LSTM+attentionlayer.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "「LSTM完成版.ipynb」的副本",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lQ8pXahqWjoy",
        "outputId": "bfa5d250-d052-44b6-a961-33801475258c"
      },
      "source": [
        "from google.colab import drive\r\n",
        "drive.mount('/content/drive', force_remount=True)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "resources": {
            "http://localhost:8080/nbextensions/google.colab/files.js": {
              "data": "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",
              "ok": true,
              "headers": [
                [
                  "content-type",
                  "application/javascript"
                ]
              ],
              "status": 200,
              "status_text": ""
            }
          },
          "base_uri": "https://localhost:8080/",
          "height": 74
        },
        "id": "vk0TV7INXZJm",
        "outputId": "2b6e3223-099a-45d7-d813-38611cf48ee6"
      },
      "source": [
        "from google.colab import files\r\n",
        "uploaded_files = files.upload()\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "     <input type=\"file\" id=\"files-73a2b4de-f038-4274-8d68-51cee2dba8d8\" name=\"files[]\" multiple disabled\n",
              "        style=\"border:none\" />\n",
              "     <output id=\"result-73a2b4de-f038-4274-8d68-51cee2dba8d8\">\n",
              "      Upload widget is only available when the cell has been executed in the\n",
              "      current browser session. Please rerun this cell to enable.\n",
              "      </output>\n",
              "      <script src=\"/nbextensions/google.colab/files.js\"></script> "
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Saving preprocess.csv to preprocess (2).csv\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "48UfiuLDWsHT"
      },
      "source": [
        "import os\r\n",
        "\r\n",
        "import torch\r\n",
        "import numpy as np\r\n",
        "import pandas as pd\r\n",
        "import seaborn as sns\r\n",
        "import matplotlib.pyplot as plt\r\n",
        "from torch import nn, optim\r\n",
        "from torch.utils.data import Dataset, DataLoader\r\n",
        "import torch.nn.functional as F\r\n",
        "from torch.autograd import Variable\r\n",
        "import torchvision.models as models  \r\n",
        "\r\n",
        "import warnings# 移除不必要的警告\r\n",
        "warnings.filterwarnings('ignore')\r\n",
        "from sklearn.preprocessing import MinMaxScaler\r\n",
        "from xgboost import XGBRegressor\r\n",
        "from xgboost import plot_importance"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WpZNUx2fW2o0"
      },
      "source": [
        "import matplotlib as mpl\r\n",
        "zhfont = mpl.font_manager.FontProperties(fname='/content/drive/MyDrive/chinese.simhei.ttf')#取得中文字型\r\n",
        "\r\n",
        "sns.set(font=zhfont.get_name())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7gIdXfRuW3sp"
      },
      "source": [
        "df = pd.read_csv('preprocess.csv')\r\n",
        "\r\n",
        "df.set_index(keys =df.columns[0],inplace=True)\r\n",
        "df['y_class']=(df.y)>1\r\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "obxuG-VTW875",
        "outputId": "acf92913-1f9f-4ce2-d76b-f3a592bb2d61"
      },
      "source": [
        "df.index=pd.to_datetime(df.index,format=\"%Y/%m/%d\")\r\n",
        "df.index.name='date'\r\n",
        "feature=df.columns[:-2]\r\n",
        "print(len(feature))\r\n",
        "\r\n",
        "feature_len=len(feature)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "221\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9-KrWDrvXAix"
      },
      "source": [
        "DAYS_FOR_TRAIN=20\r\n",
        "def deal_lstm(df,feature,n = DAYS_FOR_TRAIN):#以20天作為預測\r\n",
        "    \r\n",
        "    X = []\r\n",
        "    y = []\r\n",
        "    dataset_scaled_x = df[feature]\r\n",
        "    \r\n",
        "    for i in range(0, len(df)-n):\r\n",
        "        X.append(dataset_scaled_x[i:i+n])\r\n",
        "        y.append(df['y'].iloc[i+n-1])\r\n",
        "    \r\n",
        "    return (np.array(X),np.array(y))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YKyOIh5aXDaB"
      },
      "source": [
        "def tol_normalize(ddf):\r\n",
        "\r\n",
        "  for col in feature:\r\n",
        "    ddf[col]=MinMaxScaler().fit_transform(ddf[col].values.reshape(-1, 1))\r\n",
        "\r\n",
        "  # ddf=MinMaxScaler().fit_transform(ddf)\r\n",
        "  return ddf"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "id": "YJvul5BdXF2R",
        "outputId": "4cf5e711-16c6-4932-ec62-abcca6e8e3f3"
      },
      "source": [
        "xgb = XGBRegressor(n_estimators=100)\r\n",
        "xgb.fit(df[:'2013'][feature], df[:'2013']['y'])\r\n",
        "sorted_idx = xgb.feature_importances_.argsort()\r\n",
        "# plot_importance(xgb, max_num_features = 20)#xgb內建畫圖，但比較不漂亮\r\n",
        "sns.set(font=zhfont.get_name())\r\n",
        "plt.yticks(fontproperties=zhfont)\r\n",
        "\r\n",
        "sns.barplot(xgb.feature_importances_[sorted_idx][-20::],feature[sorted_idx][-20::] )\r\n",
        "\r\n",
        "feature_new=feature[sorted_idx][-20::]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[11:07:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Rj0UMaYlXKpC"
      },
      "source": [
        "train_x,train_y=deal_lstm(tol_normalize(df[:'2013']),feature_new)\r\n",
        "valid_x,valid_y=deal_lstm((df['2013':'2017']),feature_new)\r\n",
        "test_x,test_y=deal_lstm((df['2017':]),feature_new)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OvxdnBpTXMTT",
        "outputId": "a5525a04-9aa9-4d9f-84e9-1a501e98e5c7"
      },
      "source": [
        "print(train_x.shape)\r\n",
        "print(train_y.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(1594, 20, 20)\n",
            "(1594,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vuysBwVAXOK6"
      },
      "source": [
        "class ElecDataset(Dataset):\r\n",
        "    def __init__(self,feature,target):\r\n",
        "        self.feature = feature\r\n",
        "        self.target = target\r\n",
        "    \r\n",
        "    def __len__(self):\r\n",
        "        return len(self.feature)\r\n",
        "    \r\n",
        "    def __getitem__(self,idx):\r\n",
        "        item = self.feature[idx]\r\n",
        "        label = self.target[idx]\r\n",
        "        \r\n",
        "        return item,label"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sgnbJ0HSvTZ7"
      },
      "source": [
        "import tensorflow as tf"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Pr3Hdrh0sAie"
      },
      "source": [
        "class Attention_Layer():\r\n",
        "    def __init__(\r\n",
        "        self,\r\n",
        "        timewindow_size,\r\n",
        "        input_hidden_unit_size,\r\n",
        "        attention_size=None):\r\n",
        "\r\n",
        "        if(attention_size is None):\r\n",
        "            attention_size=input_hidden_unit_size\r\n",
        "        self.o_size = attention_size\r\n",
        "        self.h_size = input_hidden_unit_size\r\n",
        "        self.t_size = timewindow_size\r\n",
        "\r\n",
        "        self.beta_weight = tf.Variable(tf.random.normal([self.h_size,self.o_size]), name='beta_weight')\r\n",
        "        self.beta_bias = tf.Variable(tf.zeros([self.o_size]),name='beta_bias')\r\n",
        "\r\n",
        "        self.v = tf.Variable(tf.random.normal([self.o_size,1]),name='beta_v')\r\n",
        "\r\n",
        "    def __call__(self,inputs):\r\n",
        "        #temp = tanh(Y X W + b) ->shape of result = (-1, self.o_size)\r\n",
        "        temp=tf.matmul(tf.reshape(inputs.cpu().detach().numpy(),[-1,self.h_size]),self.beta_weight)\r\n",
        "        #print(temp.shape)\r\n",
        "        temp=tf.tanh(temp + self.beta_bias)\r\n",
        "        #print(temp.shape)\r\n",
        "            \r\n",
        "        #j=temp X vv\r\n",
        "        j=tf.reshape(tf.matmul(temp,self.v),[-1,self.t_size,1])\r\n",
        "\r\n",
        "        #print(j.shape)\r\n",
        "\r\n",
        "        beta=tf.nn.softmax(j)\r\n",
        "\r\n",
        "        #print(beta.shape)\r\n",
        "        #print(inputs.shape)\r\n",
        "\r\n",
        "        output = beta*(inputs.cpu().detach().numpy())\r\n",
        "        #print(output.shape)\r\n",
        "        output = output.numpy()\r\n",
        "        output = Variable(torch.from_numpy(output))\r\n",
        "        return output"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1XsrxkOOckeJ"
      },
      "source": [
        "class LSTM(nn.Module):\r\n",
        "    def __init__(self, input_dim, hidden_dim=20, num_layers=1, output_dim=1):\r\n",
        "        super(LSTM, self).__init__()\r\n",
        "        # Hidden dimensions\r\n",
        "        self.hidden_dim = hidden_dim\r\n",
        "\r\n",
        "        # Number of hidden layers\r\n",
        "        self.num_layers = num_layers\r\n",
        "\r\n",
        "        # Building your LSTM\r\n",
        "        # batch_first=True causes input/output tensors to be of shape\r\n",
        "        # (batch_dim, seq_dim, feature_dim)\r\n",
        "        self.lstm = nn.LSTM(input_dim, hidden_dim, num_layers, batch_first=True)\r\n",
        "\r\n",
        "        # Readout layer\r\n",
        "        self.fc = nn.Linear(hidden_dim, output_dim)\r\n",
        "\r\n",
        "    def forward(self, x):\r\n",
        "        # Initialize hidden state with zeros\r\n",
        "        h0 = torch.zeros(self.num_layers, x.size(0), self.hidden_dim).requires_grad_().to(device)\r\n",
        "\r\n",
        "        # Initialize cell state\r\n",
        "        c0 = torch.zeros(self.num_layers, x.size(0), self.hidden_dim).requires_grad_().to(device)\r\n",
        "\r\n",
        "        # One time step\r\n",
        "        # If we don't, we'll backprop all the way to the start even after going through another batch\r\n",
        "        out, (hn, cn) = self.lstm(x)\r\n",
        "        #print('a',out.shape)\r\n",
        "\r\n",
        "        attention = Attention_Layer(20,20)\r\n",
        "\r\n",
        "        out = attention(out)\r\n",
        "        out = out.to(device)\r\n",
        "        #print('b',out.shape)\r\n",
        "        #print(type(out))\r\n",
        "\r\n",
        "        # Index hidden state of last time step\r\n",
        "        # out[:, -1, :] --> 100, 100 --> just want last time step hidden states! \r\n",
        "        out = self.fc(out[:, -1, :]) \r\n",
        "        # out.size() --> 100, 10\r\n",
        "        return out"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0XoQ_YZrLsWm"
      },
      "source": [
        "def MAPELoss(output,target):\r\n",
        "  return torch.mean(torch.abs((target - output)/target))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Izu2FwZ_XTJq",
        "outputId": "879cd939-0efb-4f20-9c81-cd76fbc041d5"
      },
      "source": [
        "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\r\n",
        "\r\n",
        "\r\n",
        "#model = BiLSTM_Attention(input_dim=20).to(device)\r\n",
        "model = LSTM(input_dim=20)\r\n",
        "model = model.to(device)\r\n",
        "model = model.float()\r\n",
        "\r\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=1e-5)\r\n",
        "model.to(device)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "LSTM(\n",
              "  (lstm): LSTM(20, 20, batch_first=True)\n",
              "  (fc): Linear(in_features=20, out_features=1, bias=True)\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 294
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Jeg-7Fw2XVvK"
      },
      "source": [
        "train = ElecDataset(train_x.reshape(train_x.shape[0],-1,train_x.shape[1]),train_y)\r\n",
        "valid = ElecDataset(valid_x.reshape(valid_x.shape[0],-1,valid_x.shape[1]),valid_y)\r\n",
        "train_loader = torch.utils.data.DataLoader(train,batch_size=128,shuffle=False)\r\n",
        "valid_loader = torch.utils.data.DataLoader(valid,batch_size=128,shuffle=False)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JXs6M0qJXX97"
      },
      "source": [
        "train_losses = []\r\n",
        "valid_losses = []\r\n",
        "train_accs = []\r\n",
        "val_accs = []\r\n",
        "\r\n",
        "def Train():\r\n",
        "    \r\n",
        "    running_loss = .0\r\n",
        "    train_acc = 0\r\n",
        "    \r\n",
        "    model.train()\r\n",
        "    \r\n",
        "    for idx, (inputs,labels) in enumerate(train_loader):\r\n",
        "        inputs = inputs.to(device)\r\n",
        "        inputs = inputs.float()\r\n",
        "        labels = labels.to(device)\r\n",
        "        labels = labels.float()\r\n",
        "        optimizer.zero_grad()\r\n",
        "        preds = model(inputs)\r\n",
        "        loss = MAPELoss(preds,labels)\r\n",
        "        #loss = \r\n",
        "        loss.backward()\r\n",
        "        optimizer.step()\r\n",
        "        running_loss += loss\r\n",
        "        mse = nn.MSELoss()\r\n",
        "        acc = mse(preds,labels)\r\n",
        "        train_acc = train_acc + acc.item()\r\n",
        "        \r\n",
        "    train_loss = running_loss/len(train_loader)\r\n",
        "    train_losses.append(train_loss.detach().cpu().numpy())\r\n",
        "    train_acc = train_acc/len(train_loader)\r\n",
        "    train_accs.append(train_acc)\r\n",
        "    \r\n",
        "    print(f'train_loss {train_loss}')\r\n",
        "    print(f'train_acc {train_acc}')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "r-qKOi7kXcea"
      },
      "source": [
        "def Valid():\r\n",
        "    running_loss = .0\r\n",
        "    val_acc = 0\r\n",
        "    \r\n",
        "    model.eval()\r\n",
        "    \r\n",
        "    with torch.no_grad():\r\n",
        "        for idx, (inputs, labels) in enumerate(valid_loader):\r\n",
        "            inputs = inputs.to(device)\r\n",
        "            inputs = inputs.float()\r\n",
        "            labels = labels.to(device)\r\n",
        "            labels = labels.float()\r\n",
        "            optimizer.zero_grad()\r\n",
        "            preds = model(inputs)\r\n",
        "            loss = MAPELoss(preds,labels)\r\n",
        "            running_loss += loss\r\n",
        "            mse = nn.MSELoss()\r\n",
        "            acc = mse(preds,labels)\r\n",
        "            val_acc = val_acc + acc.item()\r\n",
        "            \r\n",
        "        valid_loss = running_loss/len(valid_loader)\r\n",
        "        valid_losses.append(valid_loss.detach().cpu().numpy())\r\n",
        "        val_acc = val_acc/len(valid_loader)\r\n",
        "        val_accs.append(val_acc)\r\n",
        "\r\n",
        "        print(f'valid_loss {valid_loss}')\r\n",
        "        print(f'val_acc {val_acc}')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7n6xCckHXexC",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5c79b3c6-212b-4299-d752-6c4372e79828"
      },
      "source": [
        "epochs = 1000\r\n",
        "for epoch in range(epochs):\r\n",
        "    print('epochs {}/{}'.format(epoch+1,epochs))\r\n",
        "    Train()\r\n",
        "    Valid()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[1;30;43m串流輸出內容已截斷至最後 5000 行。\u001b[0m\n",
            "epochs 1/1000\n",
            "train_loss 161.82998657226562\n",
            "train_acc 0.05009651900484012\n",
            "valid_loss 256.04241943359375\n",
            "val_acc 0.14024005085229874\n",
            "epochs 2/1000\n",
            "train_loss 161.52783203125\n",
            "train_acc 0.04991111847070547\n",
            "valid_loss 255.85948181152344\n",
            "val_acc 0.14007083252072333\n",
            "epochs 3/1000\n",
            "train_loss 161.22573852539062\n",
            "train_acc 0.049726069833223634\n",
            "valid_loss 255.67649841308594\n",
            "val_acc 0.13990178108215331\n",
            "epochs 4/1000\n",
            "train_loss 160.9236297607422\n",
            "train_acc 0.04954137452519857\n",
            "valid_loss 255.49351501464844\n",
            "val_acc 0.13973288014531135\n",
            "epochs 5/1000\n",
            "train_loss 160.6215362548828\n",
            "train_acc 0.04935702853477918\n",
            "valid_loss 255.31057739257812\n",
            "val_acc 0.13956415206193923\n",
            "epochs 6/1000\n",
            "train_loss 160.3194122314453\n",
            "train_acc 0.04917303300820864\n",
            "valid_loss 255.12759399414062\n",
            "val_acc 0.13939557746052741\n",
            "epochs 7/1000\n",
            "train_loss 160.01731872558594\n",
            "train_acc 0.048989384793318234\n",
            "valid_loss 254.94461059570312\n",
            "val_acc 0.13922713845968246\n",
            "epochs 8/1000\n",
            "train_loss 159.71524047851562\n",
            "train_acc 0.048806086755715884\n",
            "valid_loss 254.7616729736328\n",
            "val_acc 0.1390588916838169\n",
            "epochs 9/1000\n",
            "train_loss 159.41314697265625\n",
            "train_acc 0.048623136889476046\n",
            "valid_loss 254.5786895751953\n",
            "val_acc 0.13889078721404075\n",
            "epochs 10/1000\n",
            "train_loss 159.11106872558594\n",
            "train_acc 0.04844053347523396\n",
            "valid_loss 254.39578247070312\n",
            "val_acc 0.13872287273406983\n",
            "epochs 11/1000\n",
            "train_loss 158.80899047851562\n",
            "train_acc 0.048258280811401516\n",
            "valid_loss 254.2128448486328\n",
            "val_acc 0.138555109500885\n",
            "epochs 12/1000\n",
            "train_loss 158.5069122314453\n",
            "train_acc 0.048076374599566825\n",
            "valid_loss 254.02989196777344\n",
            "val_acc 0.1383875258266926\n",
            "epochs 13/1000\n",
            "train_loss 158.20486450195312\n",
            "train_acc 0.047894815842692666\n",
            "valid_loss 253.8469696044922\n",
            "val_acc 0.13822008669376373\n",
            "epochs 14/1000\n",
            "train_loss 157.9027862548828\n",
            "train_acc 0.04771360497062023\n",
            "valid_loss 253.6640625\n",
            "val_acc 0.13805282190442086\n",
            "epochs 15/1000\n",
            "train_loss 157.6007080078125\n",
            "train_acc 0.047532740550545544\n",
            "valid_loss 253.4811248779297\n",
            "val_acc 0.13788572549819947\n",
            "epochs 16/1000\n",
            "train_loss 157.29864501953125\n",
            "train_acc 0.04735222516151575\n",
            "valid_loss 253.29820251464844\n",
            "val_acc 0.1377187676727772\n",
            "epochs 17/1000\n",
            "train_loss 156.99658203125\n",
            "train_acc 0.04717205636776411\n",
            "valid_loss 253.11526489257812\n",
            "val_acc 0.13755197450518608\n",
            "epochs 18/1000\n",
            "train_loss 156.69451904296875\n",
            "train_acc 0.04699223660505735\n",
            "valid_loss 252.932373046875\n",
            "val_acc 0.1373853713274002\n",
            "epochs 19/1000\n",
            "train_loss 156.3924560546875\n",
            "train_acc 0.046812763007787556\n",
            "valid_loss 252.74949645996094\n",
            "val_acc 0.1372189112007618\n",
            "epochs 20/1000\n",
            "train_loss 156.09039306640625\n",
            "train_acc 0.04663363600579592\n",
            "valid_loss 252.56655883789062\n",
            "val_acc 0.13705262020230294\n",
            "epochs 21/1000\n",
            "train_loss 155.788330078125\n",
            "train_acc 0.04645485918109234\n",
            "valid_loss 252.38365173339844\n",
            "val_acc 0.13688648641109466\n",
            "epochs 22/1000\n",
            "train_loss 155.48631286621094\n",
            "train_acc 0.04627643239039641\n",
            "valid_loss 252.2007598876953\n",
            "val_acc 0.1367205023765564\n",
            "epochs 23/1000\n",
            "train_loss 155.18426513671875\n",
            "train_acc 0.04609835763963369\n",
            "valid_loss 252.0178680419922\n",
            "val_acc 0.13655470460653304\n",
            "epochs 24/1000\n",
            "train_loss 154.88221740722656\n",
            "train_acc 0.04592063249303745\n",
            "valid_loss 251.83499145507812\n",
            "val_acc 0.1363890714943409\n",
            "epochs 25/1000\n",
            "train_loss 154.5802001953125\n",
            "train_acc 0.04574324992986826\n",
            "valid_loss 251.652099609375\n",
            "val_acc 0.13622358143329621\n",
            "epochs 26/1000\n",
            "train_loss 154.27816772460938\n",
            "train_acc 0.04556621926335188\n",
            "valid_loss 251.4691925048828\n",
            "val_acc 0.13605823889374732\n",
            "epochs 27/1000\n",
            "train_loss 153.9761505126953\n",
            "train_acc 0.045389535335394054\n",
            "valid_loss 251.28628540039062\n",
            "val_acc 0.13589307963848113\n",
            "epochs 28/1000\n",
            "train_loss 153.67410278320312\n",
            "train_acc 0.045213195137106456\n",
            "valid_loss 251.1033935546875\n",
            "val_acc 0.1357280820608139\n",
            "epochs 29/1000\n",
            "train_loss 153.37208557128906\n",
            "train_acc 0.045037203683302954\n",
            "valid_loss 250.92051696777344\n",
            "val_acc 0.1355632334947586\n",
            "epochs 30/1000\n",
            "train_loss 153.070068359375\n",
            "train_acc 0.04486156169038553\n",
            "valid_loss 250.73765563964844\n",
            "val_acc 0.1353985693305731\n",
            "epochs 31/1000\n",
            "train_loss 152.76805114746094\n",
            "train_acc 0.04468626944491497\n",
            "valid_loss 250.55479431152344\n",
            "val_acc 0.13523405641317368\n",
            "epochs 32/1000\n",
            "train_loss 152.4660186767578\n",
            "train_acc 0.04451132322160097\n",
            "valid_loss 250.3719024658203\n",
            "val_acc 0.1350696813315153\n",
            "epochs 33/1000\n",
            "train_loss 152.1640167236328\n",
            "train_acc 0.04433672416668672\n",
            "valid_loss 250.1890411376953\n",
            "val_acc 0.13490550331771373\n",
            "epochs 34/1000\n",
            "train_loss 151.8620147705078\n",
            "train_acc 0.04416247571890171\n",
            "valid_loss 250.0061798095703\n",
            "val_acc 0.13474148660898208\n",
            "epochs 35/1000\n",
            "train_loss 151.5600128173828\n",
            "train_acc 0.04398857673200277\n",
            "valid_loss 249.82334899902344\n",
            "val_acc 0.1345776416361332\n",
            "epochs 36/1000\n",
            "train_loss 151.25802612304688\n",
            "train_acc 0.04381502333741922\n",
            "valid_loss 249.64048767089844\n",
            "val_acc 0.13441395349800586\n",
            "epochs 37/1000\n",
            "train_loss 150.95602416992188\n",
            "train_acc 0.0436418172545158\n",
            "valid_loss 249.4575958251953\n",
            "val_acc 0.13425041884183883\n",
            "epochs 38/1000\n",
            "train_loss 150.65402221679688\n",
            "train_acc 0.043468957766890526\n",
            "valid_loss 249.2747802734375\n",
            "val_acc 0.13408706188201905\n",
            "epochs 39/1000\n",
            "train_loss 150.35203552246094\n",
            "train_acc 0.04329644487454341\n",
            "valid_loss 249.09190368652344\n",
            "val_acc 0.13392385542392732\n",
            "epochs 40/1000\n",
            "train_loss 150.05003356933594\n",
            "train_acc 0.04312427958043722\n",
            "valid_loss 248.90904235839844\n",
            "val_acc 0.13376080840826035\n",
            "epochs 41/1000\n",
            "train_loss 149.74803161621094\n",
            "train_acc 0.04295245787272087\n",
            "valid_loss 248.72618103027344\n",
            "val_acc 0.1335979148745537\n",
            "epochs 42/1000\n",
            "train_loss 149.44602966308594\n",
            "train_acc 0.042780980754357115\n",
            "valid_loss 248.54336547851562\n",
            "val_acc 0.13343519531190395\n",
            "epochs 43/1000\n",
            "train_loss 149.14402770996094\n",
            "train_acc 0.04260984808206558\n",
            "valid_loss 248.36050415039062\n",
            "val_acc 0.1332726262509823\n",
            "epochs 44/1000\n",
            "train_loss 148.842041015625\n",
            "train_acc 0.04243906386769735\n",
            "valid_loss 248.1776123046875\n",
            "val_acc 0.13311021812260151\n",
            "epochs 45/1000\n",
            "train_loss 148.5400390625\n",
            "train_acc 0.042268626391887665\n",
            "valid_loss 247.99478149414062\n",
            "val_acc 0.1329479832202196\n",
            "epochs 46/1000\n",
            "train_loss 148.238037109375\n",
            "train_acc 0.04209853465167376\n",
            "valid_loss 247.8119354248047\n",
            "val_acc 0.13278589956462383\n",
            "epochs 47/1000\n",
            "train_loss 147.93603515625\n",
            "train_acc 0.041928788647055626\n",
            "valid_loss 247.62905883789062\n",
            "val_acc 0.13262398391962052\n",
            "epochs 48/1000\n",
            "train_loss 147.63404846191406\n",
            "train_acc 0.04175939081380001\n",
            "valid_loss 247.4462432861328\n",
            "val_acc 0.1324622269719839\n",
            "epochs 49/1000\n",
            "train_loss 147.33206176757812\n",
            "train_acc 0.04159033771317739\n",
            "valid_loss 247.2633819580078\n",
            "val_acc 0.1323006346821785\n",
            "epochs 50/1000\n",
            "train_loss 147.03005981445312\n",
            "train_acc 0.04142163135111332\n",
            "valid_loss 247.08056640625\n",
            "val_acc 0.1321392022073269\n",
            "epochs 51/1000\n",
            "train_loss 146.72805786132812\n",
            "train_acc 0.041253270724645026\n",
            "valid_loss 246.89773559570312\n",
            "val_acc 0.13197792395949365\n",
            "epochs 52/1000\n",
            "train_loss 146.42605590820312\n",
            "train_acc 0.04108525655017449\n",
            "valid_loss 246.71487426757812\n",
            "val_acc 0.1318168170750141\n",
            "epochs 53/1000\n",
            "train_loss 146.1240692138672\n",
            "train_acc 0.04091758753817815\n",
            "valid_loss 246.531982421875\n",
            "val_acc 0.13165586963295936\n",
            "epochs 54/1000\n",
            "train_loss 145.82208251953125\n",
            "train_acc 0.040750265264740355\n",
            "valid_loss 246.3491973876953\n",
            "val_acc 0.1314950942993164\n",
            "epochs 55/1000\n",
            "train_loss 145.5200653076172\n",
            "train_acc 0.0405832901597023\n",
            "valid_loss 246.1663818359375\n",
            "val_acc 0.13133447840809823\n",
            "epochs 56/1000\n",
            "train_loss 145.2180938720703\n",
            "train_acc 0.04041666250962477\n",
            "valid_loss 245.98355102539062\n",
            "val_acc 0.13117402270436287\n",
            "epochs 57/1000\n",
            "train_loss 144.91610717773438\n",
            "train_acc 0.040250382744348966\n",
            "valid_loss 245.8007354736328\n",
            "val_acc 0.13101372718811036\n",
            "epochs 58/1000\n",
            "train_loss 144.61412048339844\n",
            "train_acc 0.04008445029075329\n",
            "valid_loss 245.6178741455078\n",
            "val_acc 0.1308535896241665\n",
            "epochs 59/1000\n",
            "train_loss 144.3121337890625\n",
            "train_acc 0.03991886428915537\n",
