02wgQAImlJRNIEBCQkJgIZDd/DZLFjaF7JJs2u6SJZSEJNRgSiCGODHBGINjbCz3bstFluSiXqyumfP7Y0ZCZWTLxuORfD+v59GjmTtXc7/XV56Pzjn3nmvOOURExLt8sS5ARERiS0EgIuJxCgIREY9TEIiIeJyCQETE4+JiXcCxysnJcePGjYt1GSIiA8rq1avLnXO5kV4bcEEwbtw48vPzY12GiMiAYmaFvb2mriEREY9TEIiIeJyCQETE4xQEIiIepyAQEfE4BYGIiMcpCEREPM4zQbBqbyU/f2M7rYFgrEsREelXPBMEawqr+OVbBbS0KQhERDrzTBD4fQZAQDfiERHpwjNB4LNQEASDCgIRkc48EwQdLQIFgYhIF54JAp+6hkREIvJMEPg7uoZiXIiISD/jnSAI76laBCIiXXkmCDRYLCISmWeCQIPFIiKReS8I1DUkItKFZ4JAXUMiIpF5JgjUIhARicwzQdDeItAYgYhIV54JgvYWga4jEBHpykNBEPquriERka48EwTqGhIRiSxqQWBmvzWzUjPb1MvrZmYPmVmBmW0wsxnRqgU6dQ2pRSAi0kU0WwS/B644wutXAnnhr9uAR6JYS8dcQ2oRiIh0FbUgcM69A1QeYZVrgadcyAogy8yGR6sen0/XEYiIRBLLMYKRQFGn58XhZVGh6whERCIbEIPFZnabmeWbWX5ZWdlxvYcGi0VEIotlEJQAozs9HxVe1oNz7nHn3Czn3Kzc3Nzj2pgGi0VEIotlECwAvhg+e+hcoMY5dyBaG/tgsDhaWxARGZjiovXGZvYH4GIgx8yKge8B8QDOuUeBhcA8oABoAL4crVoAfOHIU4tARKSrqAWBc+6Go7zugDujtf3u/DprSEQkogExWHwidHQNqUUgItKFZ4LApzuUiYhE5JkgaG8RaIxARKQr7wSBT2cNiYhE4pkg0BQTIiKReScIQjmgwWIRkW48EwSafVREJDLPBIFPU0yIiETkmSBQi0BEJDLPBIGuIxARicwzQaDZR0VEIvNOEGj2URGRiDwTBJp9VEQkMs8EgQaLRUQi804QaLBYRCQizwSBmWGmriERke48EwQQ6h5Si0BEpCtPBYHPZ5prSESkG08Fgd9Ms4+KiHTjrSDwma4jEBHpxlNB4NNgsYhID54KglCLQEEgItKZ94JALQIRkS48FQQ+DRaLiPTgqSBQ15CISE+eCgKfqWtIRKQ7TwWB36euIRGR7jwXBAHlgIhIF54KAp+hFoGISDeeCgINFouI9OSpINBgsYhIT54KAg0Wi4j05LkgUItARKQrTwWBTzemERHpwVNB4PeZZh8VEenGW0GgFoGISA+eCgKfD4K6MY2ISBeeCgINFouI9BTVIDCzK8xsu5kVmNm9EV4fY2ZLzGytmW0ws3nRrEeDxSIiPUUtCMzMDzwMXAlMBm4ws8ndVrsfeME5dzZwPfCraNUDGiwWEYkkmi2Cc4AC59xu51wL8Dxwbbd1HJARfpwJ7I9iPRosFhGJIJpBMBIo6vS8OLyss+8DN5lZMbAQ+EakNzKz28ws38zyy8rKjrsgn+YaEhHpIdaDxTcAv3fOjQLmAU+bWY+anHOPO+dmOedm5ebmHvfG/GaoZ0hEpKtoBkEJMLrT81HhZZ3dCrwA4Jx7D0gCcqJVkM4aEhHpKZpBsArIM7PxZpZAaDB4Qbd19gGXApjZJEJBcPx9P0fh06RzIiI9RC0InHNtwF3AImArobODNpvZA2Z2TXi1bwFfM7P1wB+AW5yL3p/sfkMtAhGRbuKi+ebOuYWEBoE7L/tup8dbgAuiWUNnGiwWEekp1oPFJ5Xf1DUkItKdt4JAg8UiIj14KghCXUOxrkJEpH/xVBD4TVNMiIh0560g0GCxiEgPxxQEZuYzs4yjr9k/+TRYLCLSw1GDwMyeM7MMM0sFNgFbzOxfol/aief36ToCEZHu+tIimOycqwWuA/4CjAdujmpVUaLrCEREeupLEMSbWTyhIFjgnGslNH30gKPBYhGRnvoSBI8Be4FU4B0zGwvURrOoaNFgsYhIT0cNAufcQ865kc65eS6kELjkJNR2wvnMCDqI4nRGIiIDTl8Gi+8JDxabmT1hZmuAj52E2k44v88AUKNAROQDfeka+kp4sPjjwCBCA8U/jmpVUdIeBOoeEhH5QF+CwMLf5wFPO+c2d1o2oPisvUWgIBARadeXIFhtZm8QCoJFZpYODMgZe/zhvVWLQETkA325H8GtwHRgt3OuwcwGA1+OblnRER9OgprGVlITo3orBhGRAaMvZw0FCd1v+H4z+xlwvnNuQ9Qri4I5eaEb37+6rvutk0VEvKsvZw39GLgH2BL+utvMHox2YdFw+pA0Zo/P5rmV+6iqb4l1OSIi/YId7Zx6M9sATA+3DDAzP7DWOTf1JNTXw6xZs1x+fv5x//zb20v56pP5JCf4mZuXy4isJHLSEjktN41LJw3BbECOg4uIHJGZrXbOzYr0Wl87yrOAyvDjzBNSVYxcfMYQXvvGhfz6nd2sKqxk8bZmmlpDY9/fv3oyt1wwPsYVioicXH0Jgh8Ba81sCaHTRucC90a1qiibNDyDX3x+OhC6yri+JcDdf1jLg3/ZxrwpwxmSkRTjCkVETp6+DBb/ATgX+CPwMnCec25+tAs7WcyMtMQ4br1wPC1tQXaV1ce6JBGRk6rXFoGZzei2qDj8fYSZjXDOrYleWSdfTloiABX1zTGuRETk5DpS19DPj/CaY4DON9SbnLQEAMrrFAQi4i29BoFzbkDOMHq8BqUk4DMoP6zTSkXEWzx18/oj8fmM7NREdQ2JiOcoCDrJSUugrE4tAhHxFgVBJzlpahGIiPccy1lDXZxqZw1BqEVQuE+nj4qIt/TlrKEkYBawntAFZVOBfOC86JZ28g1OS6RCg8Ui4jG9dg055y4Jnzl0AJjhnJvlnJsJnA2cktN35qQl0tASoKGlLdaliIicNH0ZIzjDObex/YlzbhMwKXolxU77tQRqFYiIl/QlCDaY2W/M7OLw16+BAXk/gqPJTg0HgaaoFhEP6cukc18G7iB0TwKAd4BHolZRDKUkhP45GlsCMa5EROTkOWoQOOeagP8Of53SkhP8ADS1KghExDuOdProC865z5nZRkJzC3URqxvTRFNyfCgIGtQiEBEPOVKLoL0r6JMno5D+oD0IGtUiEBEPOdKkcwfC3wtPXjmx1d41pCAQES/p9awhM6szs9oIX3VmVtuXNzezK8xsu5kVmFnEu5qZ2efMbIuZbTaz5453R06EjjECdQ2JiIccqUWQ/mHeOHyT+4eBywnd1GaVmS1wzm3ptE4e8B3gAudclZkN+TDb/LA0RiAiXtTXm9cT/pDuuJmvc27fUX7kHKDAObc7/PPPA9cCWzqt8zXgYedcVfg9S/taTzT4fUZCnE9dQyLiKUe9oMzMrjGzncAeYCmwF/hLH957JFDU6XlxeFlnE4GJZvZ3M1thZlf0UsNtZpZvZvllZWV92PTxS4736/RREfGUvlxZ/ENCN6/f4ZwbD1wKrDhB248D8oCLgRuAX5tZVveVnHOPh+c6mpWbm3uCNh1Zcrxfcw2JiKf0JQhanXMVgM/MfM65JYRmIz2aEmB0p+ej6DlZXTGwwDnX6pzbA+wgFAwxk5Lgp7E1GMsSREROqr4EQbWZpRGaWuJZM/tfoC+T9q8C8sxsvJklANcDC7qt8yqh1gBmlkOoq2h3H2uPiqR4v6aYEBFP6UsQXAs0AP8E/BXYBVx9tB9yzrUBdwGLgK3AC865zWb2gJldE15tEVBhZluAJcC/hFsfMZOc4KexVV1DIuIdfTlr6OvAfOdcCfDksby5c24hsLDbsu92euyAfw5/9QsaIxARr+lLiyAdeMPM3jWzu8xsaLSLiqVkjRGIiMccNQiccz9wzp0J3AkMB5aa2ZtRryxGkuP9NKpFICIe0pcWQbtS4CBQAcT0CuBoSo7364IyEfGUvlxQ9o9m9jawGBgMfO1UnIK6XXKCzhoSEW/py2DxaOCbzrl10S6mPwiNESgIRMQ7+nKHsu+cjEL6i+R4P60BR2sgSLz/WHrOREQGJn3SddM+A6nmGxIRr1AQdNNxcxqNE4iIRygIutHtKkXEaxQE3eh2lSLiNQqCbtqDoL5ZQSAi3qAg6CY9MXQi1eFmXV0sIt6gIOgmLSkcBE0KAhHxBgVBN+lJ8QDUNbXGuBIRkZNDQdBNerhFUKcWgYh4hIKgm7SEOMzUIhAR71AQdOPzGWkJcdSqRSAiHqEgiCA9KU5dQyLiGQqCCNKT4tU1JCKeoSCIQC0CEfESBUEE6Ulx1DWrRSAi3qAgiCDUNaQWgYh4g4IgAnUNiYiXKAgiaB8sds7FuhQRkahTEESQnhRHa8DR3BaMdSkiIlGnIIggIzzNRK1OIRURD1AQRPDBxHMaJxCRU5+CIAJNPCciXqIgiCArJdQiqG5oiXElIiLRpyCIYHBqIgAVhxUEInLqUxBEMDgtAYCK+uYYVyIiEn0KggjSEuNIiPOpRSAinqAgiMDMyElNoFxBICIeoCDoxeC0RHUNiYgnKAh6MTgtQV1DIuIJCoJeDE5NpOKwWgQicupTEPQiJy2B8voWTTwnIqc8BUEvBqcl0NIW5HCzri4WkVNbVIPAzK4ws+1mVmBm9x5hvc+YmTOzWdGs51joojIR8YqoBYGZ+YGHgSuBycANZjY5wnrpwD3AymjVcjzaLyor1ziBiJziotkiOAcocM7tds61AM8D10ZY74fAfwFNUazlmI0dnArArrLDMa5ERCS6ohkEI4GiTs+Lw8s6mNkMYLRz7s9HeiMzu83M8s0sv6ys7MRXGsHY7BTSEuPYVFJ7UrYnIhIrMRssNjMf8AvgW0db1zn3uHNulnNuVm5ubvSLA3w+Y/KIDDbtrzkp2xMRiZVoBkEJMLrT81HhZe3SgbOAt81sL3AusKA/DRhPGZnJ1gO1tAV0y0oROXVFMwhWAXlmNt7MEoDrgQXtLzrnapxzOc65cc65ccAK4BrnXH4UazomZ43MoKk1yK6y+liXIiISNVELAudcG3AXsAjYCrzgnNtsZg+Y2TXR2u6JNHl4JgDbDmqcQEROXXHRfHPn3EJgYbdl3+1l3YujWcvxGJ+TSpzP2H6wLtaliIhEja4sPoKEOB/jc1LZcUinkIrIqUtBcBQTh6Wz45BaBCJy6lIQHMUZQ9MpqmqgoUVzDonIqUlBcBQTh6bhHOxU95CInKIUBEcxY8wg4nzGgvX7Y12KiEhUKAiOYkhGEldPG8Hz7++jprE11uWIiJxwCoI++NqcCdS3BHgxv+joK4uIDDAKgj6YPCKDmWMH8dzKfbpjmYicchQEffSFc8awu7yeL/1uFfurG2NdjojICaMg6KOrp43g63MnsHJ3Bf/z5o5YlyMicsIoCPooIc7Hd+ZN4nOzRvPq2v2U1var++iIiBw3BcEx+uqc8QSd46eLtnOwpkljBiIy4CkIjtHYwal8dc4EXlxdzLk/WswzK/fFuiQRkQ9FQXAc7rk0jxtnj2Hs4BSefm+vWgUiMqApCI5DcoKf//zUFO646DR2HDrM6sKqjtfmr9pHYYVuZCMiA4eC4EO4etoIBqcm8MDrW2gLBNm8v4Z/fXkj97+6KdaliYj0mYLgQ0hNjOMH157JhuIaPv/4Cn7+Rui00nd3lndpJYiI9GcKgg/pqinDefBTUyipauStbaVcPnko2akJPLR45wl5/4aWNirrW07Ie4mIRBLVW1V6gZnxhdljuHb6COavKuLjZw7ltfUH+K+/buO9XRWcd9rgPr3P7/++B7/PuPm8cV2W3/2HtRRXNfLXb86NQvUiImoRnDCpiXF85cLxjBqUwhfPG8uwjCRu/M0KvvGHtWzZX0tNQyvBYOSzi4JBx0NvFfDI27u6LC8orePNraVsP1TH4eaBdWOcjcU17KtoiHUZItIHahFEQWpiHAvvmcMjbxfw/KoiXt+wH+dg2ugs/u+Gs1mzr4phGUnMnhBqLeworevo/impbmRkVjIATyzbC4BzsP1gLTPHZsdkf47HPz63molD0nnilo/GupST