            "valid_loss 245.43504333496094\n",
            "val_acc 0.13069361187517642\n",
            "epochs 60/1000\n",
            "train_loss 144.01014709472656\n",
            "train_acc 0.039753625742517985\n",
            "valid_loss 245.25222778320312\n",
            "val_acc 0.13053381219506263\n",
            "epochs 61/1000\n",
            "train_loss 143.70816040039062\n",
            "train_acc 0.03958873493740192\n",
            "valid_loss 245.0694122314453\n",
            "val_acc 0.13037416487932205\n",
            "epochs 62/1000\n",
            "train_loss 143.4061737060547\n",
            "train_acc 0.039424190297722816\n",
            "valid_loss 244.8865966796875\n",
            "val_acc 0.13021467998623848\n",
            "epochs 63/1000\n",
            "train_loss 143.1042022705078\n",
            "train_acc 0.039259991536919885\n",
            "valid_loss 244.7038116455078\n",
            "val_acc 0.1300553608685732\n",
            "epochs 64/1000\n",
            "train_loss 142.80221557617188\n",
            "train_acc 0.03909613894155392\n",
            "valid_loss 244.52099609375\n",
            "val_acc 0.12989620566368104\n",
            "epochs 65/1000\n",
            "train_loss 142.50025939941406\n",
            "train_acc 0.038932631795222945\n",
            "valid_loss 244.3381805419922\n",
            "val_acc 0.1297371990978718\n",
            "epochs 66/1000\n",
            "train_loss 142.19825744628906\n",
            "train_acc 0.038769472103852495\n",
            "valid_loss 244.15538024902344\n",
            "val_acc 0.12957835495471953\n",
            "epochs 67/1000\n",
            "train_loss 141.8962860107422\n",
            "train_acc 0.03860665800479742\n",
            "valid_loss 243.9725799560547\n",
            "val_acc 0.12941967099905013\n",
            "epochs 68/1000\n",
            "train_loss 141.5943145751953\n",
            "train_acc 0.03844418949805773\n",
            "valid_loss 243.7897491455078\n",
            "val_acc 0.1292611502110958\n",
            "epochs 69/1000\n",
            "train_loss 141.29232788085938\n",
            "train_acc 0.03828206830299818\n",
            "valid_loss 243.60693359375\n",
            "val_acc 0.12910277843475343\n",
            "epochs 70/1000\n",
            "train_loss 140.99037170410156\n",
            "train_acc 0.038120292700254\n",
            "valid_loss 243.4241180419922\n",
            "val_acc 0.1289445847272873\n",
            "epochs 71/1000\n",
            "train_loss 140.68836975097656\n",
            "train_acc 0.03795886197342323\n",
            "valid_loss 243.24136352539062\n",
            "val_acc 0.1287865586578846\n",
            "epochs 72/1000\n",
            "train_loss 140.3863983154297\n",
            "train_acc 0.03779777970451575\n",
            "valid_loss 243.0585479736328\n",
            "val_acc 0.128628671169281\n",
            "epochs 73/1000\n",
            "train_loss 140.08444213867188\n",
            "train_acc 0.03763704231152168\n",
            "valid_loss 242.875732421875\n",
            "val_acc 0.12847096621990203\n",
            "epochs 74/1000\n",
            "train_loss 139.78244018554688\n",
            "train_acc 0.037476652230207734\n",
            "valid_loss 242.6929473876953\n",
            "val_acc 0.12831342369318008\n",
            "epochs 75/1000\n",
            "train_loss 139.48046875\n",
            "train_acc 0.03731660774120918\n",
            "valid_loss 242.51016235351562\n",
            "val_acc 0.12815603241324425\n",
            "epochs 76/1000\n",
            "train_loss 139.17849731445312\n",
            "train_acc 0.03715690984748877\n",
            "valid_loss 242.3273468017578\n",
            "val_acc 0.12799879722297192\n",
            "epochs 77/1000\n",
            "train_loss 138.8765106201172\n",
            "train_acc 0.03699755625656018\n",
            "valid_loss 242.1445770263672\n",
            "val_acc 0.12784174010157584\n",
            "epochs 78/1000\n",
            "train_loss 138.5745391845703\n",
            "train_acc 0.0368385508369941\n",
            "valid_loss 241.96177673339844\n",
            "val_acc 0.1276848278939724\n",
            "epochs 79/1000\n",
            "train_loss 138.27256774902344\n",
            "train_acc 0.03667989015006102\n",
            "valid_loss 241.7790069580078\n",
            "val_acc 0.12752808555960654\n",
            "epochs 80/1000\n",
            "train_loss 137.97059631347656\n",
            "train_acc 0.03652157634496689\n",
            "valid_loss 241.59622192382812\n",
            "val_acc 0.12737149894237518\n",
            "epochs 81/1000\n",
            "train_loss 137.66860961914062\n",
            "train_acc 0.03636360698594497\n",
            "valid_loss 241.41343688964844\n",
            "val_acc 0.12721507325768472\n",
            "epochs 82/1000\n",
            "train_loss 137.3666534423828\n",
            "train_acc 0.03620598680124833\n",
            "valid_loss 241.2306365966797\n",
            "val_acc 0.12705882117152215\n",
            "epochs 83/1000\n",
            "train_loss 137.06468200683594\n",
            "train_acc 0.0360487113491847\n",
            "valid_loss 241.04788208007812\n",
            "val_acc 0.12690271362662314\n",
            "epochs 84/1000\n",
            "train_loss 136.76271057128906\n",
            "train_acc 0.035891784355044365\n",
            "valid_loss 240.8651123046875\n",
            "val_acc 0.126746778935194\n",
            "epochs 85/1000\n",
            "train_loss 136.46072387695312\n",
            "train_acc 0.03573520309649981\n",
            "valid_loss 240.68235778808594\n",
            "val_acc 0.1265910230576992\n",
            "epochs 86/1000\n",
            "train_loss 136.15875244140625\n",
            "train_acc 0.035578966713868655\n",
            "valid_loss 240.4996337890625\n",
            "val_acc 0.12643543258309364\n",
            "epochs 87/1000\n",
            "train_loss 135.8568115234375\n",
            "train_acc 0.03542307749963724\n",
            "valid_loss 240.31689453125\n",
            "val_acc 0.1262799885123968\n",
            "epochs 88/1000\n",
            "train_loss 135.55482482910156\n",
            "train_acc 0.03526753488068397\n",
            "valid_loss 240.13418579101562\n",
            "val_acc 0.12612469792366027\n",
            "epochs 89/1000\n",
            "train_loss 135.2528533935547\n",
            "train_acc 0.035112336421242125\n",
            "valid_loss 239.95144653320312\n",
            "val_acc 0.12596957609057427\n",
            "epochs 90/1000\n",
            "train_loss 134.9508819580078\n",
            "train_acc 0.034957484843639225\n",
            "valid_loss 239.7687530517578\n",
            "val_acc 0.12581462264060975\n",
            "epochs 91/1000\n",
            "train_loss 134.64889526367188\n",
            "train_acc 0.03480298057771646\n",
            "valid_loss 239.5860595703125\n",
            "val_acc 0.12565984539687633\n",
            "epochs 92/1000\n",
            "train_loss 134.34693908691406\n",
            "train_acc 0.03464882176082868\n",
            "valid_loss 239.4034423828125\n",
            "val_acc 0.1255052089691162\n",
            "epochs 93/1000\n",
            "train_loss 134.0449676513672\n",
            "train_acc 0.03449500953921905\n",
            "valid_loss 239.22085571289062\n",
            "val_acc 0.12535074576735497\n",
            "epochs 94/1000\n",
            "train_loss 133.7429962158203\n",
            "train_acc 0.03434154448600916\n",
            "valid_loss 239.03831481933594\n",
            "val_acc 0.12519644796848298\n",
            "epochs 95/1000\n",
            "train_loss 133.4410400390625\n",
            "train_acc 0.03418842559823623\n",
            "valid_loss 238.85569763183594\n",
            "val_acc 0.12504228949546814\n",
            "epochs 96/1000\n",
            "train_loss 133.13906860351562\n",
            "train_acc 0.0340356517296571\n",
            "valid_loss 238.67312622070312\n",
            "val_acc 0.12488831467926502\n",
            "epochs 97/1000\n",
            "train_loss 132.8371124267578\n",
            "train_acc 0.03388322588916008\n",
            "valid_loss 238.4905548095703\n",
            "val_acc 0.12473449930548668\n",
            "epochs 98/1000\n",
            "train_loss 132.53515625\n",
            "train_acc 0.03373114635738043\n",
            "valid_loss 238.3079833984375\n",
            "val_acc 0.12458084151148796\n",
            "epochs 99/1000\n",
            "train_loss 132.23318481445312\n",
            "train_acc 0.03357941256119655\n",
            "valid_loss 238.12547302246094\n",
            "val_acc 0.12442734874784947\n",
            "epochs 100/1000\n",
            "train_loss 131.9312286376953\n",
            "train_acc 0.033428024930449635\n",
            "valid_loss 237.94290161132812\n",
            "val_acc 0.1242740087211132\n",
            "epochs 101/1000\n",
            "train_loss 131.62924194335938\n",
            "train_acc 0.033276982605457306\n",
            "valid_loss 237.7603759765625\n",
            "val_acc 0.12412084192037583\n",
            "epochs 102/1000\n",
            "train_loss 131.32730102539062\n",
            "train_acc 0.033126286875743136\n",
            "valid_loss 237.577880859375\n",
            "val_acc 0.12396783269941807\n",
            "epochs 103/1000\n",
            "train_loss 131.0253448486328\n",
            "train_acc 0.03297593659506394\n",
            "valid_loss 237.39541625976562\n",
            "val_acc 0.12381499260663986\n",
            "epochs 104/1000\n",
            "train_loss 130.72337341308594\n",
            "train_acc 0.03282593405590607\n",
            "valid_loss 237.21304321289062\n",
            "val_acc 0.12366230487823486\n",
            "epochs 105/1000\n",
            "train_loss 130.42140197753906\n",
            "train_acc 0.0326762768225028\n",
            "valid_loss 237.03079223632812\n",
            "val_acc 0.12350979559123516\n",
            "epochs 106/1000\n",
            "train_loss 130.1194610595703\n",
            "train_acc 0.03252696589781688\n",
            "valid_loss 236.8485107421875\n",
            "val_acc 0.12335742637515068\n",
            "epochs 107/1000\n",
            "train_loss 129.8175048828125\n",
            "train_acc 0.03237800056544634\n",
            "valid_loss 236.666259765625\n",
            "val_acc 0.12320521846413612\n",
            "epochs 108/1000\n",
            "train_loss 129.51553344726562\n",
            "train_acc 0.032229382831316725\n",
            "valid_loss 236.4840850830078\n",
            "val_acc 0.1230531945824623\n",
            "epochs 109/1000\n",
            "train_loss 129.2135772705078\n",
            "train_acc 0.032081109686539724\n",
            "valid_loss 236.30186462402344\n",
            "val_acc 0.1229013204574585\n",
            "epochs 110/1000\n",
            "train_loss 128.91159057617188\n",
            "train_acc 0.031933183710162454\n",
            "valid_loss 236.11962890625\n",
            "val_acc 0.12274960279464722\n",
            "epochs 111/1000\n",
            "train_loss 128.60964965820312\n",
            "train_acc 0.03178560289625938\n",
            "valid_loss 235.9374542236328\n",
            "val_acc 0.12259805575013161\n",
            "epochs 112/1000\n",
            "train_loss 128.3076934814453\n",
            "train_acc 0.0316383671015501\n",
            "valid_loss 235.7552490234375\n",
            "val_acc 0.12244666777551175\n",
            "epochs 113/1000\n",
            "train_loss 128.00572204589844\n",
            "train_acc 0.03149147904836214\n",
            "valid_loss 235.5730743408203\n",
            "val_acc 0.1222954373806715\n",
            "epochs 114/1000\n",
            "train_loss 127.70376586914062\n",
            "train_acc 0.031344935011405214\n",
            "valid_loss 235.39089965820312\n",
            "val_acc 0.12214438877999782\n",
            "epochs 115/1000\n",
            "train_loss 127.40180206298828\n",
            "train_acc 0.031198738429408807\n",
            "valid_loss 235.20872497558594\n",
            "val_acc 0.12199348062276841\n",
            "epochs 116/1000\n",
            "train_loss 127.09983825683594\n",
            "train_acc 0.03105288786956897\n",
            "valid_loss 235.0265655517578\n",
            "val_acc 0.12184274047613144\n",
            "epochs 117/1000\n",
            "train_loss 126.79788208007812\n",
            "train_acc 0.03090738419156808\n",
            "valid_loss 234.8444061279297\n",
            "val_acc 0.12169215939939022\n",
            "epochs 118/1000\n",
            "train_loss 126.49591064453125\n",
            "train_acc 0.03076222467307861\n",
            "valid_loss 234.6622772216797\n",
            "val_acc 0.12154175117611885\n",
            "epochs 119/1000\n",
            "train_loss 126.19395446777344\n",
            "train_acc 0.03061741189314769\n",
            "valid_loss 234.4801788330078\n",
            "val_acc 0.12139149680733681\n",
            "epochs 120/1000\n",
            "train_loss 125.89201354980469\n",
            "train_acc 0.030472944848812543\n",
            "valid_loss 234.2981719970703\n",
            "val_acc 0.12124141305685043\n",
            "epochs 121/1000\n",
            "train_loss 125.59003448486328\n",
            "train_acc 0.030328823826633967\n",
            "valid_loss 234.1162109375\n",
            "val_acc 0.12109146416187286\n",
            "epochs 122/1000\n",
            "train_loss 125.28807067871094\n",
            "train_acc 0.030185048969892356\n",
            "valid_loss 233.934326171875\n",
            "val_acc 0.1209416925907135\n",
            "epochs 123/1000\n",
            "train_loss 124.98611450195312\n",
            "train_acc 0.030041620278587707\n",
            "valid_loss 233.75244140625\n",
            "val_acc 0.1207920990884304\n",
            "epochs 124/1000\n",
            "train_loss 124.68415069580078\n",
            "train_acc 0.02989853760943963\n",
            "valid_loss 233.57058715820312\n",
            "val_acc 0.12064264081418515\n",
            "epochs 125/1000\n",
            "train_loss 124.38220977783203\n",
            "train_acc 0.029755800675887328\n",
            "valid_loss 233.38870239257812\n",
            "val_acc 0.12049334906041623\n",
            "epochs 126/1000\n",
            "train_loss 124.08023071289062\n",
            "train_acc 0.02961341220025833\n",
            "valid_loss 233.2068328857422\n",
            "val_acc 0.12034423351287842\n",
            "epochs 127/1000\n",
            "train_loss 123.77826690673828\n",
            "train_acc 0.029471368600542728\n",
            "valid_loss 233.02500915527344\n",
            "val_acc 0.12019527107477188\n",
            "epochs 128/1000\n",
            "train_loss 123.47632598876953\n",
            "train_acc 0.02932967173938568\n",
            "valid_loss 232.84317016601562\n",
            "val_acc 0.1200464703142643\n",
            "epochs 129/1000\n",
            "train_loss 123.17437744140625\n",
            "train_acc 0.029188320613824405\n",
            "valid_loss 232.6613311767578\n",
            "val_acc 0.11989782825112343\n",
            "epochs 130/1000\n",
            "train_loss 122.87242126464844\n",
            "train_acc 0.029047315653700095\n",
            "valid_loss 232.4794921875\n",
            "val_acc 0.11974935233592987\n",
            "epochs 131/1000\n",
            "train_loss 122.57046508789062\n",
            "train_acc 0.028906657002293147\n",
            "valid_loss 232.2976837158203\n",
            "val_acc 0.11960105895996094\n",
            "epochs 132/1000\n",
            "train_loss 122.26850128173828\n",
            "train_acc 0.02876634394320158\n",
            "valid_loss 232.11582946777344\n",
            "val_acc 0.11945289447903633\n",
            "epochs 133/1000\n",
            "train_loss 121.96653747558594\n",
            "train_acc 0.028626376333144996\n",
            "valid_loss 231.9340057373047\n",
            "val_acc 0.11930491030216217\n",
            "epochs 134/1000\n",
            "train_loss 121.66459655761719\n",
            "train_acc 0.02848675546164696\n",
            "valid_loss 231.75218200683594\n",
            "val_acc 0.1191570833325386\n",
            "epochs 135/1000\n",
            "train_loss 121.36262512207031\n",
            "train_acc 0.02834748075558589\n",
            "valid_loss 231.5703887939453\n",
            "val_acc 0.11900941692292691\n",
            "epochs 136/1000\n",
            "train_loss 121.06068420410156\n",
            "train_acc 0.02820855092543822\n",
            "valid_loss 231.3885498046875\n",
            "val_acc 0.11886190809309483\n",
            "epochs 137/1000\n",
            "train_loss 120.75872802734375\n",
            "train_acc 0.02806996812040989\n",
            "valid_loss 231.2067413330078\n",
            "val_acc 0.1187145620584488\n",
            "epochs 138/1000\n",
            "train_loss 120.4567642211914\n",
            "train_acc 0.027931730191294964\n",
            "valid_loss 231.02490234375\n",
            "val_acc 0.11856739334762097\n",
            "epochs 139/1000\n",
            "train_loss 120.15481567382812\n",
            "train_acc 0.027793840003701355\n",
            "valid_loss 230.8431396484375\n",
            "val_acc 0.11842037923634052\n",
            "epochs 140/1000\n",
            "train_loss 119.85287475585938\n",
            "train_acc 0.027656295265142735\n",
            "valid_loss 230.6613006591797\n",
            "val_acc 0.11827351450920105\n",
            "epochs 141/1000\n",
            "train_loss 119.55091094970703\n",
            "train_acc 0.0275190958323387\n",
            "valid_loss 230.4794921875\n",
            "val_acc 0.11812681928277016\n",
            "epochs 142/1000\n",
            "train_loss 119.24895477294922\n",
            "train_acc 0.027382242851532422\n",
            "valid_loss 230.2976837158203\n",
            "val_acc 0.11798029541969299\n",
            "epochs 143/1000\n",
            "train_loss 118.94699096679688\n",
            "train_acc 0.027245735749602318\n",
            "valid_loss 230.1158905029297\n",
            "val_acc 0.11783393435180187\n",
            "epochs 144/1000\n",
            "train_loss 118.6450424194336\n",
            "train_acc 0.027109574813109178\n",
            "valid_loss 229.93406677246094\n",
            "val_acc 0.11768771968781948\n",
            "epochs 145/1000\n",
            "train_loss 118.34307098388672\n",
            "train_acc 0.026973760615174588\n",
            "valid_loss 229.7522735595703\n",
            "val_acc 0.11754167526960373\n",
            "epochs 146/1000\n",
            "train_loss 118.0411148071289\n",
            "train_acc 0.026838291293153398\n",
            "valid_loss 229.5704345703125\n",
            "val_acc 0.11739578656852245\n",
            "epochs 147/1000\n",
            "train_loss 117.73916625976562\n",
            "train_acc 0.026703168996251546\n",
            "valid_loss 229.38865661621094\n",
            "val_acc 0.11725007258355617\n",
            "epochs 148/1000\n",
            "train_loss 117.43721008300781\n",
            "train_acc 0.026568391288702305\n",
            "valid_loss 229.2068328857422\n",
            "val_acc 0.11710450388491153\n",
            "epochs 149/1000\n",
            "train_loss 117.13526916503906\n",
            "train_acc 0.026433959746590026\n",
            "valid_loss 229.02505493164062\n",
            "val_acc 0.1169590998440981\n",
            "epochs 150/1000\n",
            "train_loss 116.83331298828125\n",
            "train_acc 0.026299875372877486\n",
            "valid_loss 228.84326171875\n",
            "val_acc 0.11681386791169643\n",
            "epochs 151/1000\n",
            "train_loss 116.53135681152344\n",
            "train_acc 0.02616613616163914\n",
            "valid_loss 228.66148376464844\n",
            "val_acc 0.11666878908872605\n",
            "epochs 152/1000\n",
            "train_loss 116.22940826416016\n",
            "train_acc 0.026032743259118155\n",
            "valid_loss 228.4796905517578\n",
            "val_acc 0.11652389504015445\n",
            "epochs 153/1000\n",
            "train_loss 115.92743682861328\n",
            "train_acc 0.025899696378753737\n",
            "valid_loss 228.29788208007812\n",
            "val_acc 0.11637914218008519\n",
            "epochs 154/1000\n",
            "train_loss 115.62549591064453\n",
            "train_acc 0.025766995663826283\n",
            "valid_loss 228.11611938476562\n",
            "val_acc 0.11623455211520195\n",
            "epochs 155/1000\n",
            "train_loss 115.32353973388672\n",
            "train_acc 0.02563464197401817\n",
            "valid_loss 227.93435668945312\n",
            "val_acc 0.11609011590480804\n",
            "epochs 156/1000\n",
            "train_loss 115.0215835571289\n",
            "train_acc 0.025502632587001875\n",
            "valid_loss 227.7525634765625\n",
            "val_acc 0.11594585329294205\n",
            "epochs 157/1000\n",
            "train_loss 114.71963500976562\n",
            "train_acc 0.02537097051166571\n",
            "valid_loss 227.57080078125\n",
            "val_acc 0.11580175273120404\n",
            "epochs 158/1000\n",
            "train_loss 114.41769409179688\n",
            "train_acc 0.02523965431520572\n",
            "valid_loss 227.3890380859375\n",
            "val_acc 0.11565780490636826\n",
            "epochs 159/1000\n",
            "train_loss 114.11573791503906\n",
            "train_acc 0.025108683854341507\n",
            "valid_loss 227.20730590820312\n",
            "val_acc 0.11551402434706688\n",
            "epochs 160/1000\n",
            "train_loss 113.81378173828125\n",
            "train_acc 0.02497805869923188\n",
            "valid_loss 227.02554321289062\n",
            "val_acc 0.11537040509283543\n",
            "epochs 161/1000\n",
            "train_loss 113.51182556152344\n",
            "train_acc 0.024847780497601397\n",
            "valid_loss 226.8438262939453\n",
            "val_acc 0.11522694528102875\n",
            "epochs 162/1000\n",
            "train_loss 113.20986938476562\n",
            "train_acc 0.024717847673365705\n",
            "valid_loss 226.662109375\n",
            "val_acc 0.11508365012705327\n",
            "epochs 163/1000\n",
            "train_loss 112.90792846679688\n",
            "train_acc 0.024588261372767962\n",
            "valid_loss 226.48037719726562\n",
            "val_acc 0.11494050733745098\n",
            "epochs 164/1000\n",
            "train_loss 112.6059799194336\n",
            "train_acc 0.024459020592845403\n",
            "valid_loss 226.29869079589844\n",
            "val_acc 0.1147975355386734\n",
            "epochs 165/1000\n",
            "train_loss 112.30402374267578\n",
            "train_acc 0.024330126050000008\n",
            "valid_loss 226.11705017089844\n",
            "val_acc 0.11465472914278507\n",
            "epochs 166/1000\n",
            "train_loss 112.00206756591797\n",
            "train_acc 0.024201577099469993\n",
            "valid_loss 225.9353790283203\n",
            "val_acc 0.1145120844244957\n",
            "epochs 167/1000\n",
            "train_loss 111.70010375976562\n",
            "train_acc 0.024073374242736742\n",
            "valid_loss 225.7537078857422\n",
            "val_acc 0.11436959728598595\n",
            "epochs 168/1000\n",
            "train_loss 111.39816284179688\n",
            "train_acc 0.023945517408160064\n",
            "valid_loss 225.5721435546875\n",
            "val_acc 0.1142272673547268\n",
            "epochs 169/1000\n",
            "train_loss 111.09620666503906\n",
            "train_acc 0.023818007598702725\n",
            "valid_loss 225.39065551757812\n",
            "val_acc 0.11408510729670525\n",
            "epochs 170/1000\n",
            "train_loss 110.79425811767578\n",
            "train_acc 0.023690842020397004\n",
            "valid_loss 225.20957946777344\n",
            "val_acc 0.1139431033283472\n",
            "epochs 171/1000\n",
            "train_loss 110.49231719970703\n",
            "train_acc 0.023564023180649832\n",
            "valid_loss 225.0286407470703\n",
            "val_acc 0.11380127407610416\n",
            "epochs 172/1000\n",
            "train_loss 110.19034576416016\n",
            "train_acc 0.023437549718297444\n",
            "valid_loss 224.8477783203125\n",
            "val_acc 0.11365959644317628\n",
            "epochs 173/1000\n",
            "train_loss 109.88839721679688\n",
            "train_acc 0.023311422493022222\n",
            "valid_loss 224.6669464111328\n",
            "val_acc 0.11351808048784733\n",
            "epochs 174/1000\n",
            "train_loss 109.5864486694336\n",
            "train_acc 0.023185641433183964\n",
            "valid_loss 224.4861297607422\n",
            "val_acc 0.11337671279907227\n",
            "epochs 175/1000\n",
            "train_loss 109.28448486328125\n",
            "train_acc 0.02306020625222188\n",
            "valid_loss 224.3053436279297\n",
            "val_acc 0.11323552876710892\n",
            "epochs 176/1000\n",
            "train_loss 108.98253631591797\n",
            "train_acc 0.022935116878495768\n",
            "valid_loss 224.1245880126953\n",
            "val_acc 0.11309449560940266\n",
            "epochs 177/1000\n",
            "train_loss 108.68059539794922\n",
            "train_acc 0.02281037402840761\n",
            "valid_loss 223.94387817382812\n",
            "val_acc 0.11295363567769527\n",
            "epochs 178/1000\n",
            "train_loss 108.3786392211914\n",
            "train_acc 0.022685979421322163\n",
            "valid_loss 223.76316833496094\n",
            "val_acc 0.11281292513012886\n",
            "epochs 179/1000\n",
            "train_loss 108.07669830322266\n",
            "train_acc 0.022561931839356057\n",
            "valid_loss 223.58255004882812\n",
            "val_acc 0.11267239786684513\n",
            "epochs 180/1000\n",
            "train_loss 107.7747573852539\n",
            "train_acc 0.022438231569070082\n",
            "valid_loss 223.40188598632812\n",
            "val_acc 0.11253200247883796\n",
            "epochs 181/1000\n",
            "train_loss 107.47281646728516\n",
            "train_acc 0.022314876676178895\n",
            "valid_loss 223.2213134765625\n",
            "val_acc 0.11239179670810699\n",
            "epochs 182/1000\n",
            "train_loss 107.17086791992188\n",
            "train_acc 0.02219186845020606\n",
            "valid_loss 223.04080200195312\n",
            "val_acc 0.11225174441933632\n",
            "epochs 183/1000\n",
            "train_loss 106.86892700195312\n",
            "train_acc 0.02206920488522603\n",
            "valid_loss 222.86038208007812\n",
            "val_acc 0.11211185082793236\n",
            "epochs 184/1000\n",
            "train_loss 106.5669937133789\n",
            "train_acc 0.021946888130444746\n",
            "valid_loss 222.6800079345703\n",
            "val_acc 0.11197212040424347\n",
            "epochs 185/1000\n",
            "train_loss 106.26505279541016\n",
            "train_acc 0.021824916967978843\n",
            "valid_loss 222.499755859375\n",
            "val_acc 0.11183254942297935\n",
            "epochs 186/1000\n",
            "train_loss 105.96311950683594\n",
            "train_acc 0.021703292185870502\n",
            "valid_loss 222.31961059570312\n",
            "val_acc 0.11169314831495285\n",
            "epochs 187/1000\n",
            "train_loss 105.66117095947266\n",
            "train_acc 0.021582013210998133\n",
            "valid_loss 222.1396026611328\n",
            "val_acc 0.11155391447246074\n",
            "epochs 188/1000\n",
            "train_loss 105.35923767089844\n",
            "train_acc 0.02146108068812352\n",
            "valid_loss 221.95994567871094\n",
            "val_acc 0.11141483373939991\n",
            "epochs 189/1000\n",
            "train_loss 105.05728912353516\n",
            "train_acc 0.0213404932560829\n",
            "valid_loss 221.7815399169922\n",
            "val_acc 0.11127590611577035\n",
            "epochs 190/1000\n",
            "train_loss 104.7553482055664\n",
            "train_acc 0.021220251774558656\n",
            "valid_loss 221.6053466796875\n",
            "val_acc 0.1111371573060751\n",
            "epochs 191/1000\n",
            "train_loss 104.45340728759766\n",
            "train_acc 0.021100356315190975\n",
            "valid_loss 221.4295196533203\n",
            "val_acc 0.11099854707717896\n",
            "epochs 192/1000\n",
            "train_loss 104.15147399902344\n",
            "train_acc 0.020980807594381846\n",
            "valid_loss 221.254150390625\n",
            "val_acc 0.11086011305451393\n",
            "epochs 193/1000\n",
            "train_loss 103.84954071044922\n",
            "train_acc 0.020861603462925322\n",
            "valid_loss 221.0792236328125\n",
            "val_acc 0.11072184145450592\n",
            "epochs 194/1000\n",
            "train_loss 103.54759216308594\n",
            "train_acc 0.020742746213307746\n",
            "valid_loss 220.9044189453125\n",
            "val_acc 0.11058372743427754\n",
            "epochs 195/1000\n",
            "train_loss 103.24566650390625\n",
            "train_acc 0.02062423570224872\n",
            "valid_loss 220.729736328125\n",
            "val_acc 0.11044578179717064\n",
            "epochs 196/1000\n",
            "train_loss 102.94371032714844\n",
            "train_acc 0.020506069923822697\n",
            "valid_loss 220.55503845214844\n",
            "val_acc 0.11030798833817243\n",
            "epochs 197/1000\n",
            "train_loss 102.64178466796875\n",
            "train_acc 0.02038825088395522\n",
            "valid_loss 220.3804931640625\n",
            "val_acc 0.1101703729480505\n",
            "epochs 198/1000\n",
            "train_loss 102.33982849121094\n",
            "train_acc 0.020270777866244316\n",
            "valid_loss 220.2059326171875\n",
            "val_acc 0.11003291215747595\n",
            "epochs 199/1000\n",
            "train_loss 102.03790283203125\n",
            "train_acc 0.020153651443811562\n",
            "valid_loss 220.0314178466797\n",
            "val_acc 0.10989561788737774\n",
            "epochs 200/1000\n",
            "train_loss 101.7359619140625\n",
            "train_acc 0.020036869539091222\n",
            "valid_loss 219.8568878173828\n",
            "val_acc 0.10975848808884621\n",
            "epochs 201/1000\n",
            "train_loss 101.43402862548828\n",
            "train_acc 0.01992043465949022\n",
            "valid_loss 219.68240356445312\n",
            "val_acc 0.10962151251733303\n",
            "epochs 202/1000\n",
            "train_loss 101.132080078125\n",
            "train_acc 0.019804345085643806\n",
            "valid_loss 219.50791931152344\n",
            "val_acc 0.10948468912392854\n",
            "epochs 203/1000\n",
            "train_loss 100.83015441894531\n",
            "train_acc 0.019688601462313764\n",
            "valid_loss 219.33348083496094\n",
            "val_acc 0.10934804677963257\n",
            "epochs 204/1000\n",
            "train_loss 100.52821350097656\n",
            "train_acc 0.019573204434261873\n",
            "valid_loss 219.1590118408203\n",
            "val_acc 0.10921155121177435\n",
            "epochs 205/1000\n",
            "train_loss 100.22628784179688\n",
            "train_acc 0.01945815307016556\n",
            "valid_loss 218.9845733642578\n",
            "val_acc 0.10907523427158594\n",
            "epochs 206/1000\n",
            "train_loss 99.9243392944336\n",
            "train_acc 0.01934344772822582\n",
            "valid_loss 218.81016540527344\n",
            "val_acc 0.10893906243145465\n",
            "epochs 207/1000\n",
            "train_loss 99.62239074707031\n",
            "train_acc 0.01922908783532106\n",
            "valid_loss 218.63572692871094\n",
            "val_acc 0.10880306176841259\n",
            "epochs 208/1000\n",
            "train_loss 99.32047271728516\n",
            "train_acc 0.019115074537694454\n",
            "valid_loss 218.4613037109375\n",
            "val_acc 0.10866721775382757\n",
            "epochs 209/1000\n",
            "train_loss 99.01852416992188\n",
            "train_acc 0.019001406617462635\n",
            "valid_loss 218.28689575195312\n",
            "val_acc 0.1085315439850092\n",
            "epochs 210/1000\n",
            "train_loss 98.71659088134766\n",
            "train_acc 0.018888084719387386\n",
            "valid_loss 218.1125030517578\n",
            "val_acc 0.10839602295309306\n",
            "epochs 211/1000\n",
            "train_loss 98.41466522216797\n",
            "train_acc 0.018775108843468703\n",
            "valid_loss 217.9380645751953\n",
            "val_acc 0.10826067198067904\n",
            "epochs 212/1000\n",
            "train_loss 98.11270904541016\n",
            "train_acc 0.018662479204627182\n",
            "valid_loss 217.763671875\n",
            "val_acc 0.10812549013644457\n",
            "epochs 213/1000\n",
            "train_loss 97.81078338623047\n",
            "train_acc 0.01855019501482065\n",
            "valid_loss 217.58926391601562\n",
            "val_acc 0.10799044128507376\n",
            "epochs 214/1000\n",
            "train_loss 97.50885772705078\n",
            "train_acc 0.018438257133731477\n",
            "valid_loss 217.4148712158203\n",
            "val_acc 0.10785558000206948\n",
            "epochs 215/1000\n",
            "train_loss 97.20692443847656\n",
            "train_acc 0.01832666563299986\n",
            "valid_loss 217.2404327392578\n",
            "val_acc 0.10772086791694165\n",
            "epochs 216/1000\n",
            "train_loss 96.90498352050781\n",
            "train_acc 0.018215418936541446\n",
            "valid_loss 217.0660400390625\n",
            "val_acc 0.10758632235229015\n",
            "epochs 217/1000\n",
            "train_loss 96.6030502319336\n",
            "train_acc 0.01810451940848277\n",
            "valid_loss 216.8916473388672\n",
            "val_acc 0.10745194610208272\n",
            "epochs 218/1000\n",
            "train_loss 96.30110168457031\n",
            "train_acc 0.01799396497125809\n",
            "valid_loss 216.7172393798828\n",
            "val_acc 0.10731771755963564\n",
            "epochs 219/1000\n",
            "train_loss 95.99917602539062\n",
            "train_acc 0.017883757272591956\n",
            "valid_loss 216.54287719726562\n",
            "val_acc 0.10718366149812937\n",
            "epochs 220/1000\n",
            "train_loss 95.6972427368164\n",
            "train_acc 0.017773895022960808\n",
            "valid_loss 216.3684844970703\n",
            "val_acc 0.10704976674169302\n",
            "epochs 221/1000\n",
            "train_loss 95.39530181884766\n",
            "train_acc 0.01766437843728524\n",
            "valid_loss 216.19412231445312\n",
            "val_acc 0.1069160457700491\n",
            "epochs 222/1000\n",
            "train_loss 95.09337615966797\n",
            "train_acc 0.01755520923493\n",
            "valid_loss 216.0197296142578\n",
            "val_acc 0.1067824624478817\n",
            "epochs 223/1000\n",
            "train_loss 94.79145050048828\n",
            "train_acc 0.017446385338329352\n",
            "valid_loss 215.8453369140625\n",
            "val_acc 0.10664904117584229\n",
            "epochs 224/1000\n",
            "train_loss 94.489501953125\n",
            "train_acc 0.01733790789372646\n",
            "valid_loss 215.6709747314453\n",
            "val_acc 0.1065158024430275\n",
            "epochs 225/1000\n",
            "train_loss 94.18757629394531\n",
            "train_acc 0.017229776113079145\n",
            "valid_loss 215.49661254882812\n",
            "val_acc 0.1063827246427536\n",
            "epochs 226/1000\n",
            "train_loss 93.88565063476562\n",
            "train_acc 0.017121990784429587\n",
            "valid_loss 215.322265625\n",
            "val_acc 0.10624979361891747\n",
            "epochs 227/1000\n",
            "train_loss 93.5837173461914\n",
            "train_acc 0.017014550618254222\n",
            "valid_loss 215.1478729248047\n",
            "val_acc 0.10611704085022211\n",
            "epochs 228/1000\n",
            "train_loss 93.28179168701172\n",
            "train_acc 0.016907457190637406\n",
            "valid_loss 214.9735565185547\n",
            "val_acc 0.10598444566130638\n",
            "epochs 229/1000\n",
            "train_loss 92.9798583984375\n",
            "train_acc 0.016800709928457554\n",
            "valid_loss 214.79922485351562\n",
            "val_acc 0.10585201177746058\n",
            "epochs 230/1000\n",
            "train_loss 92.67792510986328\n",
            "train_acc 0.01669430797203229\n",
            "valid_loss 214.6248779296875\n",
            "val_acc 0.10571973752230406\n",
            "epochs 231/1000\n",
            "train_loss 92.3759994506836\n",
            "train_acc 0.0165882519661234\n",
            "valid_loss 214.45054626464844\n",
            "val_acc 0.10558762736618518\n",
            "epochs 232/1000\n",
            "train_loss 92.0740737915039\n",
            "train_acc 0.01648254219729167\n",
            "valid_loss 214.2761993408203\n",
            "val_acc 0.10545567572116851\n",
            "epochs 233/1000\n",
            "train_loss 91.77216339111328\n",
            "train_acc 0.016377180313261654\n",
            "valid_loss 214.1020050048828\n",
            "val_acc 0.1053239043802023\n",
            "epochs 234/1000\n",
            "train_loss 91.4702377319336\n",
            "train_acc 0.016272166672234353\n",
            "valid_loss 213.9277801513672\n",
            "val_acc 0.10519226789474487\n",
            "epochs 235/1000\n",
            "train_loss 91.16831970214844\n",
            "train_acc 0.016167498695162628\n",
            "valid_loss 213.753662109375\n",
            "val_acc 0.10506081432104111\n",
            "epochs 236/1000\n",
            "train_loss 90.86640930175781\n",
            "train_acc 0.01606317674024747\n",
            "valid_loss 213.5795135498047\n",
            "val_acc 0.1049295162782073\n",
            "epochs 237/1000\n",
            "train_loss 90.56450653076172\n",
            "train_acc 0.015959201094049674\n",
            "valid_loss 213.4053955078125\n",
            "val_acc 0.10479839183390141\n",
            "epochs 238/1000\n",
            "train_loss 90.26258850097656\n",
            "train_acc 0.015855570753606465\n",
            "valid_loss 213.2312469482422\n",
            "val_acc 0.1046674270182848\n",
            "epochs 239/1000\n",
            "train_loss 89.96067810058594\n",
            "train_acc 0.01575228636367963\n",
            "valid_loss 213.05715942382812\n",
            "val_acc 0.10453661382198334\n",
            "epochs 240/1000\n",
            "train_loss 89.65876007080078\n",
            "train_acc 0.01564934878395154\n",
            "valid_loss 212.8830108642578\n",
            "val_acc 0.1044059544801712\n",
            "epochs 241/1000\n",
            "train_loss 89.35686492919922\n",
            "train_acc 0.015546756868179027\n",
            "valid_loss 212.7089080810547\n",
            "val_acc 0.10427547432482243\n",
            "epochs 242/1000\n",
            "train_loss 89.0549545288086\n",
            "train_acc 0.015444510473081699\n",
            "valid_loss 212.53482055664062\n",
            "val_acc 0.10414514802396298\n",
            "epochs 243/1000\n",
            "train_loss 88.75304412841797\n",
            "train_acc 0.015342610458341928\n",
            "valid_loss 212.3607177734375\n",
            "val_acc 0.10401499532163143\n",
            "epochs 244/1000\n",
            "train_loss 88.45114135742188\n",
            "train_acc 0.015241056322478332\n",
            "valid_loss 212.18663024902344\n",
            "val_acc 0.10388498492538929\n",
            "epochs 245/1000\n",
            "train_loss 88.14923095703125\n",
            "train_acc 0.015139848065490905\n",
            "valid_loss 212.01255798339844\n",
            "val_acc 0.10375515334308147\n",
            "epochs 246/1000\n",
            "train_loss 87.84732818603516\n",
            "train_acc 0.015038986045580644\n",
            "valid_loss 211.83848571777344\n",
            "val_acc 0.10362547505646943\n",
            "epochs 247/1000\n",
            "train_loss 87.54541015625\n",
            "train_acc 0.014938469403065167\n",
            "valid_loss 211.66441345214844\n",
            "val_acc 0.10349596627056598\n",
            "epochs 248/1000\n",
            "train_loss 87.24352264404297\n",
            "train_acc 0.014838299714028835\n",
            "valid_loss 211.4903106689453\n",
            "val_acc 0.10336662344634533\n",
            "epochs 249/1000\n",
            "train_loss 86.94163513183594\n",
            "train_acc 0.014738476548630457\n",
            "valid_loss 211.31626892089844\n",
            "val_acc 0.10323743708431721\n",
            "epochs 250/1000\n",
            "train_loss 86.63973236083984\n",
            "train_acc 0.014638998545706272\n",
            "valid_loss 211.1422119140625\n",
            "val_acc 0.10310842543840408\n",
            "epochs 251/1000\n",
            "train_loss 86.33783721923828\n",
            "train_acc 0.014539867639541626\n",
            "valid_loss 210.96817016601562\n",
            "val_acc 0.10297956727445126\n",
            "epochs 252/1000\n",
            "train_loss 86.03595733642578\n",
            "train_acc 0.014441081680930577\n",
            "valid_loss 210.79409790039062\n",
            "val_acc 0.1028508685529232\n",
            "epochs 253/1000\n",
            "train_loss 85.73408508300781\n",
            "train_acc 0.014342643033999663\n",
            "valid_loss 210.6200408935547\n",
            "val_acc 0.10272234268486499\n",
            "epochs 254/1000\n",
            "train_loss 85.43221282958984\n",
            "train_acc 0.014244551340547891\n",
            "valid_loss 210.4459991455078\n",
            "val_acc 0.1025939654558897\n",
            "epochs 255/1000\n",
            "train_loss 85.1303482055664\n",
            "train_acc 0.01414680566925269\n",
            "valid_loss 210.27197265625\n",
            "val_acc 0.10246577039361\n",
            "epochs 256/1000\n",
            "train_loss 84.82854461669922\n",
            "train_acc 0.014049416551223168\n",
            "valid_loss 210.0979461669922\n",
            "val_acc 0.10233775228261947\n",
            "epochs 257/1000\n",
            "train_loss 84.52684783935547\n",
            "train_acc 0.013952386708786855\n",
            "valid_loss 209.92393493652344\n",
            "val_acc 0.10220992900431156\n",
            "epochs 258/1000\n",
            "train_loss 84.2251968383789\n",
            "train_acc 0.013855707688400378\n",
            "valid_loss 209.7499542236328\n",
            "val_acc 0.10208228472620248\n",
            "epochs 259/1000\n",
            "train_loss 83.92353820800781\n",
            "train_acc 0.01375937626625483\n",
            "valid_loss 209.5759735107422\n",
            "val_acc 0.10195479784160852\n",
            "epochs 260/1000\n",
            "train_loss 83.62188720703125\n",
            "train_acc 0.013663391009546243\n",
            "valid_loss 209.4019775390625\n",
            "val_acc 0.1018274761736393\n",
            "epochs 261/1000\n",
            "train_loss 83.32022857666016\n",
            "train_acc 0.013567751918274622\n",
            "valid_loss 209.22802734375\n",
            "val_acc 0.10170031897723675\n",
            "epochs 262/1000\n",
            "train_loss 83.01859283447266\n",
            "train_acc 0.013472458562598778\n",
            "valid_loss 209.0540771484375\n",
            "val_acc 0.10157333444803954\n",
            "epochs 263/1000\n",
            "train_loss 82.71694946289062\n",
            "train_acc 0.013377511587280493\n",
            "valid_loss 208.880126953125\n",
            "val_acc 0.10144650526344776\n",
            "epochs 264/1000\n",
            "train_loss 82.41531372070312\n",
            "train_acc 0.01328291013263739\n",
            "valid_loss 208.7062530517578\n",
            "val_acc 0.10131985154002905\n",
            "epochs 265/1000\n",
            "train_loss 82.11367797851562\n",
            "train_acc 0.013188655058351846\n",
            "valid_loss 208.53237915039062\n",
            "val_acc 0.1011933546513319\n",
            "epochs 266/1000\n",
            "train_loss 81.81204986572266\n",
            "train_acc 0.013094745862942476\n",
            "valid_loss 208.3585205078125\n",
            "val_acc 0.10106702186167241\n",
            "epochs 267/1000\n",
            "train_loss 81.51042938232422\n",
            "train_acc 0.013001182474769078\n",
            "valid_loss 208.1846923828125\n",
            "val_acc 0.1009408300742507\n",
            "epochs 268/1000\n",
            "train_loss 81.20880126953125\n",
            "train_acc 0.012907964607270865\n",
            "valid_loss 208.0109405517578\n",
            "val_acc 0.10081483032554388\n",
            "epochs 269/1000\n",
            "train_loss 80.90716552734375\n",
            "train_acc 0.012815091651506148\n",
            "valid_loss 207.8373260498047\n",
            "val_acc 0.10068897064775229\n",
            "epochs 270/1000\n",
            "train_loss 80.60552978515625\n",
            "train_acc 0.012722564252236715\n",
            "valid_loss 207.66380310058594\n",
            "val_acc 0.10056327618658542\n",
            "epochs 271/1000\n",
            "train_loss 80.30390167236328\n",
            "train_acc 0.012630381800520878\n",
            "valid_loss 207.4903106689453\n",
            "val_acc 0.10043774545192719\n",
            "epochs 272/1000\n",
            "train_loss 80.00225830078125\n",
            "train_acc 0.012538545156041017\n",
            "valid_loss 207.3168182373047\n",
            "val_acc 0.10031237564980984\n",
            "epochs 273/1000\n",
            "train_loss 79.70063018798828\n",
            "train_acc 0.012447053674035348\n",
            "valid_loss 207.1433868408203\n",
            "val_acc 0.10018717125058174\n",
            "epochs 274/1000\n",
            "train_loss 79.39899444580078\n",
            "train_acc 0.012355908393286742\n",
            "valid_loss 206.9700469970703\n",
            "val_acc 0.10006212145090103\n",
            "epochs 275/1000\n",
            "train_loss 79.09737396240234\n",
            "train_acc 0.012265108597393218\n",
            "valid_loss 206.796875\n",
            "val_acc 0.09993723221123219\n",
            "epochs 276/1000\n",
            "train_loss 78.79573822021484\n",
            "train_acc 0.012174654178894483\n",
            "valid_loss 206.6237030029297\n",
            "val_acc 0.09981250874698162\n",
            "epochs 277/1000\n",
            "train_loss 78.4941177368164\n",
            "train_acc 0.012084545854192514\n",
            "valid_loss 206.45059204101562\n",
            "val_acc 0.09968794398009777\n",
            "epochs 278/1000\n",
            "train_loss 78.19248962402344\n",
            "train_acc 0.011994783265086321\n",
            "valid_loss 206.27749633789062\n",
            "val_acc 0.09956354405730963\n",
            "epochs 279/1000\n",
            "train_loss 77.890869140625\n",
            "train_acc 0.011905366339935707\n",
            "valid_loss 206.10440063476562\n",
            "val_acc 0.09943930897861719\n",
            "epochs 280/1000\n",
            "train_loss 77.58924102783203\n",
            "train_acc 0.011816294648899483\n",
            "valid_loss 205.9313201904297\n",
            "val_acc 0.09931522831320763\n",
            "epochs 281/1000\n",
            "train_loss 77.2876205444336\n",
            "train_acc 0.011727568621818837\n",
            "valid_loss 205.7582550048828\n",
            "val_acc 0.09919130783528089\n",
            "epochs 282/1000\n",
            "train_loss 76.9859848022461\n",
            "train_acc 0.01163918818705357\n",
            "valid_loss 205.585205078125\n",
            "val_acc 0.09906756468117237\n",
            "epochs 283/1000\n",
            "train_loss 76.68437194824219\n",
            "train_acc 0.011551153595344378\n",
            "valid_loss 205.4121551513672\n",
            "val_acc 0.09894395601004362\n",
            "epochs 284/1000\n",
            "train_loss 76.38275146484375\n",
            "train_acc 0.011463464775051061\n",
            "valid_loss 205.2391815185547\n",
            "val_acc 0.09882054068148136\n",
            "epochs 285/1000\n",
            "train_loss 76.08113098144531\n",
            "train_acc 0.011376121833633918\n",
            "valid_loss 205.0662384033203\n",
            "val_acc 0.09869726374745369\n",
            "epochs 286/1000\n",
            "train_loss 75.77950286865234\n",
            "train_acc 0.011289124018870868\n",
            "valid_loss 204.89341735839844\n",
            "val_acc 0.09857417345046997\n",
            "epochs 287/1000\n",
            "train_loss 75.47789001464844\n",
            "train_acc 0.011202472548645277\n",
            "valid_loss 204.72056579589844\n",
            "val_acc 0.09845121931284666\n",
            "epochs 288/1000\n",
            "train_loss 75.17626953125\n",
            "train_acc 0.011116166419994373\n",
            "valid_loss 204.5477752685547\n",
            "val_acc 0.09832843728363513\n",
            "epochs 289/1000\n",
            "train_loss 74.8746566772461\n",
            "train_acc 0.011030205812018651\n",
            "valid_loss 204.37498474121094\n",
            "val_acc 0.09820581786334515\n",
            "epochs 290/1000\n",
            "train_loss 74.57303619384766\n",
            "train_acc 0.0109445911903794\n",
            "valid_loss 204.2021484375\n",
            "val_acc 0.0980833474546671\n",
            "epochs 291/1000\n",
            "train_loss 74.27142333984375\n",
            "train_acc 0.010859322017775131\n",
            "valid_loss 204.02940368652344\n",
            "val_acc 0.09796105939894914\n",
            "epochs 292/1000\n",
            "train_loss 73.96981048583984\n",
            "train_acc 0.010774399118068127\n",
            "valid_loss 203.85659790039062\n",
            "val_acc 0.09783891756087541\n",
            "epochs 293/1000\n",
            "train_loss 73.6681900024414\n",
            "train_acc 0.010689821631576005\n",
            "valid_loss 203.683837890625\n",
            "val_acc 0.09771694056689739\n",
            "epochs 294/1000\n",
            "train_loss 73.36659240722656\n",
            "train_acc 0.010605590095600257\n",
            "valid_loss 203.51109313964844\n",
            "val_acc 0.09759513102471828\n",
            "epochs 295/1000\n",
            "train_loss 73.06498718261719\n",
            "train_acc 0.010521704223580085\n",
            "valid_loss 203.33831787109375\n",
            "val_acc 0.09747346565127372\n",
            "epochs 296/1000\n",
            "train_loss 72.76337432861328\n",
            "train_acc 0.01043816408715569\n",
            "valid_loss 203.1655731201172\n",
            "val_acc 0.09735198635607958\n",
            "epochs 297/1000\n",
            "train_loss 72.46177673339844\n",
            "train_acc 0.010354969686327072\n",
            "valid_loss 202.99282836914062\n",
            "val_acc 0.0972306551411748\n",
            "epochs 298/1000\n",
            "train_loss 72.16017150878906\n",
            "train_acc 0.010272121164374627\n",
            "valid_loss 202.82009887695312\n",
            "val_acc 0.09710949640721082\n",
            "epochs 299/1000\n",
            "train_loss 71.85857391357422\n",
            "train_acc 0.010189617769076275\n",
            "valid_loss 202.6473388671875\n",
            "val_acc 0.09698849059641361\n",
            "epochs 300/1000\n",
            "train_loss 71.55696868896484\n",
            "train_acc 0.010107459930273203\n",
            "valid_loss 202.47462463378906\n",
            "val_acc 0.09686764478683471\n",
            "epochs 301/1000\n",
            "train_loss 71.25537109375\n",
            "train_acc 0.010025648292727195\n",
            "valid_loss 202.30189514160156\n",
            "val_acc 0.09674696810543537\n",
            "epochs 302/1000\n",
            "train_loss 70.95377349853516\n",
            "train_acc 0.009944181566914687\n",
            "valid_loss 202.12918090820312\n",
            "val_acc 0.09662644788622857\n",
            "epochs 303/1000\n",
            "train_loss 70.65217590332031\n",
            "train_acc 0.009863060612518054\n",
            "valid_loss 201.9564666748047\n",
            "val_acc 0.09650609213858843\n",
            "epochs 304/1000\n",
            "train_loss 70.35057830810547\n",
            "train_acc 0.00978228568027799\n",
            "valid_loss 201.78378295898438\n",
            "val_acc 0.09638589732348919\n",
            "epochs 305/1000\n",
            "train_loss 70.04898834228516\n",
            "train_acc 0.009701856017972414\n",
            "valid_loss 201.61106872558594\n",
            "val_acc 0.09626586521044374\n",
            "epochs 306/1000\n",
            "train_loss 69.74739837646484\n",
            "train_acc 0.009621773130045487\n",
            "valid_loss 201.4384002685547\n",
            "val_acc 0.09614599738270044\n",
            "epochs 307/1000\n",
            "train_loss 69.4458236694336\n",
            "train_acc 0.009542036371735426\n",
            "valid_loss 201.26573181152344\n",
            "val_acc 0.0960262936539948\n",
            "epochs 308/1000\n",
            "train_loss 69.14424133300781\n",
            "train_acc 0.009462645241560845\n",
            "valid_loss 201.09304809570312\n",
            "val_acc 0.09590674974024296\n",
            "epochs 309/1000\n",
            "train_loss 68.8426742553711\n",
            "train_acc 0.009383602318568872\n",
            "valid_loss 200.92039489746094\n",
            "val_acc 0.09578738156706094\n",
            "epochs 310/1000\n",
            "train_loss 68.54112243652344\n",
            "train_acc 0.009304906169955548\n",
            "valid_loss 200.74777221679688\n",
            "val_acc 0.09566817013546824\n",
            "epochs 311/1000\n",
            "train_loss 68.23956298828125\n",
            "train_acc 0.009226555900218395\n",
            "valid_loss 200.5751953125\n",
            "val_acc 0.09554911777377129\n",
            "epochs 312/1000\n",
            "train_loss 67.93802642822266\n",
            "train_acc 0.009148553407822665\n",
            "valid_loss 200.4026336669922\n",
            "val_acc 0.0954302353784442\n",
            "epochs 313/1000\n",
            "train_loss 67.6364974975586\n",
            "train_acc 0.009070896830123205\n",
            "valid_loss 200.23007202148438\n",
            "val_acc 0.09531151158735156\n",
            "epochs 314/1000\n",
            "train_loss 67.3349838256836\n",
            "train_acc 0.008993587170082789\n",
            "valid_loss 200.0575408935547\n",
            "val_acc 0.09519294817000627\n",
            "epochs 315/1000\n",
            "train_loss 67.03349304199219\n",
            "train_acc 0.008916626003785776\n",
            "valid_loss 199.88510131835938\n",
            "val_acc 0.09507457222789525\n",
            "epochs 316/1000\n",
            "train_loss 66.73208618164062\n",
            "train_acc 0.008840016232660184\n",
            "valid_loss 199.7127227783203\n",
            "val_acc 0.09495638627558947\n",
            "epochs 317/1000\n",
            "train_loss 66.43074035644531\n",
            "train_acc 0.008763761581996312\n",
            "valid_loss 199.5403594970703\n",
            "val_acc 0.09483839087188244\n",
            "epochs 318/1000\n",
            "train_loss 66.12942504882812\n",
            "train_acc 0.008687855102694951\n",
            "valid_loss 199.3680419921875\n",
            "val_acc 0.094720552675426\n",
            "epochs 319/1000\n",
            "train_loss 65.828125\n",
            "train_acc 0.008612297618618378\n",
            "valid_loss 199.1957244873047\n",
            "val_acc 0.094602888263762\n",
            "epochs 320/1000\n",
            "train_loss 65.52684020996094\n",
            "train_acc 0.008537085655216988\n",
            "valid_loss 199.02340698242188\n",
            "val_acc 0.09448538888245821\n",
            "epochs 321/1000\n",
            "train_loss 65.22554016113281\n",
            "train_acc 0.008462218961750086\n",
            "valid_loss 198.85113525390625\n",
            "val_acc 0.09436805211007596\n",
            "epochs 322/1000\n",
            "train_loss 64.92425537109375\n",
            "train_acc 0.008387698254619654\n",
            "valid_loss 198.67886352539062\n",
            "val_acc 0.09425087943673134\n",
            "epochs 323/1000\n",
            "train_loss 64.62297821044922\n",
            "train_acc 0.008313523354725195\n",
            "valid_loss 198.506591796875\n",
            "val_acc 0.09413386778905988\n",
            "epochs 324/1000\n",
            "train_loss 64.32170867919922\n",
            "train_acc 0.008239694620267702\n",
            "valid_loss 198.33433532714844\n",
            "val_acc 0.09401702862232923\n",
            "epochs 325/1000\n",
            "train_loss 64.02046203613281\n",
            "train_acc 0.008166213125850145\n",
            "valid_loss 198.16209411621094\n",
            "val_acc 0.09390035457909107\n",
            "epochs 326/1000\n",
            "train_loss 63.71921157836914\n",
            "train_acc 0.00809307829835094\n",
            "valid_loss 197.98985290527344\n",
            "val_acc 0.09378384407609701\n",
            "epochs 327/1000\n",
            "train_loss 63.4179801940918\n",
            "train_acc 0.008020289313907806\n",
            "valid_loss 197.81765747070312\n",
            "val_acc 0.09366749860346317\n",
            "epochs 328/1000\n",
            "train_loss 63.116756439208984\n",
            "train_acc 0.007947845778499659\n",
            "valid_loss 197.6454620361328\n",
            "val_acc 0.09355130139738321\n",
            "epochs 329/1000\n",
            "train_loss 62.81556701660156\n",
            "train_acc 0.007875751095035901\n",
            "valid_loss 197.4733123779297\n",
            "val_acc 0.09343529092147947\n",
            "epochs 330/1000\n",
            "train_loss 62.514461517333984\n",
            "train_acc 0.007804012140975549\n",
            "valid_loss 197.3012237548828\n",
            "val_acc 0.09331946447491646\n",
            "epochs 331/1000\n",
            "train_loss 62.213436126708984\n",
            "train_acc 0.007732627626795035\n",
            "valid_loss 197.1291961669922\n",
            "val_acc 0.09320380287244916\n",
            "epochs 332/1000\n",
            "train_loss 61.91246032714844\n",
            "train_acc 0.007661594830166835\n",
            "valid_loss 196.95718383789062\n",
            "val_acc 0.09308833125978708\n",
            "epochs 333/1000\n",
            "train_loss 61.611488342285156\n",
            "train_acc 0.007590908736277085\n",
            "valid_loss 196.7852020263672\n",
            "val_acc 0.09297301825135947\n",
            "epochs 334/1000\n",
            "train_loss 61.31053161621094\n",
            "train_acc 0.007520568306342914\n",
            "valid_loss 196.61322021484375\n",
            "val_acc 0.09285786589607596\n",
            "epochs 335/1000\n",
            "train_loss 61.00957107543945\n",
            "train_acc 0.0074505727165020425\n",
            "valid_loss 196.4412384033203\n",
            "val_acc 0.09274289011955261\n",
            "epochs 336/1000\n",
            "train_loss 60.708614349365234\n",
            "train_acc 0.007380922289135365\n",
            "valid_loss 196.26925659179688\n",
            "val_acc 0.09262806130573153\n",
            "epochs 337/1000\n",
            "train_loss 60.407657623291016\n",
            "train_acc 0.007311616845142383\n",
            "valid_loss 196.0972900390625\n",
            "val_acc 0.09251339677721263\n",
            "epochs 338/1000\n",
            "train_loss 60.10671615600586\n",
            "train_acc 0.007242656886004484\n",
            "valid_loss 195.9253387451172\n",
            "val_acc 0.09239890500903129\n",
            "epochs 339/1000\n",
            "train_loss 59.805789947509766\n",
            "train_acc 0.007174043378864343\n",
            "valid_loss 195.75338745117188\n",
            "val_acc 0.0922845745459199\n",
            "epochs 340/1000\n",
            "train_loss 59.50486755371094\n",
            "train_acc 0.007105776359542058\n",
            "valid_loss 195.58143615722656\n",
            "val_acc 0.09217040613293648\n",
            "epochs 341/1000\n",
            "train_loss 59.20394515991211\n",
            "train_acc 0.007037853535551291\n",
            "valid_loss 195.40948486328125\n",
            "val_acc 0.0920563898049295\n",
            "epochs 342/1000\n",
            "train_loss 58.90302658081055\n",
            "train_acc 0.006970275874034717\n",
            "valid_loss 195.237548828125\n",
            "val_acc 0.09194255089387297\n",
            "epochs 343/1000\n",
            "train_loss 58.602108001708984\n",
            "train_acc 0.0069030428735109475\n",
            "valid_loss 195.0655975341797\n",
            "val_acc 0.09182887468487025\n",
            "epochs 344/1000\n",
            "train_loss 58.30120086669922\n",
            "train_acc 0.00683615467726038\n",
            "valid_loss 194.89366149902344\n",
            "val_acc 0.0917153438553214\n",
            "epochs 345/1000\n",
            "train_loss 58.000274658203125\n",
            "train_acc 0.006769608562955489\n",
            "valid_loss 194.7217254638672\n",
            "val_acc 0.09160200133919716\n",
            "epochs 346/1000\n",
            "train_loss 57.69935989379883\n",
            "train_acc 0.006703408470807167\n",
            "valid_loss 194.5498046875\n",
            "val_acc 0.09148879330605268\n",
            "epochs 347/1000\n",
            "train_loss 57.39845275878906\n",
            "train_acc 0.006637553666503384\n",
            "valid_loss 194.3778839111328\n",
            "val_acc 0.09137575170025229\n",
            "epochs 348/1000\n",
            "train_loss 57.09754943847656\n",
            "train_acc 0.006572045153006911\n",
            "valid_loss 194.20594787597656\n",
            "val_acc 0.09126286990940571\n",
            "epochs 349/1000\n",
            "train_loss 56.796661376953125\n",
            "train_acc 0.006506882643756958\n",
            "valid_loss 194.0340118408203\n",
            "val_acc 0.09115016236901283\n",
            "epochs 350/1000\n",
            "train_loss 56.495784759521484\n",
            "train_acc 0.0064420655298118405\n",
            "valid_loss 193.86207580566406\n",
            "val_acc 0.09103760169818997\n",
            "epochs 351/1000\n",
            "train_loss 56.19490432739258\n",
            "train_acc 0.006377593829081609\n",
            "valid_loss 193.69017028808594\n",
            "val_acc 0.09092520549893379\n",
            "epochs 352/1000\n",
            "train_loss 55.89403533935547\n",
            "train_acc 0.006313467487836113\n",
            "valid_loss 193.5182342529297\n",
            "val_acc 0.09081296576187015\n",
            "epochs 353/1000\n",
            "train_loss 55.593170166015625\n",
            "train_acc 0.006249685753853275\n",
            "valid_loss 193.34629821777344\n",
            "val_acc 0.09070088630542159\n",
            "epochs 354/1000\n",
            "train_loss 55.292301177978516\n",
            "train_acc 0.0061862488062335895\n",
            "valid_loss 193.1743927001953\n",
            "val_acc 0.09058896638453007\n",
            "epochs 355/1000\n",
            "train_loss 54.99144744873047\n",
            "train_acc 0.006123157522569482\n",
            "valid_loss 193.00245666503906\n",
            "val_acc 0.0904772212728858\n",
            "epochs 356/1000\n",
            "train_loss 54.69060516357422\n",
            "train_acc 0.006060413192384518\n",
            "valid_loss 192.83055114746094\n",
            "val_acc 0.09036563839763403\n",
            "epochs 357/1000\n",
            "train_loss 54.3897819519043\n",
            "train_acc 0.005998014579885281\n",
            "valid_loss 192.6586456298828\n",
            "val_acc 0.09025420900434256\n",
            "epochs 358/1000\n",
            "train_loss 54.088958740234375\n",
            "train_acc 0.005935961810442118\n",
            "valid_loss 192.48672485351562\n",
            "val_acc 0.09014294669032097\n",
            "epochs 359/1000\n",
            "train_loss 53.78816604614258\n",
            "train_acc 0.005874255421356513\n",
            "valid_loss 192.3148651123047\n",
            "val_acc 0.0900318706408143\n",
            "epochs 360/1000\n",
            "train_loss 53.487369537353516\n",
            "train_acc 0.00581289457085614\n",
            "valid_loss 192.1429901123047\n",
            "val_acc 0.08992093428969383\n",
            "epochs 361/1000\n",
            "train_loss 53.18659210205078\n",
            "train_acc 0.005751879115660603\n",
            "valid_loss 191.97109985351562\n",
            "val_acc 0.08981016594916583\n",
            "epochs 362/1000\n",
            "train_loss 52.885833740234375\n",
            "train_acc 0.0056912100050025265\n",
            "valid_loss 191.79922485351562\n",
            "val_acc 0.08969956375658512\n",
            "epochs 363/1000\n",
            "train_loss 52.5850944519043\n",
            "train_acc 0.00563088829557483\n",
            "valid_loss 191.62738037109375\n",
            "val_acc 0.08958913460373878\n",
            "epochs 364/1000\n",
            "train_loss 52.28437423706055\n",
            "train_acc 0.005570912751584099\n",
            "valid_loss 191.4555206298828\n",
            "val_acc 0.08947888119146227\n",
            "epochs 365/1000\n",
            "train_loss 51.98369216918945\n",
            "train_acc 0.005511284787924244\n",
            "valid_loss 191.28369140625\n",
            "val_acc 0.08936878573149443\n",
            "epochs 366/1000\n",
            "train_loss 51.68305206298828\n",
            "train_acc 0.005452003938933978\n",
            "valid_loss 191.1118621826172\n",
            "val_acc 0.08925886880606412\n",
            "epochs 367/1000\n",
            "train_loss 51.38249206542969\n",
            "train_acc 0.005393074323924689\n",
            "valid_loss 190.9401092529297\n",
            "val_acc 0.08914914969354867\n",
            "epochs 368/1000\n",
            "train_loss 51.082061767578125\n",
            "train_acc 0.005334500868159991\n",
            "valid_loss 190.76837158203125\n",
            "val_acc 0.08903963025659323\n",
            "epochs 369/1000\n",
            "train_loss 50.781768798828125\n",
            "train_acc 0.005276282479126866\n",
            "valid_loss 190.5966796875\n",
            "val_acc 0.08893030239269137\n",
            "epochs 370/1000\n",
            "train_loss 50.4815673828125\n",
            "train_acc 0.005218417527010808\n",
            "valid_loss 190.42503356933594\n",
            "val_acc 0.08882114905864\n",
            "epochs 371/1000\n",
            "train_loss 50.18141555786133\n",
            "train_acc 0.005160900764167309\n",
            "valid_loss 190.2534637451172\n",
            "val_acc 0.08871217593550682\n",
            "epochs 372/1000\n",
            "train_loss 49.88128662109375\n",
            "train_acc 0.005103729683189438\n",
            "valid_loss 190.0819091796875\n",
            "val_acc 0.08860336020588874\n",
            "epochs 373/1000\n",
            "train_loss 49.581214904785156\n",
            "train_acc 0.005046905484050512\n",
            "valid_loss 189.91050720214844\n",
            "val_acc 0.08849472142755985\n",
            "epochs 374/1000\n",
            "train_loss 49.28117370605469\n",
            "train_acc 0.004990429760744939\n",
            "valid_loss 189.73919677734375\n",
            "val_acc 0.08838626192882656\n",
            "epochs 375/1000\n",
            "train_loss 48.981231689453125\n",
            "train_acc 0.004934307796737323\n",
            "valid_loss 189.5679168701172\n",
            "val_acc 0.08827799847349524\n",
            "epochs 376/1000\n",
            "train_loss 48.681453704833984\n",
            "train_acc 0.0048785417770537045\n",
            "valid_loss 189.39669799804688\n",
            "val_acc 0.08816995024681092\n",
            "epochs 377/1000\n",
            "train_loss 48.38195037841797\n",
            "train_acc 0.004823143719337308\n",
            "valid_loss 189.22555541992188\n",
            "val_acc 0.08806214574724436\n",
            "epochs 378/1000\n",
            "train_loss 48.08263397216797\n",
            "train_acc 0.004768104292452335\n",
            "valid_loss 189.054443359375\n",
            "val_acc 0.08795454110950232\n",
            "epochs 379/1000\n",
            "train_loss 47.783424377441406\n",
            "train_acc 0.004713417496532202\n",
            "valid_loss 188.88339233398438\n",
            "val_acc 0.08784713167697192\n",
            "epochs 380/1000\n",
            "train_loss 47.48435974121094\n",
            "train_acc 0.004659085856893888\n",
            "valid_loss 188.7123565673828\n",
            "val_acc 0.08773990906774998\n",
            "epochs 381/1000\n",
            "train_loss 47.18538284301758\n",
            "train_acc 0.004605106633299818\n",
            "valid_loss 188.5413360595703\n",
            "val_acc 0.08763286266475916\n",
            "epochs 382/1000\n",
            "train_loss 46.886444091796875\n",
            "train_acc 0.004551473109481426\n",
            "valid_loss 188.37034606933594\n",
            "val_acc 0.08752599898725748\n",
            "epochs 383/1000\n",
            "train_loss 46.5875244140625\n",
            "train_acc 0.004498185070518117\n",
            "valid_loss 188.19935607910156\n",
            "val_acc 0.08741930276155471\n",
            "epochs 384/1000\n",
            "train_loss 46.28864669799805\n",
            "train_acc 0.004445240635854693\n",
            "valid_loss 188.02838134765625\n",
            "val_acc 0.08731277324259282\n",
            "epochs 385/1000\n",
            "train_loss 45.98978042602539\n",
            "train_acc 0.00439263959057056\n",
            "valid_loss 187.85739135742188\n",
            "val_acc 0.08720640949904919\n",
            "epochs 386/1000\n",
            "train_loss 45.69094467163086\n",
            "train_acc 0.004340382865988291\n",
            "valid_loss 187.68643188476562\n",
            "val_acc 0.08710021637380123\n",
            "epochs 387/1000\n",
            "train_loss 45.39212417602539\n",
            "train_acc 0.004288469262134571\n",
            "valid_loss 187.51548767089844\n",
            "val_acc 0.08699419116601348\n",
            "epochs 388/1000\n",
            "train_loss 45.09332275390625\n",
            "train_acc 0.004236898725279248\n",
            "valid_loss 187.34451293945312\n",
            "val_acc 0.08688831557519734\n",
            "epochs 389/1000\n",
            "train_loss 44.79454803466797\n",
            "train_acc 0.004185672527035842\n",
            "valid_loss 187.17359924316406\n",
            "val_acc 0.08678262531757355\n",
            "epochs 390/1000\n",
            "train_loss 44.49580764770508\n",
            "train_acc 0.004134791473356577\n",
            "valid_loss 187.00265502929688\n",
            "val_acc 0.08667708686552941\n",
            "epochs 391/1000\n",
            "train_loss 44.19709396362305\n",
            "train_acc 0.0040842552239505146\n",
            "valid_loss 186.83172607421875\n",
            "val_acc 0.08657171167433261\n",
            "epochs 392/1000\n",
            "train_loss 43.898475646972656\n",
            "train_acc 0.0040340689906420615\n",
            "valid_loss 186.66082763671875\n",
            "val_acc 0.08646654086187482\n",
            "epochs 393/1000\n",
            "train_loss 43.59999084472656\n",
            "train_acc 0.003984237089753151\n",
            "valid_loss 186.49000549316406\n",
            "val_acc 0.08636156837455929\n",
            "epochs 394/1000\n",
            "train_loss 43.30158615112305\n",
            "train_acc 0.0039347529841157105\n",
            "valid_loss 186.31919860839844\n",
            "val_acc 0.08625677935779094\n",
            "epochs 395/1000\n",
            "train_loss 43.003231048583984\n",
            "train_acc 0.0038856146678041955\n",
            "valid_loss 186.14837646484375\n",
            "val_acc 0.08615215411409735\n",
            "epochs 396/1000\n",
            "train_loss 42.70490646362305\n",
            "train_acc 0.0038368191140202377\n",
            "valid_loss 185.97755432128906\n",
            "val_acc 0.0860476822592318\n",
            "epochs 397/1000\n",
            "train_loss 42.406612396240234\n",
            "train_acc 0.0037883675316921794\n",
            "valid_loss 185.8067626953125\n",
            "val_acc 0.08594339941628278\n",
            "epochs 398/1000\n",
            "train_loss 42.10841751098633\n",
            "train_acc 0.0037402618640603926\n",
            "valid_loss 185.63601684570312\n",
            "val_acc 0.0858393028844148\n",
            "epochs 399/1000\n",
            "train_loss 41.81035232543945\n",
            "train_acc 0.0036925082542718602\n",
            "valid_loss 185.46530151367188\n",
            "val_acc 0.08573539201170206\n",
            "epochs 400/1000\n",
            "train_loss 41.51238250732422\n",
            "train_acc 0.0036451048396814326\n",
            "valid_loss 185.29458618164062\n",
            "val_acc 0.0856316703837365\n",
            "epochs 401/1000\n",
            "train_loss 41.214534759521484\n",
            "train_acc 0.003598051593424036\n",
            "valid_loss 185.12393188476562\n",
            "val_acc 0.08552815280854702\n",
            "epochs 402/1000\n",
            "train_loss 40.916839599609375\n",
            "train_acc 0.003551351748263607\n",
            "valid_loss 184.95330810546875\n",
            "val_acc 0.08542483970522881\n",
            "epochs 403/1000\n",
            "train_loss 40.61930465698242\n",
            "train_acc 0.0035050043549675206\n",
            "valid_loss 184.78274536132812\n",
            "val_acc 0.08532172869890928\n",
            "epochs 404/1000\n",
            "train_loss 40.32189178466797\n",
            "train_acc 0.0034590055986952325\n",
            "valid_loss 184.61219787597656\n",
            "val_acc 0.08521880572661758\n",
            "epochs 405/1000\n",
            "train_loss 40.02456283569336\n",
            "train_acc 0.0034133544048437704\n",
            "valid_loss 184.4416961669922\n",
            "val_acc 0.08511605504900217\n",
            "epochs 406/1000\n",
            "train_loss 39.727352142333984\n",
            "train_acc 0.003368053791256478\n",
            "valid_loss 184.2711944580078\n",
            "val_acc 0.08501348672434687\n",
            "epochs 407/1000\n",
            "train_loss 39.43033981323242\n",
            "train_acc 0.0033231071966628614\n",
            "valid_loss 184.1007843017578\n",
            "val_acc 0.08491116371005773\n",
            "epochs 408/1000\n",
            "train_loss 39.1335334777832\n",
            "train_acc 0.0032785141106265094\n",
            "valid_loss 183.93038940429688\n",
            "val_acc 0.08480900144204498\n",
            "epochs 409/1000\n",
            "train_loss 38.836856842041016\n",
            "train_acc 0.003234271560079203\n",
            "valid_loss 183.76002502441406\n",
            "val_acc 0.08470701826736331\n",
            "epochs 410/1000\n",
            "train_loss 38.54035186767578\n",
            "train_acc 0.003190380404703319\n",
            "valid_loss 183.58970642089844\n",
            "val_acc 0.0846052368171513\n",
            "epochs 411/1000\n",
            "train_loss 38.243953704833984\n",
            "train_acc 0.003146836758018113\n",
            "valid_loss 183.41940307617188\n",
            "val_acc 0.08450360768474638\n",
            "epochs 412/1000\n",
            "train_loss 37.94761657714844\n",
            "train_acc 0.003103635972365737\n",
            "valid_loss 183.24911499023438\n",
            "val_acc 0.08440216956660151\n",
            "epochs 413/1000\n",
            "train_loss 37.65134811401367\n",
            "train_acc 0.0030607806178382956\n",
            "valid_loss 183.07884216308594\n",
            "val_acc 0.08430090257897974\n",
            "epochs 414/1000\n",
            "train_loss 37.3552131652832\n",
            "train_acc 0.0030182734257183396\n",
            "valid_loss 182.90875244140625\n",
            "val_acc 0.08419984644278884\n",
            "epochs 415/1000\n",
            "train_loss 37.05923843383789\n",
            "train_acc 0.002976118022790895\n",
            "valid_loss 182.7387237548828\n",
            "val_acc 0.08409897862002254\n",
            "epochs 416/1000\n",
            "train_loss 36.76335525512695\n",
            "train_acc 0.002934307773382618\n",
            "valid_loss 182.5687713623047\n",
            "val_acc 0.08399830241687596\n",
            "epochs 417/1000\n",
            "train_loss 36.46753692626953\n",
            "train_acc 0.0028928402506817994\n",
            "valid_loss 182.39883422851562\n",
            "val_acc 0.08389779971912503\n",
            "epochs 418/1000\n",
            "train_loss 36.1717529296875\n",
            "train_acc 0.0028517136457734383\n",
            "valid_loss 182.2288818359375\n",
            "val_acc 0.08379745609126985\n",
            "epochs 419/1000\n",
            "train_loss 35.87601852416992\n",
            "train_acc 0.0028109276541866935\n",
            "valid_loss 182.0589599609375\n",
            "val_acc 0.08369727982208133\n",
            "epochs 420/1000\n",
            "train_loss 35.58033752441406\n",
            "train_acc 0.0027704847833284964\n",
            "valid_loss 181.88905334472656\n",
            "val_acc 0.08359725689515471\n",
            "epochs 421/1000\n",
            "train_loss 35.28473663330078\n",
            "train_acc 0.0027303853197596394\n",
            "valid_loss 181.7191619873047\n",
            "val_acc 0.08349741194397212\n",
            "epochs 422/1000\n",
            "train_loss 34.989219665527344\n",
            "train_acc 0.0026906296306361374\n",
            "valid_loss 181.54930114746094\n",
            "val_acc 0.08339774822816252\n",
            "epochs 423/1000\n",
            "train_loss 34.69382858276367\n",
            "train_acc 0.0026512206979812337\n",
            "valid_loss 181.3794708251953\n",
            "val_acc 0.08329824674874545\n",
            "epochs 424/1000\n",
            "train_loss 34.398555755615234\n",
            "train_acc 0.002612156945710572\n",
            "valid_loss 181.20970153808594\n",
            "val_acc 0.08319894662126899\n",
            "epochs 425/1000\n",
            "train_loss 34.10352325439453\n",
            "train_acc 0.002573445931865046\n",
            "valid_loss 181.03997802734375\n",
            "val_acc 0.08309987280517817\n",
            "epochs 426/1000\n",
            "train_loss 33.80879592895508\n",
            "train_acc 0.0025350915608354486\n",
            "valid_loss 180.870361328125\n",
            "val_acc 0.0830010381527245\n",
            "epochs 427/1000\n",
            "train_loss 33.51432800292969\n",
            "train_acc 0.002497090805823413\n",
            "valid_loss 180.70079040527344\n",
            "val_acc 0.08290240047499538\n",
            "epochs 428/1000\n",
            "train_loss 33.22003936767578\n",
            "train_acc 0.0024594390549911903\n",
            "valid_loss 180.53128051757812\n",
            "val_acc 0.08280396405607462\n",
            "epochs 429/1000\n",
            "train_loss 32.92589569091797\n",
            "train_acc 0.0024221347322544227\n",
            "valid_loss 180.3617706298828\n",
            "val_acc 0.08270569881424308\n",
            "epochs 430/1000\n",
            "train_loss 32.631935119628906\n",
            "train_acc 0.0023851786346103135\n",
            "valid_loss 180.19236755371094\n",
            "val_acc 0.08260764051228761\n",
            "epochs 431/1000\n",
            "train_loss 32.33883285522461\n",
            "train_acc 0.002348604012065782\n",
            "valid_loss 180.0231170654297\n",
            "val_acc 0.08250988624058664\n",
            "epochs 432/1000\n",
            "train_loss 32.04622268676758\n",
            "train_acc 0.002312394351555178\n",
            "valid_loss 179.85398864746094\n",
            "val_acc 0.08241236330941319\n",
            "epochs 433/1000\n",
            "train_loss 31.753887176513672\n",
            "train_acc 0.0022765403308977303\n",
            "valid_loss 179.68492126464844\n",
            "val_acc 0.08231502454727888\n",
            "epochs 434/1000\n",
            "train_loss 31.46182632446289\n",
            "train_acc 0.0022410393352262103\n",
            "valid_loss 179.51593017578125\n",
            "val_acc 0.08221788052469492\n",
            "epochs 435/1000\n",
            "train_loss 31.169939041137695\n",
            "train_acc 0.0022058832154680905\n",
            "valid_loss 179.3469696044922\n",
            "val_acc 0.08212091727182269\n",
            "epochs 436/1000\n",
            "train_loss 30.87837028503418\n",
            "train_acc 0.002171088995125431\n",
            "valid_loss 179.17822265625\n",
            "val_acc 0.08202424724586307\n",
            "epochs 437/1000\n",
            "train_loss 30.587541580200195\n",
            "train_acc 0.0021366564055474904\n",
            "valid_loss 179.00978088378906\n",
            "val_acc 0.08192783622071147\n",
            "epochs 438/1000\n",
            "train_loss 30.297082901000977\n",
            "train_acc 0.002102579227469575\n",
            "valid_loss 178.84153747558594\n",
            "val_acc 0.08183166743256151\n",
            "epochs 439/1000\n",
            "train_loss 30.00692367553711\n",
            "train_acc 0.0020688519356414103\n",
            "valid_loss 178.67335510253906\n",
            "val_acc 0.08173568206839263\n",
            "epochs 440/1000\n",
            "train_loss 29.716999053955078\n",
            "train_acc 0.002035467760064281\n",
            "valid_loss 178.5052032470703\n",
            "val_acc 0.08163988990709185\n",
            "epochs 441/1000\n",
            "train_loss 29.427486419677734\n",
            "train_acc 0.0020024411899682423\n",
            "valid_loss 178.33726501464844\n",
            "val_acc 0.08154438105411828\n",
            "epochs 442/1000\n",
            "train_loss 29.138484954833984\n",
            "train_acc 0.001969769064229555\n",
            "valid_loss 178.1693572998047\n",
            "val_acc 0.08144904840737581\n",
            "epochs 443/1000\n",
            "train_loss 28.85013771057129\n",
            "train_acc 0.0019374595095331853\n",
            "valid_loss 178.00160217285156\n",
            "val_acc 0.08135399930179119\n",
            "epochs 444/1000\n",
            "train_loss 28.562334060668945\n",
            "train_acc 0.0019055129154227101\n",
            "valid_loss 177.83396911621094\n",
            "val_acc 0.08125916437711567\n",
            "epochs 445/1000\n",
            "train_loss 28.275211334228516\n",
            "train_acc 0.0018739316952773011\n",
            "valid_loss 177.6665496826172\n",
            "val_acc 0.0811646577436477\n",
            "epochs 446/1000\n",
            "train_loss 27.98888397216797\n",
            "train_acc 0.0018427184550091624\n",
            "valid_loss 177.49929809570312\n",
            "val_acc 0.08107043774798513\n",
            "epochs 447/1000\n",
            "train_loss 27.703046798706055\n",
            "train_acc 0.0018118583516647609\n",
            "valid_loss 177.3321075439453\n",
            "val_acc 0.080976424459368\n",
            "epochs 448/1000\n",
            "train_loss 27.417545318603516\n",
            "train_acc 0.0017813428108078928\n",
            "valid_loss 177.16500854492188\n",
            "val_acc 0.08088261554948986\n",
            "epochs 449/1000\n",