7r5XNtLYGuAXn5sOwNp9VZQfbuHyyUNjXJlIZAqCKMlOTeC+qyZz1yV5/G75Hg43tfHUe4XM+cmSjnWunjaCG2ePYX1RdceyN7ccYs2+KloDQRZvLeWiibks3VHGlv2hICira6a4qoGzxwyKxW71SWNLgKLKRppbg7EuJSZW7qmkuS3Q8fx/F+9k+8G6fhUE64qqOS03lfSk+FiXIv2AgiDKMlPi+eZlEwH4/EdH8/eCcsbmpLJqTyVPLt/La+E7n43OTqaqvpXvLdiM32cEwt1I9181iXVF1Ww5UEcw6Pjqk6vYvL+Wv35zDqcPSQegLRDkcHMbWSkJ/HnDAR5avJO5E3O498pJ+H3Wa21V9S0MSk044fu8pzx0HUVpXTNldc3kpicC8N0/bSLoHP9x3ZQTvs3OKutb+Pkb2/m3eZNITTy5v+LOOYqrGggEHcGgw+cz9lU0cKi2idZAkHh/7HtjG1sC/MOjy/nGx/K4+9K8WJfTLz2zohDnXI8xu1OVguAkyhuaTt7Q0If3JWcM4Y6LT+O9XRUs31XBR8dl88aWg2wsruF/rp/O4q2l1De3kTc0ncnDM/jzhv1s2V/D+uIa4nzGnc+u5fLJQ0lPiuOVtSUUVjTwg2vP5Lt/2kRaYjy/fncPM8cO4oqzhkesZcm2Ur7y5Cr+/arJzBw7iLe3l3HL+ePITPnwfyHuLv/g/s6b99dw8RlDcM7x2vr9NLQEuG/eZJIT/Px100F2lx/m9rmn4esUWFX1LRRVNTB1VNZxbf/NrYd4duU+5k7M5RNnDvvQ+3MsKutbaAq3hMoPNzM4LZHiqkaCDg7VNjFqUMpJrSeSkupGWgOO3WWn/n24nXN85pHlfGH2WD47c1Sff+7J5XsxQ0Eg0ZeeFM/HzxzGx8MfVleeNQyz0JlInT8EvzB7DE+vKKQtEOS66SO4bPJQfrpoO48s3UUg6BiTnUJqop9vv7SB0dnJvHT7+XzmkeU8unQ3mckJVNQ3k5oQh8Px/PtF3HvlR3jk7V04Bw+8vqVjOxX1zTxw7Vldatxf3ciwjKQuH9Sd1TS08n9LdjJ3Yi5z8nIB2F32wZXVm/fXMml4BiXVjVQ1hG71uaygnMsnD+Xu59fS0hbkQHUTP7zug+3e/+om/rzxAPde+RFuv+i0Y/533XagruN7b0HQ1Brg2y9t4O5L8zh9SFqP1xLjfJj13prqTXHVB/eqKKlupC3oaAmEgqGkqrHfBAFAUdWx31ejLbwvcf2gZdMXVQ2trNlXzdjBqX0OgrZAkL0V9ST4fTjnjuv3YKBREPQjvX3YXj1tBFdPG9Fl2SenjqClLUhja4DM5HhWF1bxytpi/vnyM8hOTeD2i07j/lc3ccOvV/R4v2UF5TS0BPjW5RNJTvCTlZLAe7sqeG7lPhL8Pv6y6SBTRmbSFgzy5tZSPjdrFHMn5jIkPYlxg1P4vyUFXHB6DqMHpXDrk6s4UNPE86uKmH/befx00TaWbC9jZFYycX7juZX7+NWSAloDoa4un8GizQf5yLB0WtpCHyrPvb+Pey7LIyctkUDQsWR7KQA//ss2po/O4twJfTsFt922g7UAbD1Q2+s6q/ZWsmD9foZnJvGdeZM6ltc0tHLxz5Zw96V5fPmC8ce0XegaBAdqmmhu+2CcZH/Nib+hUW1TKxnH2M9fEq5xX+Wxn9V1x7NriPMZj9w085h/NhaKwvu49ximfSmqCrWYWgMBahpbyUo58d2n/Y2CYABLiPOREBf6y2zm2EHMHPvBAPJN545l5thBVBxuITc9kfLDzRRWNDB1VCa/X76XEVnJfG3uBJLi/QBcNDGXHYfq+M2yPZw9Jov8wkrifD7m5OXwQn4xL+QXd9n2U+8VAjA0I5FHbpzBt1/awLyH3u143TnHTz4zjW/OX0dyQhz1h5tJjvdzzbQRvLSmGH/4r6xHb5rB7c+s4an3CrlqynBqGltpaAnw089O5ZdvFfBvf9zIwnvm8O7OchZtPsh/XHdWR80AK3ZX8ODCrQzNSOKnn51KZnI82w6GWgRbD0YOgtZAkPy9oSu/lxWUd3ntxdVFVDW08srakmMKgmdWFJIc76f8cHPHsv3VjV1O+y3p41/gBaWH2VBczafOHnnEv0ZX7q7ghl+v4Jmvzub803IirtMWCPLTN7Zz0+yxjM5O6agLoKyumcaWAMkJoX/PF/KLmDQsgymjMqltaiUpzt/x+9X+Xst2lhNwjqbWQJfjcCxqm1rZVFLTa83HyjnH29vLOGd8do8xoaKqUBAUHsOpzLtKP+gyK65qVBDIwDZpeEbH4zNI54LTQ49/9g/Teqybm57Ia9+4kKr6FrJS4js+gJxzvLymhDHZKdQ0trKxpIaLJuawdl81gaDjurNHMjQjiY8Mz+DJ5XuZNW4Qq/ZUMmPsIGZPGMw7376EoHN8/L/fYfSgFP5t3iTe3lHK/PwixmSn8IkzhzF5eAYPLd7ZcTW232d8/MxhDMtM4uYn3ueOZ1azrKCc1oCjuqGF04ekM21UJmWHm/nlWwXE+Ywt+2v5xh/WMjQjicr6UPgVVjSw9UAtYwen8PKaEi48PYeKw83c8rtVHR/Qm/fX8tdNB5g0PINhmUk8s6IQn8GG4hr2VzcyIisZ5xxPryhk0vAMZowZRFFlA8kJfoakJ2Jm1De38Z9/3kq837h88jAykuIIBB0l1Y1UN7Ti9xlpiXEs3VFGVkoC54zPZmJ4rKg1EKSyvoWhGUlUN7Tw+Du7+VX4epKJQ9M5a2Rmr8d3fn4RQQe/XFzQ8aHqnGPR5kMUVzXw+Y+OZvP+Wh5bupuWtiDf/eRknl25j1V7Kzveo7iqgbyh6eyraOBfX97A+JxUFt49h08+tAyfwW++NKvjpISdpYdpbA2dDbVqb2VHV+CxeujNnfxm2R6evvUcnIP7Xt3IvVdM4qqpkcezjubF1cV8+6UN/MsnzuDOS07v8lpRZSj0KutbqGlsJTP56K2nXZ3GTkqqGzuOQVFlA+WHmzl7zCBWF1ayeX8tX4zSGMKafVVMH5XVay/BiaYgkC66n0VkZl36VttPgex+TcP4nFS+f82ZQKjbql37WTIvfv08zIzMlHhevuN8lu+qYMrITMyMBz89hVV7KslOTWBHaR1jslPITI5nTl4un5kxipfXFHPehMGMz03luZX7WLK9rOOsqvTEOJ7/+rm8sfkQ/9tpWo9Pnz2Sx97ZzZX/+y7pSXHUNYU++NMS4zpCYEJOKrvL67n9mTUADM9M4kBNE/dfNYn/+PNW/vmFdUwbnUVxZSN/3niABL+P1ER/x1jHddNHkJYUx97yBhpbAzS2woL1JeQNSac1EGTL/lrqW9oYkZWEz4xVe6tYtbeKhDgfT3/lHGaOHcTXnspn+a4KfvG5afzfWwXsOFTHeRMG897uCn6/fC/TRmcxdWQmU0dlUlrXzBubD/L5j44hEHQs2nSQ7NQE3ttdwWvr93P1tBEs2nyI259ZDUBdUxut4T79NzYfYvrorI4JEZPj/TS2BigKB8GzKwtxLjS+c9vTq9lX2UBinI97X97IS3ecD9DlNOd3d5b3CIL26djb/4j4y8YD/G3LIR647izSEuNoCwQxM/4UPlPuvlc2MXFoGkWVjdz53Bpy08/jnPHHdq3Mku2lfH/BZgDe2lbaMwiqPmgJ7KtoYMqoTP625RDjBqd0nLjR7uElBaQm+NlVdpikeB9NrcGO1lMw6PjaU/nsOFTHLz43nedX7WPF7kpOH5J21JbNuqJqnn6vkO9dM5lDNU3kDU0nEHTsKa/vMT4FsHRHGV/67fv86NNTuOGcMcf073G8bKDNpT9r1iyXn58f6zLkJGkNBCmra+74y7yqoZV4v1FQepiRg5LJSIonKd5PMOjYerCWrJQElheU86mzR/LWtlL2VTbw+oYDfOXC8ewuO8wbmw9x3dkjeHDhNh69aSYl1Y2MyEyisLKBJ5bt4asXjufrF53Gr94u4On3CqmobwEHN5wzmpLqJpLifczNy2XrwVp+9/e9HXWOzk4mOyWBXWX1PPjpKfx5w34WbT5EQpyPn