            "train_loss 27.132631301879883\n",
            "train_acc 0.0017511830486070651\n",
            "valid_loss 176.99806213378906\n",
            "val_acc 0.08078906058799476\n",
            "epochs 450/1000\n",
            "train_loss 26.848237991333008\n",
            "train_acc 0.0017213768397386258\n",
            "valid_loss 176.8312225341797\n",
            "val_acc 0.08069572409149259\n",
            "epochs 451/1000\n",
            "train_loss 26.5645809173584\n",
            "train_acc 0.0016919319571640629\n",
            "valid_loss 176.66455078125\n",
            "val_acc 0.08060266594402492\n",
            "epochs 452/1000\n",
            "train_loss 26.281930923461914\n",
            "train_acc 0.00166286341398238\n",
            "valid_loss 176.49815368652344\n",
            "val_acc 0.08050998726394028\n",
            "epochs 453/1000\n",
            "train_loss 26.00029182434082\n",
            "train_acc 0.0016341634730521876\n",
            "valid_loss 176.3319549560547\n",
            "val_acc 0.08041755426675082\n",
            "epochs 454/1000\n",
            "train_loss 25.71904945373535\n",
            "train_acc 0.0016058047185651958\n",
            "valid_loss 176.1658477783203\n",
            "val_acc 0.08032532306388021\n",
            "epochs 455/1000\n",
            "train_loss 25.438119888305664\n",
            "train_acc 0.0015777837252244353\n",
            "valid_loss 175.99986267089844\n",
            "val_acc 0.08023332168813795\n",
            "epochs 456/1000\n",
            "train_loss 25.157732009887695\n",
            "train_acc 0.001550117194151076\n",
            "valid_loss 175.83409118652344\n",
            "val_acc 0.08014159086160362\n",
            "epochs 457/1000\n",
            "train_loss 24.8781795501709\n",
            "train_acc 0.0015228012612519355\n",
            "valid_loss 175.6685028076172\n",
            "val_acc 0.08005009940825403\n",
            "epochs 458/1000\n",
            "train_loss 24.59928321838379\n",
            "train_acc 0.0014958366563615317\n",
            "valid_loss 175.5030975341797\n",
            "val_acc 0.07995888944715261\n",
            "epochs 459/1000\n",
            "train_loss 24.32139778137207\n",
            "train_acc 0.0014692352627977156\n",
            "valid_loss 175.3380126953125\n",
            "val_acc 0.07986803038511425\n",
            "epochs 460/1000\n",
            "train_loss 24.044273376464844\n",
            "train_acc 0.0014429853942531806\n",
            "valid_loss 175.17308044433594\n",
            "val_acc 0.07977740338537842\n",
            "epochs 461/1000\n",
            "train_loss 23.7679500579834\n",
            "train_acc 0.0014170942953429543\n",
            "valid_loss 175.0084686279297\n",
            "val_acc 0.07968711375724524\n",
            "epochs 462/1000\n",
            "train_loss 23.4936580657959\n",
            "train_acc 0.001391601784584614\n",
            "valid_loss 174.84445190429688\n",
            "val_acc 0.0795973330270499\n",
            "epochs 463/1000\n",
            "train_loss 23.220840454101562\n",
            "train_acc 0.0013664858236622352\n",
            "valid_loss 174.6807098388672\n",
            "val_acc 0.07950784389395267\n",
            "epochs 464/1000\n",
            "train_loss 22.94873809814453\n",
            "train_acc 0.001341720989260536\n",
            "valid_loss 174.5172119140625\n",
            "val_acc 0.07941863513551653\n",
            "epochs 465/1000\n",
            "train_loss 22.677597045898438\n",
            "train_acc 0.0013173046441247256\n",
            "valid_loss 174.35406494140625\n",
            "val_acc 0.07932972903363407\n",
            "epochs 466/1000\n",
            "train_loss 22.407264709472656\n",
            "train_acc 0.001293232082389295\n",
            "valid_loss 174.19126892089844\n",
            "val_acc 0.0792411086615175\n",
            "epochs 467/1000\n",
            "train_loss 22.137662887573242\n",
            "train_acc 0.0012695071681474263\n",
            "valid_loss 174.0289764404297\n",
            "val_acc 0.07915278119035066\n",
            "epochs 468/1000\n",
            "train_loss 21.869380950927734\n",
            "train_acc 0.001246136196781523\n",
            "valid_loss 173.8673095703125\n",
            "val_acc 0.07906476266216486\n",
            "epochs 469/1000\n",
            "train_loss 21.602991104125977\n",
            "train_acc 0.0012231403961777687\n",
            "valid_loss 173.70706176757812\n",
            "val_acc 0.07897716923616827\n",
            "epochs 470/1000\n",
            "train_loss 21.33779525756836\n",
            "train_acc 0.0012005042607107987\n",
            "valid_loss 173.5484619140625\n",
            "val_acc 0.07888988261111081\n",
            "epochs 471/1000\n",
            "train_loss 21.07396125793457\n",
            "train_acc 0.001178223106902666\n",
            "valid_loss 173.39149475097656\n",
            "val_acc 0.0788029579911381\n",
            "epochs 472/1000\n",
            "train_loss 20.811092376708984\n",
            "train_acc 0.00115628793495349\n",
            "valid_loss 173.23504638671875\n",
            "val_acc 0.07871629986912013\n",
            "epochs 473/1000\n",
            "train_loss 20.54926109313965\n",
            "train_acc 0.0011346925703737026\n",
            "valid_loss 173.0792236328125\n",
            "val_acc 0.07862996598705649\n",
            "epochs 474/1000\n",
            "train_loss 20.2884521484375\n",
            "train_acc 0.0011134397063869983\n",
            "valid_loss 172.9240264892578\n",
            "val_acc 0.07854391778819263\n",
            "epochs 475/1000\n",
            "train_loss 20.028827667236328\n",
            "train_acc 0.0010925324615807487\n",
            "valid_loss 172.76954650878906\n",
            "val_acc 0.07845820351503789\n",
            "epochs 476/1000\n",
            "train_loss 19.77132797241211\n",
            "train_acc 0.0010719957824151677\n",
            "valid_loss 172.61595153808594\n",
            "val_acc 0.07837297781370581\n",
            "epochs 477/1000\n",
            "train_loss 19.516090393066406\n",
            "train_acc 0.001051834668032825\n",
            "valid_loss 172.46408081054688\n",
            "val_acc 0.07828822117298842\n",
            "epochs 478/1000\n",
            "train_loss 19.262863159179688\n",
            "train_acc 0.0010320538780293786\n",
            "valid_loss 172.31365966796875\n",
            "val_acc 0.07820400795899332\n",
            "epochs 479/1000\n",
            "train_loss 19.012569427490234\n",
            "train_acc 0.0010126642457460268\n",
            "valid_loss 172.1641082763672\n",
            "val_acc 0.07812029621563851\n",
            "epochs 480/1000\n",
            "train_loss 18.76431655883789\n",
            "train_acc 0.000993635910652721\n",
            "valid_loss 172.01507568359375\n",
            "val_acc 0.07803698242641985\n",
            "epochs 481/1000\n",
            "train_loss 18.51837158203125\n",
            "train_acc 0.0009749777024038709\n",
            "valid_loss 171.86683654785156\n",
            "val_acc 0.07795423725619913\n",
            "epochs 482/1000\n",
            "train_loss 18.274991989135742\n",
            "train_acc 0.0009566853516914237\n",
            "valid_loss 171.71913146972656\n",
            "val_acc 0.07787190834060312\n",
            "epochs 483/1000\n",
            "train_loss 18.033063888549805\n",
            "train_acc 0.0009387294009614449\n",
            "valid_loss 171.57183837890625\n",
            "val_acc 0.07778993495739997\n",
            "epochs 484/1000\n",
            "train_loss 17.7928524017334\n",
            "train_acc 0.0009211127135831004\n",
            "valid_loss 171.42503356933594\n",
            "val_acc 0.07770840290468187\n",
            "epochs 485/1000\n",
            "train_loss 17.55509376525879\n",
            "train_acc 0.0009038481064355717\n",
            "valid_loss 171.27891540527344\n",
            "val_acc 0.0776274102856405\n",
            "epochs 486/1000\n",
            "train_loss 17.320016860961914\n",
            "train_acc 0.0008869614953605029\n",
            "valid_loss 171.13365173339844\n",
            "val_acc 0.07754701088415458\n",
            "epochs 487/1000\n",
            "train_loss 17.087783813476562\n",
            "train_acc 0.0008704286211958298\n",
            "valid_loss 170.98892211914062\n",
            "val_acc 0.07746711194049567\n",
            "epochs 488/1000\n",
            "train_loss 16.85814094543457\n",
            "train_acc 0.0008542436922792918\n",
            "valid_loss 170.8448944091797\n",
            "val_acc 0.07738782422384247\n",
            "epochs 489/1000\n",
            "train_loss 16.631723403930664\n",
            "train_acc 0.0008384344221737522\n",
            "valid_loss 170.70167541503906\n",
            "val_acc 0.07730922282207757\n",
            "epochs 490/1000\n",
            "train_loss 16.408432006835938\n",
            "train_acc 0.0008229877768406788\n",
            "valid_loss 170.55914306640625\n",
            "val_acc 0.07723123981850222\n",
            "epochs 491/1000\n",
            "train_loss 16.187969207763672\n",
            "train_acc 0.0008078954236295361\n",
            "valid_loss 170.41729736328125\n",
            "val_acc 0.0771538260160014\n",
            "epochs 492/1000\n",
            "train_loss 15.970559120178223\n",
            "train_acc 0.000793148499304572\n",
            "valid_loss 170.27610778808594\n",
            "val_acc 0.07707699871389195\n",
            "epochs 493/1000\n",
            "train_loss 15.755684852600098\n",
            "train_acc 0.0007787327116462761\n",
            "valid_loss 170.1355743408203\n",
            "val_acc 0.07700073393061757\n",
            "epochs 494/1000\n",
            "train_loss 15.544382095336914\n",
            "train_acc 0.0007646737481431606\n",
            "valid_loss 169.99594116210938\n",
            "val_acc 0.0769252248806879\n",
            "epochs 495/1000\n",
            "train_loss 15.338314056396484\n",
            "train_acc 0.0007510145817202731\n",
            "valid_loss 169.8575897216797\n",
            "val_acc 0.07685067879501731\n",
            "epochs 496/1000\n",
            "train_loss 15.13577651977539\n",
            "train_acc 0.0007377087156503246\n",
            "valid_loss 169.7199249267578\n",
            "val_acc 0.07677672202698886\n",
            "epochs 497/1000\n",
            "train_loss 14.936434745788574\n",
            "train_acc 0.0007247552007006911\n",
            "valid_loss 169.58335876464844\n",
            "val_acc 0.07670357038732618\n",
            "epochs 498/1000\n",
            "train_loss 14.741700172424316\n",
            "train_acc 0.0007121672119515446\n",
            "valid_loss 169.44790649414062\n",
            "val_acc 0.07663128251442686\n",
            "epochs 499/1000\n",
            "train_loss 14.550997734069824\n",
            "train_acc 0.0006999216633490645\n",
            "valid_loss 169.31332397460938\n",
            "val_acc 0.07655967057216913\n",
            "epochs 500/1000\n",
            "train_loss 14.363205909729004\n",
            "train_acc 0.0006879876802066484\n",
            "valid_loss 169.17947387695312\n",
            "val_acc 0.07648863096255809\n",
            "epochs 501/1000\n",
            "train_loss 14.178705215454102\n",
            "train_acc 0.0006763610827664917\n",
            "valid_loss 169.04640197753906\n",
            "val_acc 0.07641823791200295\n",
            "epochs 502/1000\n",
            "train_loss 13.997627258300781\n",
            "train_acc 0.0006650563423485997\n",
            "valid_loss 168.9143524169922\n",
            "val_acc 0.07634861052501947\n",
            "epochs 503/1000\n",
            "train_loss 13.820106506347656\n",
            "train_acc 0.0006540650636172638\n",
            "valid_loss 168.783203125\n",
            "val_acc 0.0762796810362488\n",
            "epochs 504/1000\n",
            "train_loss 13.646677017211914\n",
            "train_acc 0.0006434068647929683\n",
            "valid_loss 168.6533660888672\n",
            "val_acc 0.0762116729049012\n",
            "epochs 505/1000\n",
            "train_loss 13.47788143157959\n",
            "train_acc 0.0006330853278856152\n",
            "valid_loss 168.5248260498047\n",
            "val_acc 0.07614460292970762\n",
            "epochs 506/1000\n",
            "train_loss 13.31314754486084\n",
            "train_acc 0.0006230696498488004\n",
            "valid_loss 168.39730834960938\n",
            "val_acc 0.07607827547471971\n",
            "epochs 507/1000\n",
            "train_loss 13.151938438415527\n",
            "train_acc 0.0006133499241863879\n",
            "valid_loss 168.2708282470703\n",
            "val_acc 0.0760127121466212\n",
            "epochs 508/1000\n",
            "train_loss 12.994233131408691\n",
            "train_acc 0.0006039109139237553\n",
            "valid_loss 168.14527893066406\n",
            "val_acc 0.07594784303801135\n",
            "epochs 509/1000\n",
            "train_loss 12.83942985534668\n",
            "train_acc 0.000594734473941991\n",
            "valid_loss 168.02056884765625\n",
            "val_acc 0.07588359818328172\n",
            "epochs 510/1000\n",
            "train_loss 12.687173843383789\n",
            "train_acc 0.0005858043016185268\n",
            "valid_loss 167.89659118652344\n",
            "val_acc 0.0758198935771361\n",
            "epochs 511/1000\n",
            "train_loss 12.537311553955078\n",
            "train_acc 0.0005771200230810791\n",
            "valid_loss 167.77334594726562\n",
            "val_acc 0.07575674671679736\n",
            "epochs 512/1000\n",
            "train_loss 12.389996528625488\n",
            "train_acc 0.0005686849248237335\n",
            "valid_loss 167.65098571777344\n",
            "val_acc 0.07569418866187334\n",
            "epochs 513/1000\n",
            "train_loss 12.245312690734863\n",
            "train_acc 0.0005604903562925756\n",
            "valid_loss 167.5293731689453\n",
            "val_acc 0.0756321992026642\n",
            "epochs 514/1000\n",
            "train_loss 12.10282039642334\n",
            "train_acc 0.000552522190936053\n",
            "valid_loss 167.40843200683594\n",
            "val_acc 0.07557071299524978\n",
            "epochs 515/1000\n",
            "train_loss 11.96371078491211\n",
            "train_acc 0.0005448111086689795\n",
            "valid_loss 167.28875732421875\n",
            "val_acc 0.07551010258030147\n",
            "epochs 516/1000\n",
            "train_loss 11.829147338867188\n",
            "train_acc 0.0005373771609111617\n",
            "valid_loss 167.17041015625\n",
            "val_acc 0.07545043920399622\n",
            "epochs 517/1000\n",
            "train_loss 11.697766304016113\n",
            "train_acc 0.000530183769744606\n",
            "valid_loss 167.05303955078125\n",
            "val_acc 0.07539146073395386\n",
            "epochs 518/1000\n",
            "train_loss 11.569942474365234\n",
            "train_acc 0.0005232359779676279\n",
            "valid_loss 166.93675231933594\n",
            "val_acc 0.07533323164097964\n",
            "epochs 519/1000\n",
            "train_loss 11.44552230834961\n",
            "train_acc 0.0005165249100313163\n",
            "valid_loss 166.82164001464844\n",
            "val_acc 0.07527578361332417\n",
            "epochs 520/1000\n",
            "train_loss 11.324094772338867\n",
            "train_acc 0.0005100358752175593\n",
            "valid_loss 166.7075653076172\n",
            "val_acc 0.07521906269248575\n",
            "epochs 521/1000\n",
            "train_loss 11.206159591674805\n",
            "train_acc 0.0005037797292551169\n",
            "valid_loss 166.59494018554688\n",
            "val_acc 0.07516331436345354\n",
            "epochs 522/1000\n",
            "train_loss 11.0912504196167\n",
            "train_acc 0.0004977301499680974\n",
            "valid_loss 166.4832000732422\n",
            "val_acc 0.07510817059082911\n",
            "epochs 523/1000\n",
            "train_loss 10.979201316833496\n",
            "train_acc 0.0004918951554617917\n",
            "valid_loss 166.37277221679688\n",
            "val_acc 0.0750538601190783\n",
            "epochs 524/1000\n",
            "train_loss 10.870254516601562\n",
            "train_acc 0.0004862606729147956\n",
            "valid_loss 166.26351928710938\n",
            "val_acc 0.07500023832544685\n",
            "epochs 525/1000\n",
            "train_loss 10.763873100280762\n",
            "train_acc 0.0004808132630737069\n",
            "valid_loss 166.1551971435547\n",
            "val_acc 0.0749472311057616\n",
            "epochs 526/1000\n",
            "train_loss 10.660042762756348\n",
            "train_acc 0.0004755570799506341\n",
            "valid_loss 166.0480499267578\n",
            "val_acc 0.07489500519586727\n",
            "epochs 527/1000\n",
            "train_loss 10.559247970581055\n",
            "train_acc 0.00047048882249957667\n",
            "valid_loss 165.9419403076172\n",
            "val_acc 0.07484348114812747\n",
            "epochs 528/1000\n",
            "train_loss 10.460304260253906\n",
            "train_acc 0.00046558441625585634\n",
            "valid_loss 165.8365936279297\n",
            "val_acc 0.07479248805320822\n",
            "epochs 529/1000\n",
            "train_loss 10.363747596740723\n",
            "train_acc 0.00046085751763222594\n",
            "valid_loss 165.7322998046875\n",
            "val_acc 0.07474224481265992\n",
            "epochs 530/1000\n",
            "train_loss 10.269472122192383\n",
            "train_acc 0.00045629086134095606\n",
            "valid_loss 165.62876892089844\n",
            "val_acc 0.07469252661103383\n",
            "epochs 531/1000\n",
            "train_loss 10.176912307739258\n",
            "train_acc 0.0004518722775034033\n",
            "valid_loss 165.5259246826172\n",
            "val_acc 0.07464325954788364\n",
            "epochs 532/1000\n",
            "train_loss 10.087268829345703\n",
            "train_acc 0.000447628102847375\n",
            "valid_loss 165.42434692382812\n",
            "val_acc 0.07459480203688144\n",
            "epochs 533/1000\n",
            "train_loss 10.000435829162598\n",
            "train_acc 0.00044355287033025746\n",
            "valid_loss 165.323974609375\n",
            "val_acc 0.07454711843747645\n",
            "epochs 534/1000\n",
            "train_loss 9.915667533874512\n",
            "train_acc 0.0004396239617982736\n",
            "valid_loss 165.22439575195312\n",
            "val_acc 0.07449997538933531\n",
            "epochs 535/1000\n",
            "train_loss 9.83281135559082\n",
            "train_acc 0.0004358370508550881\n",
            "valid_loss 165.12571716308594\n",
            "val_acc 0.07445341413258574\n",
            "epochs 536/1000\n",
            "train_loss 9.752288818359375\n",
            "train_acc 0.0004321952370586447\n",
            "valid_loss 165.028076171875\n",
            "val_acc 0.07440750090172514\n",
            "epochs 537/1000\n",
            "train_loss 9.674212455749512\n",
            "train_acc 0.0004287009158780655\n",
            "valid_loss 164.93162536621094\n",
            "val_acc 0.07436231769970618\n",
            "epochs 538/1000\n",
            "train_loss 9.597989082336426\n",
            "train_acc 0.0004253322012328471\n",
            "valid_loss 164.83599853515625\n",
            "val_acc 0.074317660467932\n",
            "epochs 539/1000\n",
            "train_loss 9.52387523651123\n",
            "train_acc 0.000422095188136714\n",
            "valid_loss 164.74192810058594\n",
            "val_acc 0.07427363216411323\n",
            "epochs 540/1000\n",
            "train_loss 9.45190715789795\n",
            "train_acc 0.0004189894534647465\n",
            "valid_loss 164.64906311035156\n",
            "val_acc 0.0742303014441859\n",
            "epochs 541/1000\n",
            "train_loss 9.382272720336914\n",
            "train_acc 0.00041601335620865796\n",
            "valid_loss 164.557373046875\n",
            "val_acc 0.07418774168472737\n",
            "epochs 542/1000\n",
            "train_loss 9.315485954284668\n",
            "train_acc 0.0004131775685192014\n",
            "valid_loss 164.4672088623047\n",
            "val_acc 0.07414616433670744\n",
            "epochs 543/1000\n",
            "train_loss 9.251505851745605\n",
            "train_acc 0.0004104743700456591\n",
            "valid_loss 164.3785858154297\n",
            "val_acc 0.07410560819553211\n",
            "epochs 544/1000\n",
            "train_loss 9.190805435180664\n",
            "train_acc 0.0004079116524035971\n",
            "valid_loss 164.2916717529297\n",
            "val_acc 0.0740661948104389\n",
            "epochs 545/1000\n",
            "train_loss 9.132675170898438\n",
            "train_acc 0.0004054645620411835\n",
            "valid_loss 164.2061004638672\n",
            "val_acc 0.07402769761392847\n",
            "epochs 546/1000\n",
            "train_loss 9.076750755310059\n",
            "train_acc 0.00040312660320733604\n",
            "valid_loss 164.12181091308594\n",
            "val_acc 0.07399008927168324\n",
            "epochs 547/1000\n",
            "train_loss 9.022478103637695\n",
            "train_acc 0.0004008849635003851\n",
            "valid_loss 164.0385284423828\n",
            "val_acc 0.0739531654631719\n",
            "epochs 548/1000\n",
            "train_loss 8.969695091247559\n",
            "train_acc 0.00039873326134581404\n",
            "valid_loss 163.95620727539062\n",
            "val_acc 0.07391689154901541\n",
            "epochs 549/1000\n",
            "train_loss 8.918437004089355\n",
            "train_acc 0.0003966690956776102\n",
            "valid_loss 163.87474060058594\n",
            "val_acc 0.07388122141128406\n",
            "epochs 550/1000\n",
            "train_loss 8.868688583374023\n",
            "train_acc 0.00039469054172514006\n",
            "valid_loss 163.7942352294922\n",
            "val_acc 0.07384618892101572\n",
            "epochs 551/1000\n",
            "train_loss 8.820634841918945\n",
            "train_acc 0.00039279774780600116\n",
            "valid_loss 163.71478271484375\n",
            "val_acc 0.07381187195423991\n",
            "epochs 552/1000\n",
            "train_loss 8.774531364440918\n",
            "train_acc 0.00039099710109160066\n",
            "valid_loss 163.6366424560547\n",
            "val_acc 0.0737783882941585\n",
            "epochs 553/1000\n",
            "train_loss 8.730390548706055\n",
            "train_acc 0.00038928174986862217\n",
            "valid_loss 163.5597686767578\n",
            "val_acc 0.07374576198635623\n",
            "epochs 554/1000\n",
            "train_loss 8.688126564025879\n",
            "train_acc 0.0003876493736662759\n",
            "valid_loss 163.48419189453125\n",
            "val_acc 0.07371398062095977\n",
            "epochs 555/1000\n",
            "train_loss 8.64780330657959\n",
            "train_acc 0.00038609364128205925\n",
            "valid_loss 163.4098358154297\n",
            "val_acc 0.073683012498077\n",
            "epochs 556/1000\n",
            "train_loss 8.609411239624023\n",
            "train_acc 0.00038462090966417093\n",
            "valid_loss 163.33700561523438\n",
            "val_acc 0.07365303732221946\n",
            "epochs 557/1000\n",
            "train_loss 8.572876930236816\n",
            "train_acc 0.0003832195786974178\n",
            "valid_loss 163.2654266357422\n",
            "val_acc 0.07362387136090547\n",
            "epochs 558/1000\n",
            "train_loss 8.53762149810791\n",
            "train_acc 0.00038187750089071836\n",
            "valid_loss 163.19480895996094\n",
            "val_acc 0.07359532921691425\n",
            "epochs 559/1000\n",
            "train_loss 8.503410339355469\n",
            "train_acc 0.0003805910248327284\n",
            "valid_loss 163.12506103515625\n",
            "val_acc 0.07356738318921999\n",
            "epochs 560/1000\n",
            "train_loss 8.470184326171875\n",
            "train_acc 0.00037935837407041196\n",
            "valid_loss 163.05613708496094\n",
            "val_acc 0.07353997447353322\n",
            "epochs 561/1000\n",
            "train_loss 8.437870979309082\n",
            "train_acc 0.00037817419685709936\n",
            "valid_loss 162.9879150390625\n",
            "val_acc 0.07351306180353276\n",
            "epochs 562/1000\n",
            "train_loss 8.406277656555176\n",
            "train_acc 0.00037703426586141664\n",
            "valid_loss 162.9203338623047\n",
            "val_acc 0.07348657847614959\n",
            "epochs 563/1000\n",
            "train_loss 8.375412940979004\n",
            "train_acc 0.0003759372507024986\n",
            "valid_loss 162.85342407226562\n",
            "val_acc 0.07346049468615093\n",
            "epochs 564/1000\n",
            "train_loss 8.345441818237305\n",
            "train_acc 0.0003748871139787997\n",
            "valid_loss 162.78744506835938\n",
            "val_acc 0.07343493810622022\n",
            "epochs 565/1000\n",
            "train_loss 8.316950798034668\n",
            "train_acc 0.00037389170365014044\n",
            "valid_loss 162.7227325439453\n",
            "val_acc 0.07341014231787994\n",
            "epochs 566/1000\n",
            "train_loss 8.28967571258545\n",
            "train_acc 0.0003729419768205844\n",
            "valid_loss 162.65921020507812\n",
            "val_acc 0.07338609623257071\n",
            "epochs 567/1000\n",
            "train_loss 8.263224601745605\n",
            "train_acc 0.00037203017899838206\n",
            "valid_loss 162.59652709960938\n",
            "val_acc 0.07336262452590744\n",
            "epochs 568/1000\n",
            "train_loss 8.237563133239746\n",
            "train_acc 0.00037115431489207997\n",
            "valid_loss 162.53465270996094\n",
            "val_acc 0.07333968157763593\n",
            "epochs 569/1000\n",
            "train_loss 8.213132858276367\n",
            "train_acc 0.0003703249447146216\n",
            "valid_loss 162.47402954101562\n",
            "val_acc 0.07331748584692832\n",
            "epochs 570/1000\n",
            "train_loss 8.189634323120117\n",
            "train_acc 0.00036953090546795953\n",
            "valid_loss 162.4143829345703\n",
            "val_acc 0.07329595008632168\n",
            "epochs 571/1000\n",
            "train_loss 8.167415618896484\n",
            "train_acc 0.00036877969418688177\n",
            "valid_loss 162.35609436035156\n",
            "val_acc 0.07327528197783977\n",
            "epochs 572/1000\n",
            "train_loss 8.145902633666992\n",
            "train_acc 0.0003680538513715594\n",
            "valid_loss 162.29872131347656\n",
            "val_acc 0.07325520750309807\n",
            "epochs 573/1000\n",
            "train_loss 8.124979972839355\n",
            "train_acc 0.00036735308934182214\n",
            "valid_loss 162.2420654296875\n",
            "val_acc 0.07323559865471907\n",
            "epochs 574/1000\n",
            "train_loss 8.104522705078125\n",
            "train_acc 0.0003666769765774146\n",
            "valid_loss 162.18605041503906\n",
            "val_acc 0.07321643979812506\n",
            "epochs 575/1000\n",
            "train_loss 8.084731101989746\n",
            "train_acc 0.000366027090841761\n",
            "valid_loss 162.1308135986328\n",
            "val_acc 0.0731978126859758\n",
            "epochs 576/1000\n",
            "train_loss 8.065664291381836\n",
            "train_acc 0.0003654053283613533\n",
            "valid_loss 162.0764923095703\n",
            "val_acc 0.07317979477811605\n",
            "epochs 577/1000\n",
            "train_loss 8.047185897827148\n",
            "train_acc 0.00036480541166383773\n",
            "valid_loss 162.02297973632812\n",
            "val_acc 0.07316233157180249\n",
            "epochs 578/1000\n",
            "train_loss 8.029230117797852\n",
            "train_acc 0.0003642267676276298\n",
            "valid_loss 161.9702606201172\n",
            "val_acc 0.07314539528160821\n",
            "epochs 579/1000\n",
            "train_loss 8.011799812316895\n",
            "train_acc 0.00036366742111008184\n",
            "valid_loss 161.91822814941406\n",
            "val_acc 0.07312890492903534\n",
            "epochs 580/1000\n",
            "train_loss 7.995599269866943\n",
            "train_acc 0.0003631436831290977\n",
            "valid_loss 161.86761474609375\n",
            "val_acc 0.07311334144033026\n",
            "epochs 581/1000\n",
            "train_loss 7.9798479080200195\n",
            "train_acc 0.0003626311294143446\n",
            "valid_loss 161.8179931640625\n",
            "val_acc 0.073098411236424\n",
            "epochs 582/1000\n",
            "train_loss 7.964480400085449\n",
            "train_acc 0.0003621312901556778\n",
            "valid_loss 161.7691192626953\n",
            "val_acc 0.07308396472071763\n",
            "epochs 583/1000\n",
            "train_loss 7.9493608474731445\n",
            "train_acc 0.00036164293123874813\n",
            "valid_loss 161.7207489013672\n",
            "val_acc 0.07306985283794347\n",
            "epochs 584/1000\n",
            "train_loss 7.934426307678223\n",
            "train_acc 0.00036116587804057275\n",
            "valid_loss 161.67279052734375\n",
            "val_acc 0.07305601488915273\n",
            "epochs 585/1000\n",
            "train_loss 7.919738292694092\n",
            "train_acc 0.00036070137642897095\n",
            "valid_loss 161.62521362304688\n",
            "val_acc 0.07304244297847617\n",
            "epochs 586/1000\n",
            "train_loss 7.905251502990723\n",
            "train_acc 0.0003602479090869355\n",
            "valid_loss 161.57803344726562\n",
            "val_acc 0.07302914413739928\n",
            "epochs 587/1000\n",
            "train_loss 7.891010284423828\n",
            "train_acc 0.0003598072491093682\n",
            "valid_loss 161.53128051757812\n",
            "val_acc 0.07301615633768961\n",
            "epochs 588/1000\n",
            "train_loss 7.8770341873168945\n",
            "train_acc 0.00035937732564679417\n",
            "valid_loss 161.4849853515625\n",
            "val_acc 0.07300348412827588\n",
            "epochs 589/1000\n",
            "train_loss 7.863356590270996\n",
            "train_acc 0.0003589589670390034\n",
            "valid_loss 161.4391326904297\n",
            "val_acc 0.0729911373637151\n",
            "epochs 590/1000\n",
            "train_loss 7.84993314743042\n",
            "train_acc 0.0003585502764998147\n",
            "valid_loss 161.3938446044922\n",
            "val_acc 0.07297917631221935\n",
            "epochs 591/1000\n",
            "train_loss 7.836790561676025\n",
            "train_acc 0.00035815130328186427\n",
            "valid_loss 161.34910583496094\n",
            "val_acc 0.07296760840108618\n",
            "epochs 592/1000\n",
            "train_loss 7.8239240646362305\n",
            "train_acc 0.00035776217611363303\n",
            "valid_loss 161.304931640625\n",
            "val_acc 0.07295640765223652\n",
            "epochs 593/1000\n",
            "train_loss 7.811194896697998\n",
            "train_acc 0.0003573788109441431\n",
            "valid_loss 161.2611541748047\n",
            "val_acc 0.07294549592770636\n",
            "epochs 594/1000\n",
            "train_loss 7.7986040115356445\n",
            "train_acc 0.00035700272844853596\n",
            "valid_loss 161.2176971435547\n",
            "val_acc 0.0729348100896459\n",
            "epochs 595/1000\n",
            "train_loss 7.786150932312012\n",
            "train_acc 0.00035663372489999834\n",
            "valid_loss 161.17459106445312\n",
            "val_acc 0.07292438766744454\n",
            "epochs 596/1000\n",
            "train_loss 7.773837089538574\n",
            "train_acc 0.0003562715193346286\n",
            "valid_loss 161.1317596435547\n",
            "val_acc 0.07291416133812163\n",
            "epochs 597/1000\n",
            "train_loss 7.761703968048096\n",
            "train_acc 0.00035591724288399116\n",
            "valid_loss 161.08935546875\n",
            "val_acc 0.07290423556114548\n",
            "epochs 598/1000\n",
            "train_loss 7.749795436859131\n",
            "train_acc 0.0003555705726075058\n",
            "valid_loss 161.04739379882812\n",
            "val_acc 0.07289463277265895\n",
            "epochs 599/1000\n",
            "train_loss 7.738022804260254\n",
            "train_acc 0.00035522848170680495\n",
            "valid_loss 161.0058135986328\n",
            "val_acc 0.07288531074882484\n",
            "epochs 600/1000\n",
            "train_loss 7.726309776306152\n",
            "train_acc 0.0003548901496777454\n",
            "valid_loss 160.96453857421875\n",
            "val_acc 0.07287617564143148\n",
            "epochs 601/1000\n",
            "train_loss 7.71468448638916\n",
            "train_acc 0.00035455692201279674\n",
            "valid_loss 160.92344665527344\n",
            "val_acc 0.07286723132128828\n",
            "epochs 602/1000\n",
            "train_loss 7.703141689300537\n",
            "train_acc 0.0003542287388252309\n",
            "valid_loss 160.882568359375\n",
            "val_acc 0.0728584258176852\n",
            "epochs 603/1000\n",
            "train_loss 7.691771984100342\n",
            "train_acc 0.0003539076659273213\n",
            "valid_loss 160.84207153320312\n",
            "val_acc 0.07284987405291758\n",
            "epochs 604/1000\n",
            "train_loss 7.680507183074951\n",
            "train_acc 0.0003535909675590049\n",
            "valid_loss 160.80186462402344\n",
            "val_acc 0.07284157057874836\n",
            "epochs 605/1000\n",
            "train_loss 7.6693949699401855\n",
            "train_acc 0.0003532797429495706\n",
            "valid_loss 160.76206970214844\n",
            "val_acc 0.07283355394902173\n",
            "epochs 606/1000\n",
            "train_loss 7.658429145812988\n",
            "train_acc 0.00035297207796247676\n",
            "valid_loss 160.72268676757812\n",
            "val_acc 0.07282581391045824\n",
            "epochs 607/1000\n",
            "train_loss 7.647584438323975\n",
            "train_acc 0.0003526679267032215\n",
            "valid_loss 160.68362426757812\n",
            "val_acc 0.07281830084975809\n",
            "epochs 608/1000\n",
            "train_loss 7.63681173324585\n",
            "train_acc 0.00035236671157270815\n",
            "valid_loss 160.64483642578125\n",
            "val_acc 0.07281098319508601\n",
            "epochs 609/1000\n",
            "train_loss 7.626116752624512\n",
            "train_acc 0.00035206868555038597\n",
            "valid_loss 160.6063690185547\n",
            "val_acc 0.07280389243387617\n",
            "epochs 610/1000\n",
            "train_loss 7.615610122680664\n",
            "train_acc 0.00035177606612426013\n",
            "valid_loss 160.56825256347656\n",
            "val_acc 0.07279703010572121\n",
            "epochs 611/1000\n",
            "train_loss 7.605210781097412\n",
            "train_acc 0.0003514859557841331\n",
            "valid_loss 160.5305633544922\n",
            "val_acc 0.07279045280301943\n",
            "epochs 612/1000\n",
            "train_loss 7.594911098480225\n",
            "train_acc 0.0003511982985611002\n",
            "valid_loss 160.49319458007812\n",
            "val_acc 0.07278409710852429\n",
            "epochs 613/1000\n",
            "train_loss 7.584692001342773\n",
            "train_acc 0.00035091294054067333\n",
            "valid_loss 160.4561309814453\n",
            "val_acc 0.0727779696957441\n",
            "epochs 614/1000\n",
            "train_loss 7.574563980102539\n",
            "train_acc 0.0003506300026156868\n",
            "valid_loss 160.41941833496094\n",
            "val_acc 0.07277207282895688\n",
            "epochs 615/1000\n",
            "train_loss 7.564627170562744\n",
            "train_acc 0.00035035168604316335\n",
            "valid_loss 160.38316345214844\n",
            "val_acc 0.07276649129635189\n",
            "epochs 616/1000\n",
            "train_loss 7.554705619812012\n",
            "train_acc 0.000350072031813817\n",
            "valid_loss 160.3472137451172\n",
            "val_acc 0.07276112420950084\n",
            "epochs 617/1000\n",
            "train_loss 7.544859409332275\n",
            "train_acc 0.00034979421112578933\n",
            "valid_loss 160.3115692138672\n",
            "val_acc 0.07275598406558856\n",
            "epochs 618/1000\n",
            "train_loss 7.535216808319092\n",
            "train_acc 0.00034952058250872564\n",
            "valid_loss 160.2764129638672\n",
            "val_acc 0.0727511545876041\n",
            "epochs 619/1000\n",
            "train_loss 7.525635242462158\n",
            "train_acc 0.0003492463203236604\n",
            "valid_loss 160.24168395996094\n",
            "val_acc 0.0727466237382032\n",
            "epochs 620/1000\n",
            "train_loss 7.51613187789917\n",
            "train_acc 0.00034897202847507567\n",
            "valid_loss 160.20730590820312\n",
            "val_acc 0.07274232374911663\n",
            "epochs 621/1000\n",
            "train_loss 7.506688117980957\n",
            "train_acc 0.0003486982655326406\n",
            "valid_loss 160.17323303222656\n",
            "val_acc 0.07273823343566618\n",
            "epochs 622/1000\n",
            "train_loss 7.497359275817871\n",
            "train_acc 0.00034842632717649953\n",
            "valid_loss 160.13958740234375\n",
            "val_acc 0.07273443984449841\n",
            "epochs 623/1000\n",
            "train_loss 7.488061904907227\n",
            "train_acc 0.0003481533864172749\n",
            "valid_loss 160.1062469482422\n",
            "val_acc 0.07273086499772034\n",
            "epochs 624/1000\n",
            "train_loss 7.478840351104736\n",
            "train_acc 0.0003478817765986045\n",
            "valid_loss 160.07322692871094\n",
            "val_acc 0.07272751782729756\n",
            "epochs 625/1000\n",
            "train_loss 7.469659805297852\n",
            "train_acc 0.0003476098466377992\n",
            "valid_loss 160.0404815673828\n",
            "val_acc 0.07272435621125624\n",
            "epochs 626/1000\n",
            "train_loss 7.460515022277832\n",
            "train_acc 0.00034733824073695217\n",
            "valid_loss 160.0079803466797\n",
            "val_acc 0.07272136609535665\n",
            "epochs 627/1000\n",
            "train_loss 7.451387405395508\n",
            "train_acc 0.0003470663387605992\n",
            "valid_loss 159.97560119628906\n",
            "val_acc 0.07271848735690582\n",
            "epochs 628/1000\n",
            "train_loss 7.442276954650879\n",
            "train_acc 0.0003467951274405305\n",
            "valid_loss 159.9434356689453\n",
            "val_acc 0.07271576237399131\n",
            "epochs 629/1000\n",
            "train_loss 7.433226585388184\n",
            "train_acc 0.0003465256181348545\n",
            "valid_loss 159.91143798828125\n",
            "val_acc 0.07271318116690964\n",
            "epochs 630/1000\n",
            "train_loss 7.42421817779541\n",
            "train_acc 0.0003462565022905787\n",
            "valid_loss 159.8796844482422\n",
            "val_acc 0.0727107756130863\n",
            "epochs 631/1000\n",
            "train_loss 7.415217876434326\n",
            "train_acc 0.00034598735118589294\n",
            "valid_loss 159.84808349609375\n",
            "val_acc 0.07270850556378719\n",
            "epochs 632/1000\n",
            "train_loss 7.406255722045898\n",
            "train_acc 0.0003457191689029479\n",
            "valid_loss 159.8166961669922\n",
            "val_acc 0.07270636596949771\n",
            "epochs 633/1000\n",
            "train_loss 7.397431373596191\n",
            "train_acc 0.00034545365018017875\n",
            "valid_loss 159.78562927246094\n",
            "val_acc 0.07270447020418942\n",
            "epochs 634/1000\n",
            "train_loss 7.388626575469971\n",
            "train_acc 0.0003451861031443024\n",
            "valid_loss 159.7549285888672\n",
            "val_acc 0.07270281158271245\n",
            "epochs 635/1000\n",
            "train_loss 7.379820823669434\n",
            "train_acc 0.00034491708132778655\n",
            "valid_loss 159.7244415283203\n",
            "val_acc 0.07270130004035309\n",
            "epochs 636/1000\n",
            "train_loss 7.371039390563965\n",
            "train_acc 0.0003446479363796803\n",
            "valid_loss 159.69412231445312\n",
            "val_acc 0.07269994847010822\n",
            "epochs 637/1000\n",
            "train_loss 7.362281322479248\n",
            "train_acc 0.0003443788848996449\n",
            "valid_loss 159.6639862060547\n",
            "val_acc 0.07269870632444508\n",
            "epochs 638/1000\n",
            "train_loss 7.353643417358398\n",
            "train_acc 0.0003441120341169433\n",
            "valid_loss 159.6342010498047\n",
            "val_acc 0.07269768826081417\n",
            "epochs 639/1000\n",
            "train_loss 7.345029830932617\n",
            "train_acc 0.0003438435546391142\n",
            "valid_loss 159.60470581054688\n",
            "val_acc 0.07269689244276378\n",
            "epochs 640/1000\n",
            "train_loss 7.336404323577881\n",
            "train_acc 0.0003435735785527728\n",
            "valid_loss 159.5753631591797\n",
            "val_acc 0.07269620270817541\n",
            "epochs 641/1000\n",
            "train_loss 7.327822208404541\n",
            "train_acc 0.0003433039584286654\n",
            "valid_loss 159.5462188720703\n",
            "val_acc 0.07269565944443457\n",
            "epochs 642/1000\n",
            "train_loss 7.319267749786377\n",
            "train_acc 0.00034303403495309444\n",
            "valid_loss 159.51731872558594\n",
            "val_acc 0.07269529325712938\n",
            "epochs 643/1000\n",
            "train_loss 7.3107380867004395\n",
            "train_acc 0.00034276405215048447\n",
            "valid_loss 159.48863220214844\n",
            "val_acc 0.07269506339216605\n",
            "epochs 644/1000\n",
            "train_loss 7.302253246307373\n",
            "train_acc 0.00034249409845170495\n",
            "valid_loss 159.4601287841797\n",
            "val_acc 0.0726949862902984\n",
            "epochs 645/1000\n",
            "train_loss 7.293781280517578\n",
            "train_acc 0.00034222312276072515\n",
            "valid_loss 159.4318389892578\n",
            "val_acc 0.07269506438169629\n",
            "epochs 646/1000\n",
            "train_loss 7.285398960113525\n",
            "train_acc 0.0003419534097282359\n",
            "valid_loss 159.40383911132812\n",
            "val_acc 0.07269532314676326\n",
            "epochs 647/1000\n",
            "train_loss 7.277064800262451\n",
            "train_acc 0.00034168244858917134\n",
            "valid_loss 159.37620544433594\n",
            "val_acc 0.07269583406159655\n",
            "epochs 648/1000\n",
            "train_loss 7.268759250640869\n",
            "train_acc 0.00034140986659062596\n",
            "valid_loss 159.34878540039062\n",
            "val_acc 0.07269648192741443\n",
            "epochs 649/1000\n",
            "train_loss 7.260470867156982\n",
            "train_acc 0.00034113608797689754\n",
            "valid_loss 159.3216094970703\n",
            "val_acc 0.07269729913095943\n",
            "epochs 650/1000\n",
            "train_loss 7.252193450927734\n",
            "train_acc 0.0003408615397907292\n",
            "valid_loss 159.29461669921875\n",
            "val_acc 0.0726982574822614\n",
            "epochs 651/1000\n",
            "train_loss 7.243957042694092\n",
            "train_acc 0.00034058699048518273\n",
            "valid_loss 159.26788330078125\n",
            "val_acc 0.07269939589750721\n",
            "epochs 652/1000\n",
            "train_loss 7.235813617706299\n",
            "train_acc 0.00034031279882093746\n",
            "valid_loss 159.24148559570312\n",
            "val_acc 0.07270075955311768\n",
            "epochs 653/1000\n",
            "train_loss 7.227687358856201\n",
            "train_acc 0.0003400364680251536\n",
            "valid_loss 159.21543884277344\n",
            "val_acc 0.07270236250478775\n",
            "epochs 654/1000\n",
            "train_loss 7.219514846801758\n",
            "train_acc 0.0003397567253649378\n",
            "valid_loss 159.18955993652344\n",
            "val_acc 0.0727040866142488\n",
            "epochs 655/1000\n",
            "train_loss 7.211360931396484\n",
            "train_acc 0.0003394764454032366\n",
            "valid_loss 159.16384887695312\n",
            "val_acc 0.0727059279641253\n",
            "epochs 656/1000\n",
            "train_loss 7.203212261199951\n",
            "train_acc 0.00033919524083522934\n",
            "valid_loss 159.13832092285156\n",
            "val_acc 0.07270791418268345\n",
            "epochs 657/1000\n",
            "train_loss 7.195008754730225\n",
            "train_acc 0.00033891238854266703\n",
            "valid_loss 159.11279296875\n",
            "val_acc 0.07270990571123548\n",
            "epochs 658/1000\n",
            "train_loss 7.186837196350098\n",
            "train_acc 0.0003386310267020375\n",
            "valid_loss 159.08734130859375\n",
            "val_acc 0.0727119572868105\n",
            "epochs 659/1000\n",
            "train_loss 7.178683757781982\n",
            "train_acc 0.0003383504876043074\n",
            "valid_loss 159.06199645996094\n",
            "val_acc 0.07271409896493423\n",
            "epochs 660/1000\n",
            "train_loss 7.170544624328613\n",
            "train_acc 0.0003380702462611505\n",
            "valid_loss 159.0367889404297\n",
            "val_acc 0.0727163275150815\n",
            "epochs 661/1000\n",
            "train_loss 7.162419319152832\n",
            "train_acc 0.0003377901515565239\n",
            "valid_loss 159.01165771484375\n",
            "val_acc 0.07271863302885322\n",
            "epochs 662/1000\n",
            "train_loss 7.15431022644043\n",
            "train_acc 0.000337509957786936\n",
            "valid_loss 158.9866485595703\n",
            "val_acc 0.07272102731658378\n",
            "epochs 663/1000\n",
            "train_loss 7.146194934844971\n",
            "train_acc 0.00033722923958871083\n",
            "valid_loss 158.96173095703125\n",
            "val_acc 0.0727234774880344\n",
            "epochs 664/1000\n",
            "train_loss 7.138078212738037\n",
            "train_acc 0.00033694854433773656\n",
            "valid_loss 158.93685913085938\n",
            "val_acc 0.07272599827556406\n",
            "epochs 665/1000\n",
            "train_loss 7.130010604858398\n",
            "train_acc 0.0003366690994320939\n",
            "valid_loss 158.9121856689453\n",
            "val_acc 0.0727286643726984\n",
            "epochs 666/1000\n",
            "train_loss 7.121946334838867\n",
            "train_acc 0.00033638851443986193\n",
            "valid_loss 158.88763427734375\n",
            "val_acc 0.07273142048652517\n",
            "epochs 667/1000\n",
            "train_loss 7.11386251449585\n",
            "train_acc 0.0003361069712599811\n",
            "valid_loss 158.8632049560547\n",
            "val_acc 0.07273428777989466\n",
            "epochs 668/1000\n",
            "train_loss 7.105788230895996\n",
            "train_acc 0.0003358255321822631\n",
            "valid_loss 158.8388671875\n",
            "val_acc 0.07273722903191811\n",
            "epochs 669/1000\n",
            "train_loss 7.0977396965026855\n",
            "train_acc 0.000335544388620362\n",
            "valid_loss 158.81468200683594\n",
            "val_acc 0.07274027525127166\n",
            "epochs 670/1000\n",
            "train_loss 7.089703559875488\n",
            "train_acc 0.0003352629931983896\n",
            "valid_loss 158.79071044921875\n",
            "val_acc 0.07274342859309399\n",
            "epochs 671/1000\n",
            "train_loss 7.081684589385986\n",
            "train_acc 0.0003349810677508895\n",
            "valid_loss 158.7668914794922\n",
            "val_acc 0.07274666831071955\n",
            "epochs 672/1000\n",
            "train_loss 7.073674201965332\n",
            "train_acc 0.00033469893801688717\n",
            "valid_loss 158.7432098388672\n",
            "val_acc 0.07275002643436893\n",
            "epochs 673/1000\n",
            "train_loss 7.065698146820068\n",
            "train_acc 0.0003344167618286939\n",
            "valid_loss 158.71971130371094\n",
            "val_acc 0.07275351151911309\n",
            "epochs 674/1000\n",
            "train_loss 7.057741641998291\n",
            "train_acc 0.0003341338115905483\n",
            "valid_loss 158.69642639160156\n",
            "val_acc 0.07275713898270623\n",
            "epochs 675/1000\n",
            "train_loss 7.049778938293457\n",
            "train_acc 0.000333849597014845\n",
            "valid_loss 158.6732940673828\n",
            "val_acc 0.07276086818019394\n",
            "epochs 676/1000\n",
            "train_loss 7.041810989379883\n",
            "train_acc 0.0003335647645424335\n",
            "valid_loss 158.6502685546875\n",
            "val_acc 0.07276469474454643\n",
            "epochs 677/1000\n",
            "train_loss 7.033865928649902\n",
            "train_acc 0.0003332802745997189\n",
            "valid_loss 158.62744140625\n",
            "val_acc 0.07276858999102842\n",
            "epochs 678/1000\n",
            "train_loss 7.025915622711182\n",
            "train_acc 0.00033299534194296797\n",
            "valid_loss 158.6046600341797\n",
            "val_acc 0.07277253984939307\n",
            "epochs 679/1000\n",
            "train_loss 7.017953872680664\n",
            "train_acc 0.0003327102212306972\n",
            "valid_loss 158.58192443847656\n",
            "val_acc 0.07277653975179418\n",
            "epochs 680/1000\n",
            "train_loss 7.009988784790039\n",
            "train_acc 0.0003324252891336353\n",
            "valid_loss 158.55921936035156\n",
            "val_acc 0.07278058009687811\n",
            "epochs 681/1000\n",
            "train_loss 7.002040386199951\n",
            "train_acc 0.00033214118593605235\n",
            "valid_loss 158.5365447998047\n",
            "val_acc 0.07278465127456002\n",
            "epochs 682/1000\n",
            "train_loss 6.99410343170166\n",
            "train_acc 0.00033185772973900806\n",
            "valid_loss 158.5139617919922\n",
            "val_acc 0.07278879168006824\n",
            "epochs 683/1000\n",
            "train_loss 6.986182689666748\n",
            "train_acc 0.00033157451756638835\n",
            "valid_loss 158.49148559570312\n",
            "val_acc 0.07279301951639354\n",
            "epochs 684/1000\n",
            "train_loss 6.978255271911621\n",
            "train_acc 0.0003312910753615702\n",
            "valid_loss 158.46902465820312\n",
            "val_acc 0.07279727502609604\n",
            "epochs 685/1000\n",
            "train_loss 6.970335483551025\n",
            "train_acc 0.00033100820739771455\n",
            "valid_loss 158.44664001464844\n",
            "val_acc 0.07280157184286509\n",
            "epochs 686/1000\n",
            "train_loss 6.962432384490967\n",
            "train_acc 0.00033072603232889937\n",
            "valid_loss 158.42428588867188\n",
            "val_acc 0.0728059100874816\n",
            "epochs 687/1000\n",
            "train_loss 6.954525947570801\n",
            "train_acc 0.0003304438511035047\n",
            "valid_loss 158.4019775390625\n",
            "val_acc 0.07281030317390105\n",
            "epochs 688/1000\n",
            "train_loss 6.9466447830200195\n",
            "train_acc 0.000330162676758706\n",
            "valid_loss 158.37979125976562\n",
            "val_acc 0.07281477810465731\n",
            "epochs 689/1000\n",
            "train_loss 6.938777446746826\n",
            "train_acc 0.00032988149457826064\n",
            "valid_loss 158.35769653320312\n",
            "val_acc 0.07281933364283759\n",
            "epochs 690/1000\n",
            "train_loss 6.930901527404785\n",
            "train_acc 0.00032959990718294506\n",
            "valid_loss 158.3356170654297\n",
            "val_acc 0.07282391893386375\n",
            "epochs 691/1000\n",
            "train_loss 6.923023223876953\n",
            "train_acc 0.00032931837127902184\n",
            "valid_loss 158.31358337402344\n",
            "val_acc 0.07282854629884242\n",
            "epochs 692/1000\n",
            "train_loss 6.915127754211426\n",
            "train_acc 0.00032903672287760017\n",
            "valid_loss 158.29159545898438\n",
            "val_acc 0.07283321319700917\n",
            "epochs 693/1000\n",
            "train_loss 6.9072465896606445\n",
            "train_acc 0.0003287560561707673\n",
            "valid_loss 158.26963806152344\n",
            "val_acc 0.07283791042864322\n",
            "epochs 694/1000\n",
            "train_loss 6.899379730224609\n",
            "train_acc 0.0003284761668720211\n",
            "valid_loss 158.2476806640625\n",
            "val_acc 0.0728426458008471\n",
            "epochs 695/1000\n",
            "train_loss 6.89143705368042\n",
            "train_acc 0.0003281948548437168\n",
            "valid_loss 158.2256317138672\n",
            "val_acc 0.07284733458654954\n",
            "epochs 696/1000\n",
            "train_loss 6.883584022521973\n",
            "train_acc 0.0003279179419713238\n",
            "valid_loss 158.2035369873047\n",
            "val_acc 0.07285201196209527\n",
            "epochs 697/1000\n",
            "train_loss 6.875759601593018\n",
            "train_acc 0.00032764232198063\n",
            "valid_loss 158.1815185546875\n",
            "val_acc 0.07285673042060807\n",
            "epochs 698/1000\n",
            "train_loss 6.867939472198486\n",
            "train_acc 0.000327367233694531\n",
            "valid_loss 158.15953063964844\n",
            "val_acc 0.07286148639104795\n",
            "epochs 699/1000\n",
            "train_loss 6.86013126373291\n",
            "train_acc 0.0003270927118137479\n",
            "valid_loss 158.1375732421875\n",
            "val_acc 0.07286629224254285\n",
            "epochs 700/1000\n",
            "train_loss 6.852316379547119\n",
            "train_acc 0.00032681822127555136\n",
            "valid_loss 158.1156463623047\n",
            "val_acc 0.07287113080528798\n",
            "epochs 701/1000\n",
            "train_loss 6.844523906707764\n",
            "train_acc 0.00032654488649523747\n",
            "valid_loss 158.09376525878906\n",
            "val_acc 0.07287600696581649\n",
            "epochs 702/1000\n",
            "train_loss 6.836740970611572\n",
            "train_acc 0.00032627185618576524\n",
            "valid_loss 158.0719451904297\n",
            "val_acc 0.07288093037059298\n",
            "epochs 703/1000\n",
            "train_loss 6.828968524932861\n",
            "train_acc 0.00032599902120777045\n",
            "valid_loss 158.05018615722656\n",
            "val_acc 0.07288590521784499\n",
            "epochs 704/1000\n",
            "train_loss 6.821196556091309\n",
            "train_acc 0.00032572644368673745\n",
            "valid_loss 158.0284881591797\n",
            "val_acc 0.07289093207364203\n",
            "epochs 705/1000\n",
            "train_loss 6.813431739807129\n",
            "train_acc 0.00032545444152604503\n",
            "valid_loss 158.0067901611328\n",
            "val_acc 0.07289599271462066\n",
            "epochs 706/1000\n",
            "train_loss 6.805674076080322\n",
            "train_acc 0.00032518291286228655\n",
            "valid_loss 157.9851531982422\n",
            "val_acc 0.07290106212021782\n",