352Kr9asovth+p47OaZPLhwK4UVDR2nCWckxVHb1EZSvI8nvvRRLjg9hy/+9n3e2VHW8f5jB6dQ1xSad+oLs8cQCDjm5xfx1FfO4aHFO1lXVM33rp7Mq+v2c6i2ifE5qWw/WEd2agI7Sw93BGe7GWOyWLOvmlvOH8eBmkb+tuUQV5w1jJY2x5tbDzE6O5kbZ4/lx3/Zxo8/PYURWcm8tLqYpTvKmDoqkw3FNbx214UdXUdDMxL51gvrWbmnkivPGkZOeiJPLNtDWV0z00dnMSY7hRW7K/jOvI/wT/PX8/lZo5mfXwSEAvX9PZUMSk3g5TvOZ9nOcrYfquML54yhpLqR9KQ4hmUmsXZfNbPGDiLO76OxJcDqwipu/u1KJg3LYOqoTF7IL2L1/ZcTdI7rH1/BhXk5bD9Yx/qiaupbAvzyhrM5LTeNT/7yXWaNzeaF28/r+PcorKjnkp+9TdBBelIc00dnsWpvJTfNHsv9n5zMG5sPctvTqxmWkURNYystgSCBoGNkVjI/+vQU5k4MhWLnweWGljb+XlDBD17bTHFV6ELRkupG5t92Li+vCXW3/u7LHyU3LZExg1P409oSPjVjFHc+u4alO8qYNiqTP911Ycf7/n75Xj519sjj7qoys9XOuVkRX1MQiBcdbm4jNcF/3GeEOOd4eEkBQ9KTSAp3E505ItQVl54UT21TKxuLaxg9KIUxg1PYeaiOAzVNzJ2Yy6HaJv666SCHapsYkZXMxRMQVvIAAAsXSURBVGfk8veCci6aOIRhmUlA6PTeH76+hQeuPYuS6gb+tqWUtmCQrOR4Xl0X+ov6rktO5/994gxqm1q589k1vLszNN7x75+czLjBKdz6ZOj/yS3nj+Pl1cWMHJTMLeeP494/buRjHxlCRX0L64uqSY73c8sF47htzgRSE+P4nzd38NFx2UwbncW5Dy7uOOsJQi3C+6+axNW/XEZTa5DWYBC/GTPHDmLlnkqGZyZRVtdMWzh4/vHi0/j98r00tAQwC10lPyQ9kXe+fQmfe+w9NhTX8MSXZtHSFuSOZ9d0ab0lxPk6TihI8PtoCQSZk5fD5OEZPP7uboxQS/S1b1zIjkOHue7hv5OS4CcnLZHiqgbas2/uxFze2VHGnLwcyuqaO8aQvnTeWHLCH8IvrynhvV3ljMxKpi3oeObW2dz825UUVTby6RkjWby1lEEp8fzqxpkdY2H//snJPLOikKLKBuZ//VyeWbGPhRsPcM74bG6cPZaHlxSwsSR0uvfEoelsCZ+8kJUST3X4D5oEv4/6lgDpiXHUNbcxZWQmG0tqGDc4hb0VDXz9oglMHZnFgvUlLNp8iPvmTeJrcycc1++sgkDkFBEIOlbuqSAxzs+MMVkdQRYMOt4tKKe2sZUrzhqG34zn3t9HxeEWbj5vLGmJccT7Q+s+uXwvl04ayrDMJBZtDp0hNnZwasTtvZhfREsgyJD0JAJBx5y8HFIT49h6oJZX15WQFOentK6ZNYVVzJ6QzfevPpOgc7y+4QCFFQ3cc1kehRX17Dx0mE37a/j98r08c+tszhqZycrdFTyydBeP3jSTpHg/r6wt5vF39nDb3PEMSU/i0aW7uGrKcALOsfVALTlpiTzy9i6a24JcPnkoCX4fd1+axxnD0nHO8cjSXewtr+e19Qf45mV5xPl9/PD1LXzxvLE0twZ5ZW0J2akJ3HNZHt/702Zag0HaP/58FgrW2y8+DZ8ZSfF+7npuDa9vOEB2agIjs5L51Y0zGJ2dwu1Pr2ZDcTXL/vVj1Da1ctkv3uk4SeDa6SNYtrOcivoWEuN8/OSzU5k9fjAB53gxv4jG1gCPLd3dcYfDBxdu47wJg9lf08i54wczP7+IOXk5/OjTU/iXFzfw/t5KAkFHgt/Htz4+kdvmTjjuP14UBCLSL7S0BbuciXSsmloDlB9uZmRWcq8fiIGg67iifumOMs4ckUFOWiItbUHifIbPZyzZXkpGUjy5aYk0tgYYl5NCYlzXs6CaWgO0BoI9puFoaGmjvjnQccX8yt0VLNx4gKunjWDWuGxqm1rZXFLLGcPSye425tbcFuDdHeVc8pEhBJ1j6fYyLjojt2Msrbiqocu+1Te3seNQHacPSfvQ04EoCEREPO5IQTAwLg8UEZGoURCIiHicgkBExOMUBCIiHhfVIDCzK8xsu5kVmNm9EV7/ZzPbYmYbzGyxmY2NZj0iItJT1ILAzPzAw8CVwGTgBjOb3G21tcAs59xU4CXgJ9GqR0REIotmi+AcoMA5t9s51wI8D1zbeQXn3BLnXPtkICuAvt85QkRETohoBsFIoKjT8+Lwst7cCvwl0gtmdpuZ5ZtZfllZWaRVRETkOPWL2UfN7CZgFnBRpNedc48Dj4fXLTOzwuPcVA5QftS1BgbtS/+kfemftC/Q6xhsNIOgBBjd6fmo8LIuzOwy4D7gIudcc/fXu3POHd8k6KFt5fd2Zd1Ao33pn7Qv/ZP25cii2TW0Csgzs/FmlgBcDyzovIKZnQ08BlzjnCuNYi0iItKLqAWBc64NuAtYBGwFXnDObTazB8zsmvBqPwXSgBfNbJ2ZLejl7UREJEqiOkbgnFsILOy27LudHl8Wze1H8PhJ3l40aV/6J+1L/6R9OYIBN/uoiIicWJpiQkTE4xQEIiIe55kgONq8R/2dme01s43hQfX88LJsM/ubme0Mfx8U6zojMbPfmlmpmW3qtCxi7RbyUPg4bTCzGbGrvKde9uX7ZlYSPjbrzGxep9e+E96X7Wb2idhU3ZOZjTazJeG5vjab2T3h5QPuuBxhXwbicUkys/fNbH14X34QXj7ezFaGa54fPhMTM0sMPy8Ivz7uuDbsnDvlvwA/sAuYACQA64HJsa7rGPdhL5DTbdlPgHvDj+8F/ivWdfZS+1xgBrDpaLUD8whdYW7AucDKWNffh335PvD/Iqw7Ofy7lgiMD/8O+mO9D+HahgMzwo/TgR3hegfccTnCvgzE42JAWvhxPLAy/O/9AnB9ePmjwB3hx/8IPBp+fD0w/3i265UWwVHnPRqgrgWeDD9+ErguhrX0yjn3DlDZbXFvtV8LPOVCVgBZZjb85FR6dL3sS2+uBZ53zjU75/YABYR+F2POOXfAObcm/LiO0CneIxmAx+UI+9Kb/nxcnHPucPhpfPjLAR8jNDEn9Dwu7cfrJeBSs2O/u71XguBY5z3qjxzwhpmtNrPbwsuGOucOhB8fBIbGprTj0lvtA/VY3RXuMvltpy66AbEv4e6Eswn99Tmgj0u3fYEBeFzMzG9m64BS4G+EWizVLnRtFnStt2Nfwq/XAIOPdZteCYJTwYXOuRmEpvW+08zmdn7RhdqGA/Jc4IFce9gjwGnAdOAA8PPYltN3ZpYGvAx80zlX2/m1gXZcIuzLgDwuzrmAc246oWl5zgE+Eu1teiUI+jTvUX/mnCsJfy8FXiH0C3KovXke/j6QpunorfYBd6ycc4fC/3mDwK/5oJuhX++LmcUT+uB81jn3x/DiAXlcIu3LQD0u7Zxz1cAS4DxCXXHtFwB3rrdjX8KvZwIVx7otrwTBUec96s/MLNXM0tsfAx8HNhHahy+FV/sS8KfYVHhceqt9AfDF8Fkq5wI1nboq+qVufeWfInRsILQv14fP7BgP5AHvn+z6Ign3Iz8BbHXO/aLTSwPuuPS2LwP0uOSaWVb4cTJwOaExjyXAZ8OrdT8u7cfrs8Bb4ZbcsYn1KPnJ+iJ01sMOQv1t98W6nmOsfQKhsxzWA5vb6yfUF7gY2Am8CWTHutZe6v8DoaZ5K6H+zVt7q53QWRMPh4/TRkJ3sIv5PhxlX54O17oh/B9zeKf17wvvy3bgyljX36muCwl1+2wA1oW/5g3E43KEfRmIx2UqoTs3biAUXN8NL59AKKwKgBeBxPDypPDzgvDrE45nu5piQkTE47zSNSQiIr1QEIiIeJyCQETE4xQEIiIepyAQEfE4BYHISWRmF5vZ67GuQ6QzBYGIiMcpCEQiMLObwvPCrzOzx8ITgR02s/8OzxO/2Mxyw+tON7MV4cnNXuk0h//pZvZmeG75NWZ2Wvjt08zsJTPbZmbPHs9skSInkoJApBszmwR8HrjAhSb/CgA3AqlAvnPuTGAp8L3wjzwF/KtzbiqhK1nblz8LPOycmwacT+iKZAjNjvlNQvPiTwAuiPpOiRxB3NFXEfGcS4GZwKrwH+vJhCZfCwLzw+s8A/zRzDKBLOfc0vDyJ4EXw3NDjXTOvQLgnGsCCL/f+8654vDzdcA4YFn0d0skMgWBSE8GPOmc+06XhWb/3m29452fpbnT4wD6fygxpq4hkZ4WA581syHQcR/fsYT+v7TPAPkFYJlzrgaoMrM54eU3A0td6E5ZxWZ2Xfg9Es0s5aTuhUgf6S8RkW6cc1vM7H5Cd4TzEZpp9E6gHjgn/FopoXEECE0D/Gj4g3438OXw8puBx8zsgfB7/MNJ3A2RPtPsoyJ9ZGaHnXNpsa5D5ERT15CIiMepRSAi4nFqEYiIeJyCQETE4xQEIiIepyAQEfE4BYGIiMf9fyMmOqBywKxSAAAAAElFTkSuQmCC\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.plot(np.array(valid_losses)[:,0],np.array(valid_losses)[:,1])\n",
        "plt.xlabel(\"epoch\")\n",
        "plt.ylabel(\"valid loss\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 89,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 609
        },
        "id": "Or5Y-eFEJff1",
        "outputId": "55736684-8637-4cf9-f820-90258b8e4021"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.8816986855409504"
            ]
          },
          "metadata": {},
          "execution_count": 89
        }
      ],
      "source": [
        "lstm_model.accuracy(test_loader)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 90,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 366
        },
        "id": "_9kRgVB1XACk",
        "outputId": "3e8205f5-73c7-4cb2-9883-b41b1c46f8b2"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 0 Axes>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy=0.8816986855409504\n",
            "Precision=0.9826435246995995\n",
            "Recall=0.8761904761904762\n",
            "F1=0.9263687853996223\n"
          ]
        }
      ],
      "source": [
        "lstm_model.metrics(test_loader) "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "QllGxxs8XFps"
      },
      "outputs": [],
      "source": []
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "machine_shape": "hm",
      "provenance": []
    },
    "gpuClass": "premium",
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}