            "epochs 707/1000\n",
            "train_loss 6.797926902770996\n",
            "train_acc 0.0003249121078764662\n",
            "valid_loss 157.96353149414062\n",
            "val_acc 0.07290614564699353\n",
            "epochs 708/1000\n",
            "train_loss 6.790187358856201\n",
            "train_acc 0.000324642051754591\n",
            "valid_loss 157.9419403076172\n",
            "val_acc 0.07291126477066427\n",
            "epochs 709/1000\n",
            "train_loss 6.782451152801514\n",
            "train_acc 0.00032437239637007366\n",
            "valid_loss 157.92039489746094\n",
            "val_acc 0.07291641782503575\n",
            "epochs 710/1000\n",
            "train_loss 6.774730682373047\n",
            "train_acc 0.00032410347697665344\n",
            "valid_loss 157.89891052246094\n",
            "val_acc 0.07292162309749983\n",
            "epochs 711/1000\n",
            "train_loss 6.767043590545654\n",
            "train_acc 0.00032383519171092374\n",
            "valid_loss 157.87757873535156\n",
            "val_acc 0.07292691124166595\n",
            "epochs 712/1000\n",
            "train_loss 6.759343147277832\n",
            "train_acc 0.00032356627847408305\n",
            "valid_loss 157.85633850097656\n",
            "val_acc 0.07293225637404248\n",
            "epochs 713/1000\n",
            "train_loss 6.751651287078857\n",
            "train_acc 0.000323297751982374\n",
            "valid_loss 157.835205078125\n",
            "val_acc 0.07293765236536273\n",
            "epochs 714/1000\n",
            "train_loss 6.743987083435059\n",
            "train_acc 0.0003230298982368997\n",
            "valid_loss 157.81430053710938\n",
            "val_acc 0.07294310223514913\n",
            "epochs 715/1000\n",
            "train_loss 6.736300945281982\n",
            "train_acc 0.00032276150663025107\n",
            "valid_loss 157.79344177246094\n",
            "val_acc 0.07294858751556603\n",
            "epochs 716/1000\n",
            "train_loss 6.728634357452393\n",
            "train_acc 0.00032249393832619087\n",
            "valid_loss 157.772705078125\n",
            "val_acc 0.0729541227541631\n",
            "epochs 717/1000\n",
            "train_loss 6.720980167388916\n",
            "train_acc 0.0003222264780421168\n",
            "valid_loss 157.7522430419922\n",
            "val_acc 0.07295971434505191\n",
            "epochs 718/1000\n",
            "train_loss 6.713345527648926\n",
            "train_acc 0.00032195941177913203\n",
            "valid_loss 157.73187255859375\n",
            "val_acc 0.07296536122157704\n",
            "epochs 719/1000\n",
            "train_loss 6.705698490142822\n",
            "train_acc 0.00032169221790704446\n",
            "valid_loss 157.7115936279297\n",
            "val_acc 0.0729710459097987\n",
            "epochs 720/1000\n",
            "train_loss 6.698069095611572\n",
            "train_acc 0.0003214257941670859\n",
            "valid_loss 157.69139099121094\n",
            "val_acc 0.07297677562892205\n",
            "epochs 721/1000\n",
            "train_loss 6.690415859222412\n",
            "train_acc 0.0003211590536431266\n",
            "valid_loss 157.67115783691406\n",
            "val_acc 0.07298251456813887\n",
            "epochs 722/1000\n",
            "train_loss 6.682816982269287\n",
            "train_acc 0.0003208942574789174\n",
            "valid_loss 157.65106201171875\n",
            "val_acc 0.07298830290674233\n",
            "epochs 723/1000\n",
            "train_loss 6.675234317779541\n",
            "train_acc 0.00032062955814091343\n",
            "valid_loss 157.63107299804688\n",
            "val_acc 0.07299415886955103\n",
            "epochs 724/1000\n",
            "train_loss 6.667639255523682\n",
            "train_acc 0.00032036449332596164\n",
            "valid_loss 157.6112060546875\n",
            "val_acc 0.07300003327982267\n",
            "epochs 725/1000\n",
            "train_loss 6.660057544708252\n",
            "train_acc 0.0003201002154338102\n",
            "valid_loss 157.59146118164062\n",
            "val_acc 0.0730059294932289\n",
            "epochs 726/1000\n",
            "train_loss 6.65247917175293\n",
            "train_acc 0.00031983643062770937\n",
            "valid_loss 157.57180786132812\n",
            "val_acc 0.07301186035911086\n",
            "epochs 727/1000\n",
            "train_loss 6.644904613494873\n",
            "train_acc 0.0003195730499171007\n",
            "valid_loss 157.55223083496094\n",
            "val_acc 0.07301783398434054\n",
            "epochs 728/1000\n",
            "train_loss 6.637364387512207\n",
            "train_acc 0.00031931046228587197\n",
            "valid_loss 157.53277587890625\n",
            "val_acc 0.07302387185336556\n",
            "epochs 729/1000\n",
            "train_loss 6.62986421585083\n",
            "train_acc 0.00031904833248028386\n",
            "valid_loss 157.5134735107422\n",
            "val_acc 0.07303000899846665\n",
            "epochs 730/1000\n",
            "train_loss 6.622359752655029\n",
            "train_acc 0.0003187854879517825\n",
            "valid_loss 157.49424743652344\n",
            "val_acc 0.07303619867743691\n",
            "epochs 731/1000\n",
            "train_loss 6.614852428436279\n",
            "train_acc 0.00031852263166981103\n",
            "valid_loss 157.47512817382812\n",
            "val_acc 0.07304241326492047\n",
            "epochs 732/1000\n",
            "train_loss 6.607364177703857\n",
            "train_acc 0.00031826028470487264\n",
            "valid_loss 157.45603942871094\n",
            "val_acc 0.07304869814252016\n",
            "epochs 733/1000\n",
            "train_loss 6.599872589111328\n",
            "train_acc 0.00031799795453060564\n",
            "valid_loss 157.43701171875\n",
            "val_acc 0.07305500062066131\n",
            "epochs 734/1000\n",
            "train_loss 6.592381000518799\n",
            "train_acc 0.00031773587901485513\n",
            "valid_loss 157.41799926757812\n",
            "val_acc 0.07306133960228181\n",
            "epochs 735/1000\n",
            "train_loss 6.584897041320801\n",
            "train_acc 0.00031747423166122573\n",
            "valid_loss 157.39906311035156\n",
            "val_acc 0.073067730339244\n",
            "epochs 736/1000\n",
            "train_loss 6.577416896820068\n",
            "train_acc 0.00031721324641973927\n",
            "valid_loss 157.38011169433594\n",
            "val_acc 0.07307413607341004\n",
            "epochs 737/1000\n",
            "train_loss 6.5699381828308105\n",
            "train_acc 0.0003169528068750739\n",
            "valid_loss 157.3612060546875\n",
            "val_acc 0.0730805720508215\n",
            "epochs 738/1000\n",
            "train_loss 6.562475681304932\n",
            "train_acc 0.00031669317272294743\n",
            "valid_loss 157.34231567382812\n",
            "val_acc 0.07308704177412437\n",
            "epochs 739/1000\n",
            "train_loss 6.554934978485107\n",
            "train_acc 0.0003164324925119917\n",
            "valid_loss 157.32327270507812\n",
            "val_acc 0.07309342043008656\n",
            "epochs 740/1000\n",
            "train_loss 6.5474934577941895\n",
            "train_acc 0.0003161759926185298\n",
            "valid_loss 157.30422973632812\n",
            "val_acc 0.0730998059589183\n",
            "epochs 741/1000\n",
            "train_loss 6.5400776863098145\n",
            "train_acc 0.0003159207732936081\n",
            "valid_loss 157.2852020263672\n",
            "val_acc 0.07310622108925599\n",
            "epochs 742/1000\n",
            "train_loss 6.532670497894287\n",
            "train_acc 0.000315666142761564\n",
            "valid_loss 157.26622009277344\n",
            "val_acc 0.07311267668410437\n",
            "epochs 743/1000\n",
            "train_loss 6.525269985198975\n",
            "train_acc 0.0003154119487869768\n",
            "valid_loss 157.24729919433594\n",
            "val_acc 0.07311920057836688\n",
            "epochs 744/1000\n",
            "train_loss 6.517897605895996\n",
            "train_acc 0.0003151582999495216\n",
            "valid_loss 157.22848510742188\n",
            "val_acc 0.07312576099939179\n",
            "epochs 745/1000\n",
            "train_loss 6.510539531707764\n",
            "train_acc 0.0003149045604424408\n",
            "valid_loss 157.20974731445312\n",
            "val_acc 0.07313237979833502\n",
            "epochs 746/1000\n",
            "train_loss 6.503169536590576\n",
            "train_acc 0.0003146508606732823\n",
            "valid_loss 157.19102478027344\n",
            "val_acc 0.07313899642176694\n",
            "epochs 747/1000\n",
            "train_loss 6.4958086013793945\n",
            "train_acc 0.00031439797301293135\n",
            "valid_loss 157.1723175048828\n",
            "val_acc 0.07314564195403364\n",
            "epochs 748/1000\n",
            "train_loss 6.488448619842529\n",
            "train_acc 0.00031414575250192475\n",
            "valid_loss 157.15357971191406\n",
            "val_acc 0.0731523027701769\n",
            "epochs 749/1000\n",
            "train_loss 6.481098175048828\n",
            "train_acc 0.00031389434857723804\n",
            "valid_loss 157.13490295410156\n",
            "val_acc 0.0731589971925132\n",
            "epochs 750/1000\n",
            "train_loss 6.473758220672607\n",
            "train_acc 0.00031364362411505473\n",
            "valid_loss 157.1162109375\n",
            "val_acc 0.07316569248796441\n",
            "epochs 751/1000\n",
            "train_loss 6.466446399688721\n",
            "train_acc 0.0003133939708977078\n",
            "valid_loss 157.09765625\n",
            "val_acc 0.07317246354941745\n",
            "epochs 752/1000\n",
            "train_loss 6.459145545959473\n",
            "train_acc 0.00031314414081862196\n",
            "valid_loss 157.07920837402344\n",
            "val_acc 0.07317930705321488\n",
            "epochs 753/1000\n",
            "train_loss 6.451828479766846\n",
            "train_acc 0.00031289395701605827\n",
            "valid_loss 157.06082153320312\n",
            "val_acc 0.07318617848068243\n",
            "epochs 754/1000\n",
            "train_loss 6.444500923156738\n",
            "train_acc 0.00031264401667823014\n",
            "valid_loss 157.0424346923828\n",
            "val_acc 0.07319305014389102\n",
            "epochs 755/1000\n",
            "train_loss 6.437172889709473\n",
            "train_acc 0.0003123950099017328\n",
            "valid_loss 157.0239715576172\n",
            "val_acc 0.07319991380500142\n",
            "epochs 756/1000\n",
            "train_loss 6.429786205291748\n",
            "train_acc 0.00031214591245560977\n",
            "valid_loss 157.00535583496094\n",
            "val_acc 0.07320669281471055\n",
            "epochs 757/1000\n",
            "train_loss 6.422496795654297\n",
            "train_acc 0.00031190124550798477\n",
            "valid_loss 156.98670959472656\n",
            "val_acc 0.07321345264208504\n",
            "epochs 758/1000\n",
            "train_loss 6.415167808532715\n",
            "train_acc 0.00031165685336768196\n",
            "valid_loss 156.9678497314453\n",
            "val_acc 0.07322011720243608\n",
            "epochs 759/1000\n",
            "train_loss 6.407928943634033\n",
            "train_acc 0.0003114164540490422\n",
            "valid_loss 156.948974609375\n",
            "val_acc 0.07322677307238337\n",
            "epochs 760/1000\n",
            "train_loss 6.400716781616211\n",
            "train_acc 0.00031117749760876625\n",
            "valid_loss 156.9300994873047\n",
            "val_acc 0.07323344178876141\n",
            "epochs 761/1000\n",
            "train_loss 6.393494606018066\n",
            "train_acc 0.0003109387236271197\n",
            "valid_loss 156.91119384765625\n",
            "val_acc 0.07324010113952681\n",
            "epochs 762/1000\n",
            "train_loss 6.386295795440674\n",
            "train_acc 0.0003107012369302818\n",
            "valid_loss 156.8922882080078\n",
            "val_acc 0.07324678684090032\n",
            "epochs 763/1000\n",
            "train_loss 6.3791093826293945\n",
            "train_acc 0.0003104644722323148\n",
            "valid_loss 156.8734130859375\n",
            "val_acc 0.07325350460159825\n",
            "epochs 764/1000\n",
            "train_loss 6.371928691864014\n",
            "train_acc 0.0003102282039785328\n",
            "valid_loss 156.8545684814453\n",
            "val_acc 0.07326025273941923\n",
            "epochs 765/1000\n",
            "train_loss 6.364757061004639\n",
            "train_acc 0.0003099924411239604\n",
            "valid_loss 156.8357391357422\n",
            "val_acc 0.07326704181323293\n",
            "epochs 766/1000\n",
            "train_loss 6.3575334548950195\n",
            "train_acc 0.000309755682582573\n",
            "valid_loss 156.81687927246094\n",
            "val_acc 0.07327381102659274\n",
            "epochs 767/1000\n",
            "train_loss 6.35036039352417\n",
            "train_acc 0.00030952162454083847\n",
            "valid_loss 156.7979736328125\n",
            "val_acc 0.07328058961284115\n",
            "epochs 768/1000\n",
            "train_loss 6.343210697174072\n",
            "train_acc 0.00030928836797381967\n",
            "valid_loss 156.77915954589844\n",
            "val_acc 0.07328741195233306\n",
            "epochs 769/1000\n",
            "train_loss 6.336064338684082\n",
            "train_acc 0.00030905537502034207\n",
            "valid_loss 156.76039123535156\n",
            "val_acc 0.07329427872609813\n",
            "epochs 770/1000\n",
            "train_loss 6.328932285308838\n",
            "train_acc 0.000308822767132929\n",
            "valid_loss 156.74168395996094\n",
            "val_acc 0.07330120011465624\n",
            "epochs 771/1000\n",
            "train_loss 6.321788311004639\n",
            "train_acc 0.0003085900148457418\n",
            "valid_loss 156.72299194335938\n",
            "val_acc 0.07330816582543775\n",
            "epochs 772/1000\n",
            "train_loss 6.314651012420654\n",
            "train_acc 0.000308357706951658\n",
            "valid_loss 156.704345703125\n",
            "val_acc 0.07331514959805645\n",
            "epochs 773/1000\n",
            "train_loss 6.307519912719727\n",
            "train_acc 0.0003081259699404025\n",
            "valid_loss 156.6857147216797\n",
            "val_acc 0.07332216780050657\n",
            "epochs 774/1000\n",
            "train_loss 6.300424098968506\n",
            "train_acc 0.0003078952750701529\n",
            "valid_loss 156.6671600341797\n",
            "val_acc 0.07332925722002984\n",
            "epochs 775/1000\n",
            "train_loss 6.293299198150635\n",
            "train_acc 0.0003076634087707274\n",
            "valid_loss 156.64865112304688\n",
            "val_acc 0.07333634433161933\n",
            "epochs 776/1000\n",
            "train_loss 6.286200046539307\n",
            "train_acc 0.00030743278610036493\n",
            "valid_loss 156.63021850585938\n",
            "val_acc 0.0733434860536363\n",
            "epochs 777/1000\n",
            "train_loss 6.279102325439453\n",
            "train_acc 0.0003072024438342151\n",
            "valid_loss 156.61178588867188\n",
            "val_acc 0.07335066358500626\n",
            "epochs 778/1000\n",
            "train_loss 6.272032260894775\n",
            "train_acc 0.0003069727659189644\n",
            "valid_loss 156.5934295654297\n",
            "val_acc 0.07335790562210605\n",
            "epochs 779/1000\n",
            "train_loss 6.264919281005859\n",
            "train_acc 0.00030674214716642524\n",
            "valid_loss 156.5750274658203\n",
            "val_acc 0.07336509478918743\n",
            "epochs 780/1000\n",
            "train_loss 6.257839679718018\n",
            "train_acc 0.0003065138544815664\n",
            "valid_loss 156.55654907226562\n",
            "val_acc 0.07337225847732043\n",
            "epochs 781/1000\n",
            "train_loss 6.250777721405029\n",
            "train_acc 0.00030628672818868206\n",
            "valid_loss 156.5380859375\n",
            "val_acc 0.07337946653715335\n",
            "epochs 782/1000\n",
            "train_loss 6.243727207183838\n",
            "train_acc 0.00030606044702625903\n",
            "valid_loss 156.51963806152344\n",
            "val_acc 0.07338669689488597\n",
            "epochs 783/1000\n",
            "train_loss 6.236696720123291\n",
            "train_acc 0.00030583492424249504\n",
            "valid_loss 156.50120544433594\n",
            "val_acc 0.07339394246955636\n",
            "epochs 784/1000\n",
            "train_loss 6.229664325714111\n",
            "train_acc 0.00030560949492680194\n",
            "valid_loss 156.48281860351562\n",
            "val_acc 0.073401232388278\n",
            "epochs 785/1000\n",
            "train_loss 6.222634792327881\n",
            "train_acc 0.00030538446075157623\n",
            "valid_loss 156.46444702148438\n",
            "val_acc 0.07340856382070342\n",
            "epochs 786/1000\n",
            "train_loss 6.215611457824707\n",
            "train_acc 0.0003051598955957721\n",
            "valid_loss 156.4460906982422\n",
            "val_acc 0.0734159188985359\n",
            "epochs 787/1000\n",
            "train_loss 6.20859956741333\n",
            "train_acc 0.00030493592147160176\n",
            "valid_loss 156.42776489257812\n",
            "val_acc 0.07342330250248778\n",
            "epochs 788/1000\n",
            "train_loss 6.2016167640686035\n",
            "train_acc 0.0003047128366933276\n",
            "valid_loss 156.4095458984375\n",
            "val_acc 0.07343076416727853\n",
            "epochs 789/1000\n",
            "train_loss 6.1945719718933105\n",
            "train_acc 0.0003044882497096506\n",
            "valid_loss 156.39125061035156\n",
            "val_acc 0.0734381835005479\n",
            "epochs 790/1000\n",
            "train_loss 6.187597274780273\n",
            "train_acc 0.00030426679810401623\n",
            "valid_loss 156.37295532226562\n",
            "val_acc 0.07344563199294499\n",
            "epochs 791/1000\n",
            "train_loss 6.18062162399292\n",
            "train_acc 0.00030404585749448207\n",
            "valid_loss 156.35464477539062\n",
            "val_acc 0.07345305808121338\n",
            "epochs 792/1000\n",
            "train_loss 6.173621654510498\n",
            "train_acc 0.0003038256360853736\n",
            "valid_loss 156.33621215820312\n",
            "val_acc 0.07346044036239618\n",
            "epochs 793/1000\n",
            "train_loss 6.166683673858643\n",
            "train_acc 0.0003036081196534304\n",
            "valid_loss 156.31777954101562\n",
            "val_acc 0.07346784732944797\n",
            "epochs 794/1000\n",
            "train_loss 6.15975284576416\n",
            "train_acc 0.0003033912004097007\n",
            "valid_loss 156.2993621826172\n",
            "val_acc 0.07347528943646467\n",
            "epochs 795/1000\n",
            "train_loss 6.152829170227051\n",
            "train_acc 0.00030317489626423386\n",
            "valid_loss 156.2809600830078\n",
            "val_acc 0.07348274232062976\n",
            "epochs 796/1000\n",
            "train_loss 6.145913600921631\n",
            "train_acc 0.0003029591462109238\n",
            "valid_loss 156.26255798339844\n",
            "val_acc 0.07349022967682686\n",
            "epochs 797/1000\n",
            "train_loss 6.13900899887085\n",
            "train_acc 0.00030274391554904956\n",
            "valid_loss 156.2441864013672\n",
            "val_acc 0.07349774612666807\n",
            "epochs 798/1000\n",
            "train_loss 6.1321120262146\n",
            "train_acc 0.00030252907331137417\n",
            "valid_loss 156.22584533691406\n",
            "val_acc 0.07350529151008231\n",
            "epochs 799/1000\n",
            "train_loss 6.125246524810791\n",
            "train_acc 0.00030231482938129024\n",
            "valid_loss 156.20758056640625\n",
            "val_acc 0.07351288483914686\n",
            "epochs 800/1000\n",
            "train_loss 6.1183762550354\n",
            "train_acc 0.0003021000794923076\n",
            "valid_loss 156.18942260742188\n",
            "val_acc 0.07352054352086271\n",
            "epochs 801/1000\n",
            "train_loss 6.111530780792236\n",
            "train_acc 0.00030188581205528375\n",
            "valid_loss 156.17137145996094\n",
            "val_acc 0.07352827843860724\n",
            "epochs 802/1000\n",
            "train_loss 6.104745388031006\n",
            "train_acc 0.0003016721921784875\n",
            "valid_loss 156.15347290039062\n",
            "val_acc 0.07353608147095655\n",
            "epochs 803/1000\n",
            "train_loss 6.097965717315674\n",
            "train_acc 0.0003014574988180986\n",
            "valid_loss 156.1356964111328\n",
            "val_acc 0.07354396910232026\n",
            "epochs 804/1000\n",
            "train_loss 6.091160297393799\n",
            "train_acc 0.0003012418909258066\n",
            "valid_loss 156.11801147460938\n",
            "val_acc 0.07355190281523391\n",
            "epochs 805/1000\n",
            "train_loss 6.084341526031494\n",
            "train_acc 0.00030102627967538027\n",
            "valid_loss 156.1002960205078\n",
            "val_acc 0.073559825331904\n",
            "epochs 806/1000\n",
            "train_loss 6.077549457550049\n",
            "train_acc 0.0003008118723160945\n",
            "valid_loss 156.08255004882812\n",
            "val_acc 0.07356777328968747\n",
            "epochs 807/1000\n",
            "train_loss 6.0707597732543945\n",
            "train_acc 0.0003005980308024356\n",
            "valid_loss 156.0648651123047\n",
            "val_acc 0.07357574762572767\n",
            "epochs 808/1000\n",
            "train_loss 6.063992023468018\n",
            "train_acc 0.0003003849084892024\n",
            "valid_loss 156.0471954345703\n",
            "val_acc 0.07358376123738708\n",
            "epochs 809/1000\n",
            "train_loss 6.0572404861450195\n",
            "train_acc 0.0003001719826268247\n",
            "valid_loss 156.02960205078125\n",
            "val_acc 0.07359184005181305\n",
            "epochs 810/1000\n",
            "train_loss 6.050495147705078\n",
            "train_acc 0.0002999591933285746\n",
            "valid_loss 156.01206970214844\n",
            "val_acc 0.07359996276500169\n",
            "epochs 811/1000\n",
            "train_loss 6.0437493324279785\n",
            "train_acc 0.00029974651540844486\n",
            "valid_loss 155.9945526123047\n",
            "val_acc 0.07360809088859241\n",
            "epochs 812/1000\n",
            "train_loss 6.037017345428467\n",
            "train_acc 0.00029953433785032336\n",
            "valid_loss 155.97705078125\n",
            "val_acc 0.07361623704637169\n",
            "epochs 813/1000\n",
            "train_loss 6.03030252456665\n",
            "train_acc 0.0002993228190462105\n",
            "valid_loss 155.95957946777344\n",
            "val_acc 0.07362442709854804\n",
            "epochs 814/1000\n",
            "train_loss 6.0236406326293945\n",
            "train_acc 0.00029911198026429\n",
            "valid_loss 155.94227600097656\n",
            "val_acc 0.0736327043719939\n",
            "epochs 815/1000\n",
            "train_loss 6.017023086547852\n",
            "train_acc 0.00029890093047959875\n",
            "valid_loss 155.9252166748047\n",
            "val_acc 0.07364111381175462\n",
            "epochs 816/1000\n",
            "train_loss 6.010349750518799\n",
            "train_acc 0.000298687968237433\n",
            "valid_loss 155.9081268310547\n",
            "val_acc 0.07364951140189077\n",
            "epochs 817/1000\n",
            "train_loss 6.003691673278809\n",
            "train_acc 0.000298475962503849\n",
            "valid_loss 155.8910675048828\n",
            "val_acc 0.07365792732452973\n",
            "epochs 818/1000\n",
            "train_loss 5.996976852416992\n",
            "train_acc 0.0002982636472622219\n",
            "valid_loss 155.8738250732422\n",
            "val_acc 0.07366624847782077\n",
            "epochs 819/1000\n",
            "train_loss 5.9903883934021\n",
            "train_acc 0.0002980556617750643\n",
            "valid_loss 155.8566131591797\n",
            "val_acc 0.07367460049281363\n",
            "epochs 820/1000\n",
            "train_loss 5.983817100524902\n",
            "train_acc 0.0002978480854206002\n",
            "valid_loss 155.83949279785156\n",
            "val_acc 0.07368296327476856\n",
            "epochs 821/1000\n",
            "train_loss 5.977241039276123\n",
            "train_acc 0.0002976405930194932\n",
            "valid_loss 155.82232666015625\n",
            "val_acc 0.07369134291366208\n",
            "epochs 822/1000\n",
            "train_loss 5.970676422119141\n",
            "train_acc 0.0002974337951924938\n",
            "valid_loss 155.80520629882812\n",
            "val_acc 0.07369976221525576\n",
            "epochs 823/1000\n",
            "train_loss 5.964112281799316\n",
            "train_acc 0.000297227180943502\n",
            "valid_loss 155.78807067871094\n",
            "val_acc 0.07370819298666902\n",
            "epochs 824/1000\n",
            "train_loss 5.957569599151611\n",
            "train_acc 0.00029702135137855436\n",
            "valid_loss 155.77101135253906\n",
            "val_acc 0.07371666877879761\n",
            "epochs 825/1000\n",
            "train_loss 5.951017379760742\n",
            "train_acc 0.00029681542330833437\n",
            "valid_loss 155.75390625\n",
            "val_acc 0.07372515675524483\n",
            "epochs 826/1000\n",
            "train_loss 5.944468021392822\n",
            "train_acc 0.00029661006276280835\n",
            "valid_loss 155.73684692382812\n",
            "val_acc 0.07373363422957482\n",
            "epochs 827/1000\n",
            "train_loss 5.937920570373535\n",
            "train_acc 0.0002964056665615107\n",
            "valid_loss 155.7196807861328\n",
            "val_acc 0.07374210089037661\n",
            "epochs 828/1000\n",
            "train_loss 5.931346416473389\n",
            "train_acc 0.000296201779117557\n",
            "valid_loss 155.702392578125\n",
            "val_acc 0.07375048568646889\n",
            "epochs 829/1000\n",
            "train_loss 5.924815654754639\n",
            "train_acc 0.00029600075056525663\n",
            "valid_loss 155.6849822998047\n",
            "val_acc 0.07375881304178619\n",
            "epochs 830/1000\n",
            "train_loss 5.918315410614014\n",
            "train_acc 0.00029580129977638047\n",
            "valid_loss 155.66763305664062\n",
            "val_acc 0.07376717073784675\n",
            "epochs 831/1000\n",
            "train_loss 5.911850929260254\n",
            "train_acc 0.0002956025709863752\n",
            "valid_loss 155.65032958984375\n",
            "val_acc 0.07377558067673817\n",
            "epochs 832/1000\n",
            "train_loss 5.905370712280273\n",
            "train_acc 0.0002954031493013295\n",
            "valid_loss 155.63304138183594\n",
            "val_acc 0.07378399171866476\n",
            "epochs 833/1000\n",
            "train_loss 5.898931980133057\n",
            "train_acc 0.0002952054066834255\n",
            "valid_loss 155.61578369140625\n",
            "val_acc 0.07379242047754815\n",
            "epochs 834/1000\n",
            "train_loss 5.892495632171631\n",
            "train_acc 0.00029500809278733167\n",
            "valid_loss 155.59854125976562\n",
            "val_acc 0.07380087819328765\n",
            "epochs 835/1000\n",
            "train_loss 5.886072158813477\n",
            "train_acc 0.0002948114969722855\n",
            "valid_loss 155.581298828125\n",
            "val_acc 0.07380934615066508\n",
            "epochs 836/1000\n",
            "train_loss 5.879662990570068\n",
            "train_acc 0.0002946152565597843\n",
            "valid_loss 155.56410217285156\n",
            "val_acc 0.07381784997705836\n",
            "epochs 837/1000\n",
            "train_loss 5.8732829093933105\n",
            "train_acc 0.000294419523785249\n",
            "valid_loss 155.5469970703125\n",
            "val_acc 0.07382639782736078\n",
            "epochs 838/1000\n",
            "train_loss 5.866896629333496\n",
            "train_acc 0.0002942235856048333\n",
            "valid_loss 155.52992248535156\n",
            "val_acc 0.07383500758733134\n",
            "epochs 839/1000\n",
            "train_loss 5.860527038574219\n",
            "train_acc 0.000294027936223966\n",
            "valid_loss 155.51295471191406\n",
            "val_acc 0.07384367982158438\n",
            "epochs 840/1000\n",
            "train_loss 5.854155540466309\n",
            "train_acc 0.0002938320982278898\n",
            "valid_loss 155.49598693847656\n",
            "val_acc 0.07385238711722195\n",
            "epochs 841/1000\n",
            "train_loss 5.847777366638184\n",
            "train_acc 0.0002936363436254816\n",
            "valid_loss 155.4790496826172\n",
            "val_acc 0.07386110026127427\n",
            "epochs 842/1000\n",
            "train_loss 5.84141731262207\n",
            "train_acc 0.000293441295350651\n",
            "valid_loss 155.46214294433594\n",
            "val_acc 0.07386986299534329\n",
            "epochs 843/1000\n",
            "train_loss 5.8351030349731445\n",
            "train_acc 0.0002932469500452638\n",
            "valid_loss 155.4453582763672\n",
            "val_acc 0.0738786611429532\n",
            "epochs 844/1000\n",
            "train_loss 5.828774452209473\n",
            "train_acc 0.0002930521642645964\n",
            "valid_loss 155.4285888671875\n",
            "val_acc 0.0738875092822127\n",
            "epochs 845/1000\n",
            "train_loss 5.82244873046875\n",
            "train_acc 0.00029285747195199993\n",
            "valid_loss 155.41188049316406\n",
            "val_acc 0.07389639224566054\n",
            "epochs 846/1000\n",
            "train_loss 5.816143989562988\n",
            "train_acc 0.0002926633387687616\n",
            "valid_loss 155.395263671875\n",
            "val_acc 0.0739053254656028\n",
            "epochs 847/1000\n",
            "train_loss 5.809874534606934\n",
            "train_acc 0.0002924695632268245\n",
            "valid_loss 155.3787384033203\n",
            "val_acc 0.07391429511044408\n",
            "epochs 848/1000\n",
            "train_loss 5.803601264953613\n",
            "train_acc 0.000292275676306767\n",
            "valid_loss 155.36224365234375\n",
            "val_acc 0.07392328843125143\n",
            "epochs 849/1000\n",
            "train_loss 5.797330856323242\n",
            "train_acc 0.0002920821537442792\n",
            "valid_loss 155.3457794189453\n",
            "val_acc 0.07393233272305225\n",
            "epochs 850/1000\n",
            "train_loss 5.791075229644775\n",
            "train_acc 0.000291889052627644\n",
            "valid_loss 155.329345703125\n",
            "val_acc 0.07394140743417665\n",
            "epochs 851/1000\n",
            "train_loss 5.784748554229736\n",
            "train_acc 0.0002916952916376221\n",
            "valid_loss 155.3126983642578\n",
            "val_acc 0.07395034688524901\n",
            "epochs 852/1000\n",
            "train_loss 5.7784953117370605\n",
            "train_acc 0.0002915053818679343\n",
            "valid_loss 155.2959747314453\n",
            "val_acc 0.07395926932949806\n",
            "epochs 853/1000\n",
            "train_loss 5.772299289703369\n",
            "train_acc 0.0002913172160893177\n",
            "valid_loss 155.27935791015625\n",
            "val_acc 0.0739682317856932\n",
            "epochs 854/1000\n",
            "train_loss 5.766124248504639\n",
            "train_acc 0.0002911292809025886\n",
            "valid_loss 155.26280212402344\n",
            "val_acc 0.07397725808259566\n",
            "epochs 855/1000\n",
            "train_loss 5.75996732711792\n",
            "train_acc 0.0002909411302355763\n",
            "valid_loss 155.2463836669922\n",
            "val_acc 0.07398635969730094\n",
            "epochs 856/1000\n",
            "train_loss 5.753793716430664\n",
            "train_acc 0.0002907527254697365\n",
            "valid_loss 155.2299346923828\n",
            "val_acc 0.07399547111417633\n",
            "epochs 857/1000\n",
            "train_loss 5.747618198394775\n",
            "train_acc 0.00029056461398095754\n",
            "valid_loss 155.2135009765625\n",
            "val_acc 0.07400459780037635\n",
            "epochs 858/1000\n",
            "train_loss 5.741459369659424\n",
            "train_acc 0.0002903774657170288\n",
            "valid_loss 155.1970672607422\n",
            "val_acc 0.07401372231106507\n",
            "epochs 859/1000\n",
            "train_loss 5.735347747802734\n",
            "train_acc 0.00029019125605163234\n",
            "valid_loss 155.18075561523438\n",
            "val_acc 0.0740229362359969\n",
            "epochs 860/1000\n",
            "train_loss 5.729247570037842\n",
            "train_acc 0.00029000474471383943\n",
            "valid_loss 155.16452026367188\n",
            "val_acc 0.07403219272382558\n",
            "epochs 861/1000\n",
            "train_loss 5.72313117980957\n",
            "train_acc 0.0002898178091317487\n",
            "valid_loss 155.1483154296875\n",
            "val_acc 0.07404146624467103\n",
            "epochs 862/1000\n",
            "train_loss 5.717026710510254\n",
            "train_acc 0.0002896315865935041\n",
            "valid_loss 155.1321258544922\n",
            "val_acc 0.07405076465074671\n",
            "epochs 863/1000\n",
            "train_loss 5.7109293937683105\n",
            "train_acc 0.0002894458347537483\n",
            "valid_loss 155.11593627929688\n",
            "val_acc 0.07406008572725113\n",
            "epochs 864/1000\n",
            "train_loss 5.704816818237305\n",
            "train_acc 0.00028926011369689007\n",
            "valid_loss 155.09970092773438\n",
            "val_acc 0.07406938346393872\n",
            "epochs 865/1000\n",
            "train_loss 5.698708534240723\n",
            "train_acc 0.000289075862943159\n",
            "valid_loss 155.0833740234375\n",
            "val_acc 0.0740786144102458\n",
            "epochs 866/1000\n",
            "train_loss 5.692678451538086\n",
            "train_acc 0.0002888938218417864\n",
            "valid_loss 155.067138671875\n",
            "val_acc 0.07408789859327954\n",
            "epochs 867/1000\n",
            "train_loss 5.686657428741455\n",
            "train_acc 0.0002887113967937274\n",
            "valid_loss 155.051025390625\n",
            "val_acc 0.07409727874037345\n",
            "epochs 868/1000\n",
            "train_loss 5.680627346038818\n",
            "train_acc 0.0002885286739910953\n",
            "valid_loss 155.0349578857422\n",
            "val_acc 0.07410669265809702\n",
            "epochs 869/1000\n",
            "train_loss 5.674591064453125\n",
            "train_acc 0.00028834598532949504\n",
            "valid_loss 155.0188751220703\n",
            "val_acc 0.07411612581636291\n",
            "epochs 870/1000\n",
            "train_loss 5.668501377105713\n",
            "train_acc 0.00028816321943080623\n",
            "valid_loss 155.0026092529297\n",
            "val_acc 0.07412543891841779\n",
            "epochs 871/1000\n",
            "train_loss 5.662510871887207\n",
            "train_acc 0.0002879842672532854\n",
            "valid_loss 154.986328125\n",
            "val_acc 0.07413474884233437\n",
            "epochs 872/1000\n",
            "train_loss 5.65653133392334\n",
            "train_acc 0.0002878062312867349\n",
            "valid_loss 154.9700469970703\n",
            "val_acc 0.07414405317540514\n",
            "epochs 873/1000\n",
            "train_loss 5.6505656242370605\n",
            "train_acc 0.00028762898728018627\n",
            "valid_loss 154.9537353515625\n",
            "val_acc 0.07415338034916204\n",
            "epochs 874/1000\n",
            "train_loss 5.644632816314697\n",
            "train_acc 0.0002874524641531305\n",
            "valid_loss 154.9375762939453\n",
            "val_acc 0.07416277357988292\n",
            "epochs 875/1000\n",
            "train_loss 5.63872766494751\n",
            "train_acc 0.0002872762482953616\n",
            "valid_loss 154.92149353027344\n",
            "val_acc 0.07417220903153066\n",
            "epochs 876/1000\n",
            "train_loss 5.632810592651367\n",
            "train_acc 0.0002870996300211678\n",
            "valid_loss 154.90542602539062\n",
            "val_acc 0.07418169706943445\n",
            "epochs 877/1000\n",
            "train_loss 5.626904010772705\n",
            "train_acc 0.0002869231203266491\n",
            "valid_loss 154.8894500732422\n",
            "val_acc 0.07419120364502305\n",
            "epochs 878/1000\n",
            "train_loss 5.621003150939941\n",
            "train_acc 0.0002867465664166957\n",
            "valid_loss 154.87350463867188\n",
            "val_acc 0.07420074263063725\n",
            "epochs 879/1000\n",
            "train_loss 5.615143775939941\n",
            "train_acc 0.0002865708934573027\n",
            "valid_loss 154.85769653320312\n",
            "val_acc 0.07421033475256991\n",
            "epochs 880/1000\n",
            "train_loss 5.609279155731201\n",
            "train_acc 0.0002863949708765946\n",
            "valid_loss 154.84188842773438\n",
            "val_acc 0.07421995969489217\n",
            "epochs 881/1000\n",
            "train_loss 5.603419780731201\n",
            "train_acc 0.00028621935948299675\n",
            "valid_loss 154.82615661621094\n",
            "val_acc 0.074229617016681\n",
            "epochs 882/1000\n",
            "train_loss 5.597572326660156\n",
            "train_acc 0.0002860440643137106\n",
            "valid_loss 154.8104705810547\n",
            "val_acc 0.07423930733493762\n",
            "epochs 883/1000\n",
            "train_loss 5.591724395751953\n",
            "train_acc 0.0002858690070122695\n",
            "valid_loss 154.7947998046875\n",
            "val_acc 0.07424900144687854\n",
            "epochs 884/1000\n",
            "train_loss 5.585900783538818\n",
            "train_acc 0.0002856948194676079\n",
            "valid_loss 154.77920532226562\n",
            "val_acc 0.07425875312765129\n",
            "epochs 885/1000\n",
            "train_loss 5.580071926116943\n",
            "train_acc 0.00028552073490573093\n",
            "valid_loss 154.76356506347656\n",
            "val_acc 0.07426850729971193\n",
            "epochs 886/1000\n",
            "train_loss 5.574254989624023\n",
            "train_acc 0.0002853472402561098\n",
            "valid_loss 154.74801635742188\n",
            "val_acc 0.0742782949557295\n",
            "epochs 887/1000\n",
            "train_loss 5.568447113037109\n",
            "train_acc 0.0002851741138818817\n",
            "valid_loss 154.7324981689453\n",
            "val_acc 0.07428811445424799\n",
            "epochs 888/1000\n",
            "train_loss 5.562661647796631\n",
            "train_acc 0.00028500135970087006\n",
            "valid_loss 154.71705627441406\n",
            "val_acc 0.07429797826334834\n",
            "epochs 889/1000\n",
            "train_loss 5.5568695068359375\n",
            "train_acc 0.00028482856186411274\n",
            "valid_loss 154.7017364501953\n",
            "val_acc 0.07430786009354051\n",
            "epochs 890/1000\n",
            "train_loss 5.551076889038086\n",
            "train_acc 0.0002846560449912571\n",
            "valid_loss 154.68646240234375\n",
            "val_acc 0.0743177802345599\n",
            "epochs 891/1000\n",
            "train_loss 5.545257091522217\n",
            "train_acc 0.0002844834357697087\n",
            "valid_loss 154.67115783691406\n",
            "val_acc 0.07432762703829213\n",
            "epochs 892/1000\n",
            "train_loss 5.539554119110107\n",
            "train_acc 0.00028431352592843515\n",
            "valid_loss 154.65603637695312\n",
            "val_acc 0.07433754139929079\n",
            "epochs 893/1000\n",
            "train_loss 5.53383207321167\n",
            "train_acc 0.00028414315938089904\n",
            "valid_loss 154.64096069335938\n",
            "val_acc 0.0743474418122787\n",
            "epochs 894/1000\n",
            "train_loss 5.528149604797363\n",
            "train_acc 0.0002839740291864683\n",
            "valid_loss 154.62594604492188\n",
            "val_acc 0.07435737954510842\n",
            "epochs 895/1000\n",
            "train_loss 5.522498607635498\n",
            "train_acc 0.00028380526950618683\n",
            "valid_loss 154.61106872558594\n",
            "val_acc 0.07436738663673167\n",
            "epochs 896/1000\n",
            "train_loss 5.516841411590576\n",
            "train_acc 0.0002836362126310213\n",
            "valid_loss 154.5962371826172\n",
            "val_acc 0.0743774393646163\n",
            "epochs 897/1000\n",
            "train_loss 5.511170387268066\n",
            "train_acc 0.00028346723299294425\n",
            "valid_loss 154.58139038085938\n",
            "val_acc 0.07438749320572242\n",
            "epochs 898/1000\n",
            "train_loss 5.505501747131348\n",
            "train_acc 0.00028329885446090397\n",
            "valid_loss 154.5665283203125\n",
            "val_acc 0.07439754329825518\n",
            "epochs 899/1000\n",
            "train_loss 5.49986457824707\n",
            "train_acc 0.0002831315684418839\n",
            "valid_loss 154.55174255371094\n",
            "val_acc 0.07440765145001933\n",
            "epochs 900/1000\n",
            "train_loss 5.494174957275391\n",
            "train_acc 0.00028296335725686874\n",
            "valid_loss 154.53684997558594\n",
            "val_acc 0.07441770299774361\n",
            "epochs 901/1000\n",
            "train_loss 5.48854923248291\n",
            "train_acc 0.0002827979316442417\n",
            "valid_loss 154.52195739746094\n",
            "val_acc 0.07442776163225062\n",
            "epochs 902/1000\n",
            "train_loss 5.482954978942871\n",
            "train_acc 0.0002826333976162669\n",
            "valid_loss 154.50711059570312\n",
            "val_acc 0.07443784204733675\n",
            "epochs 903/1000\n",
            "train_loss 5.477364540100098\n",
            "train_acc 0.00028246916078317625\n",
            "valid_loss 154.49229431152344\n",
            "val_acc 0.0744479619039339\n",
            "epochs 904/1000\n",
            "train_loss 5.471774578094482\n",
            "train_acc 0.00028230517357140064\n",
            "valid_loss 154.47747802734375\n",
            "val_acc 0.07445809520941112\n",
            "epochs 905/1000\n",
            "train_loss 5.466248512268066\n",
            "train_acc 0.000282142429988688\n",
            "valid_loss 154.4628143310547\n",
            "val_acc 0.07446829522450571\n",
            "epochs 906/1000\n",
            "train_loss 5.460729598999023\n",
            "train_acc 0.00028197905339766294\n",
            "valid_loss 154.44825744628906\n",
            "val_acc 0.07447854600322898\n",
            "epochs 907/1000\n",
            "train_loss 5.4551920890808105\n",
            "train_acc 0.0002818152105856615\n",
            "valid_loss 154.43373107910156\n",
            "val_acc 0.07448882522585336\n",
            "epochs 908/1000\n",
            "train_loss 5.449653625488281\n",
            "train_acc 0.00028165168119952653\n",
            "valid_loss 154.41921997070312\n",
            "val_acc 0.07449911957519362\n",
            "epochs 909/1000\n",
            "train_loss 5.44413948059082\n",
            "train_acc 0.0002814888950804463\n",
            "valid_loss 154.40476989746094\n",
            "val_acc 0.07450945604869048\n",
            "epochs 910/1000\n",
            "train_loss 5.43862771987915\n",
            "train_acc 0.0002813262757487022\n",
            "valid_loss 154.39035034179688\n",
            "val_acc 0.07451982355778455\n",
            "epochs 911/1000\n",
            "train_loss 5.433120250701904\n",
            "train_acc 0.00028116403028924164\n",
            "valid_loss 154.3759307861328\n",
            "val_acc 0.07453020676257438\n",
            "epochs 912/1000\n",
            "train_loss 5.4276041984558105\n",
            "train_acc 0.0002810022404166655\n",
            "valid_loss 154.36148071289062\n",
            "val_acc 0.07454058112198254\n",
            "epochs 913/1000\n",
            "train_loss 5.4220428466796875\n",
            "train_acc 0.0002808408764674543\n",
            "valid_loss 154.34683227539062\n",
            "val_acc 0.07455084502362297\n",
            "epochs 914/1000\n",
            "train_loss 5.416553020477295\n",
            "train_acc 0.00028068276095347334\n",
            "valid_loss 154.33213806152344\n",
            "val_acc 0.07456107211328344\n",
            "epochs 915/1000\n",
            "train_loss 5.411104679107666\n",
            "train_acc 0.0002805263273050794\n",
            "valid_loss 154.31744384765625\n",
            "val_acc 0.07457132386698503\n",
            "epochs 916/1000\n",
            "train_loss 5.405706882476807\n",
            "train_acc 0.00028037081042734475\n",
            "valid_loss 154.3029022216797\n",
            "val_acc 0.07458165889765951\n",
            "epochs 917/1000\n",
            "train_loss 5.40032434463501\n",
            "train_acc 0.0002802145651143152\n",
            "valid_loss 154.2884979248047\n",
            "val_acc 0.07459208555155783\n",
            "epochs 918/1000\n",
            "train_loss 5.394931316375732\n",
            "train_acc 0.00028005780404782854\n",
            "valid_loss 154.27415466308594\n",
            "val_acc 0.0746025466840365\n",
            "epochs 919/1000\n",
            "train_loss 5.38954496383667\n",
            "train_acc 0.00027990127105462865\n",
            "valid_loss 154.25985717773438\n",
            "val_acc 0.07461302177689504\n",
            "epochs 920/1000\n",
            "train_loss 5.384191036224365\n",
            "train_acc 0.0002797452420614158\n",
            "valid_loss 154.24571228027344\n",
            "val_acc 0.07462358449702151\n",
            "epochs 921/1000\n",
            "train_loss 5.378878116607666\n",
            "train_acc 0.0002795890848994112\n",
            "valid_loss 154.2317352294922\n",
            "val_acc 0.07463424788438715\n",
            "epochs 922/1000\n",
            "train_loss 5.373542308807373\n",
            "train_acc 0.0002794320347535316\n",
            "valid_loss 154.2178192138672\n",
            "val_acc 0.07464491967111826\n",
            "epochs 923/1000\n",
            "train_loss 5.368219375610352\n",
            "train_acc 0.0002792752983133631\n",
            "valid_loss 154.2040252685547\n",
            "val_acc 0.07465567587059922\n",
            "epochs 924/1000\n",
            "train_loss 5.362895488739014\n",
            "train_acc 0.00027911845385320845\n",
            "valid_loss 154.190185546875\n",
            "val_acc 0.07466635872115149\n",
            "epochs 925/1000\n",
            "train_loss 5.357573986053467\n",
            "train_acc 0.000278962530361381\n",
            "valid_loss 154.17630004882812\n",
            "val_acc 0.07467702161811758\n",
            "epochs 926/1000\n",
            "train_loss 5.352269172668457\n",
            "train_acc 0.0002788078572148851\n",
            "valid_loss 154.16241455078125\n",
            "val_acc 0.07468767171085347\n",
            "epochs 927/1000\n",
            "train_loss 5.346991062164307\n",
            "train_acc 0.00027865449569967145\n",
            "valid_loss 154.14849853515625\n",
            "val_acc 0.07469831814159988\n",
            "epochs 928/1000\n",
            "train_loss 5.34170389175415\n",
            "train_acc 0.0002785013868840626\n",
            "valid_loss 154.13458251953125\n",
            "val_acc 0.074708943580481\n",
            "epochs 929/1000\n",
            "train_loss 5.336429595947266\n",
            "train_acc 0.00027834944474027277\n",
            "valid_loss 154.12059020996094\n",
            "val_acc 0.07471953201384167\n",
            "epochs 930/1000\n",
            "train_loss 5.331192970275879\n",
            "train_acc 0.0002781987059279345\n",
            "valid_loss 154.106689453125\n",
            "val_acc 0.07473016629810444\n",
            "epochs 931/1000\n",
            "train_loss 5.325957298278809\n",
            "train_acc 0.00027804807961309474\n",
            "valid_loss 154.0928192138672\n",
            "val_acc 0.07474080533720553\n",
            "epochs 932/1000\n",
            "train_loss 5.3207316398620605\n",
            "train_acc 0.00027789794638430554\n",
            "valid_loss 154.07899475097656\n",
            "val_acc 0.07475148358935257\n",
            "epochs 933/1000\n",
            "train_loss 5.3155388832092285\n",
            "train_acc 0.0002777485720938645\n",
            "valid_loss 154.06529235839844\n",
            "val_acc 0.07476223120756913\n",
            "epochs 934/1000\n",
            "train_loss 5.310368537902832\n",
            "train_acc 0.00027759860481287783\n",
            "valid_loss 154.05174255371094\n",
            "val_acc 0.07477307683438994\n",
            "epochs 935/1000\n",
            "train_loss 5.305238246917725\n",
            "train_acc 0.0002774483073153533\n",
            "valid_loss 154.0384979248047\n",
            "val_acc 0.07478406498412368\n",
            "epochs 936/1000\n",
            "train_loss 5.300160884857178\n",
            "train_acc 0.0002772974571248946\n",
            "valid_loss 154.02549743652344\n",
            "val_acc 0.07479516233215691\n",
            "epochs 937/1000\n",
            "train_loss 5.295040130615234\n",
            "train_acc 0.00027714473505441745\n",
            "valid_loss 154.0125732421875\n",
            "val_acc 0.07480630318896146\n",
            "epochs 938/1000\n",
            "train_loss 5.2899017333984375\n",
            "train_acc 0.0002769918691438551\n",
            "valid_loss 153.9996795654297\n",
            "val_acc 0.07481745800032513\n",
            "epochs 939/1000\n",
            "train_loss 5.284754276275635\n",
            "train_acc 0.00027683948120773915\n",
            "valid_loss 153.98675537109375\n",
            "val_acc 0.07482856916249148\n",
            "epochs 940/1000\n",
            "train_loss 5.279646396636963\n",
            "train_acc 0.0002766887930472596\n",
            "valid_loss 153.9738006591797\n",
            "val_acc 0.0748396487957507\n",
            "epochs 941/1000\n",
            "train_loss 5.274567604064941\n",
            "train_acc 0.0002765391207224009\n",
            "valid_loss 153.96092224121094\n",
            "val_acc 0.0748507300100755\n",
            "epochs 942/1000\n",
            "train_loss 5.26951265335083\n",
            "train_acc 0.00027639008000663196\n",
            "valid_loss 153.9481201171875\n",
            "val_acc 0.07486185561865569\n",
            "epochs 943/1000\n",
            "train_loss 5.264503002166748\n",
            "train_acc 0.00027624146521422797\n",
            "valid_loss 153.935546875\n",
            "val_acc 0.07487308729687356\n",
            "epochs 944/1000\n",
            "train_loss 5.259464740753174\n",
            "train_acc 0.00027609163085939\n",
            "valid_loss 153.9230499267578\n",
            "val_acc 0.07488430546145537\n",
            "epochs 945/1000\n",
            "train_loss 5.2544732093811035\n",
            "train_acc 0.0002759434582213334\n",
            "valid_loss 153.91062927246094\n",
            "val_acc 0.07489554922794923\n",
            "epochs 946/1000\n",
            "train_loss 5.249489784240723\n",
            "train_acc 0.00027579585450719326\n",
            "valid_loss 153.8982391357422\n",
            "val_acc 0.07490680754635833\n",
            "epochs 947/1000\n",
            "train_loss 5.244513034820557\n",
            "train_acc 0.0002756488513394008\n",
            "valid_loss 153.88587951660156\n",
            "val_acc 0.07491805215540807\n",
            "epochs 948/1000\n",
            "train_loss 5.239582538604736\n",
            "train_acc 0.00027550257996503764\n",
            "valid_loss 153.8736572265625\n",
            "val_acc 0.07492934425026761\n",
            "epochs 949/1000\n",
            "train_loss 5.234626770019531\n",
            "train_acc 0.00027535589470062405\n",
            "valid_loss 153.8614959716797\n",
            "val_acc 0.07494061540055555\n",
            "epochs 950/1000\n",
            "train_loss 5.229616641998291\n",
            "train_acc 0.0002752096851719006\n",
            "valid_loss 153.84912109375\n",
            "val_acc 0.07495168337554788\n",
            "epochs 951/1000\n",
            "train_loss 5.224740028381348\n",
            "train_acc 0.00027506807824605715\n",
            "valid_loss 153.83677673339844\n",
            "val_acc 0.07496274157674634\n",
            "epochs 952/1000\n",
            "train_loss 5.219883918762207\n",
            "train_acc 0.000274927475682209\n",
            "valid_loss 153.8245849609375\n",
            "val_acc 0.07497382332949201\n",
            "epochs 953/1000\n",
            "train_loss 5.215072154998779\n",
            "train_acc 0.00027478766395898466\n",
            "valid_loss 153.8130340576172\n",
            "val_acc 0.07498497495325865\n",
            "epochs 954/1000\n",
            "train_loss 5.210282802581787\n",
            "train_acc 0.0002746476381546997\n",
            "valid_loss 153.8017578125\n",
            "val_acc 0.07499619566442561\n",
            "epochs 955/1000\n",
            "train_loss 5.205515384674072\n",
            "train_acc 0.00027450714976741717\n",
            "valid_loss 153.7908172607422\n",
            "val_acc 0.07500754416614655\n",
            "epochs 956/1000\n",
            "train_loss 5.200740337371826\n",
            "train_acc 0.00027436595169442275\n",
            "valid_loss 153.78004455566406\n",
            "val_acc 0.07501893584121717\n",
            "epochs 957/1000\n",
            "train_loss 5.195953369140625\n",
            "train_acc 0.00027422440801344166\n",
            "valid_loss 153.7693328857422\n",
            "val_acc 0.07503030324005522\n",
            "epochs 958/1000\n",
            "train_loss 5.191181182861328\n",
            "train_acc 0.0002740838751308799\n",
            "valid_loss 153.75865173339844\n",
            "val_acc 0.07504168220330029\n",
            "epochs 959/1000\n",
            "train_loss 5.186431884765625\n",
            "train_acc 0.0002739440371022703\n",
            "valid_loss 153.748046875\n",
            "val_acc 0.0750530963698111\n",
            "epochs 960/1000\n",
            "train_loss 5.181684970855713\n",
            "train_acc 0.00027380437145788723\n",
            "valid_loss 153.73751831054688\n",
            "val_acc 0.07506453975584008\n",
            "epochs 961/1000\n",
            "train_loss 5.176888942718506\n",
            "train_acc 0.0002736642656180685\n",
            "valid_loss 153.72691345214844\n",
            "val_acc 0.07507588636071887\n",
            "epochs 962/1000\n",
            "train_loss 5.172194480895996\n",
            "train_acc 0.000273527662592152\n",
            "valid_loss 153.71630859375\n",
            "val_acc 0.07508722002967261\n",
            "epochs 963/1000\n",
            "train_loss 5.167542934417725\n",
            "train_acc 0.00027339159267005295\n",
            "valid_loss 153.70587158203125\n",
            "val_acc 0.0750986162609479\n",
            "epochs 964/1000\n",
            "train_loss 5.162848949432373\n",
            "train_acc 0.00027325480158861657\n",
            "valid_loss 153.6953582763672\n",
            "val_acc 0.07510994586991729\n",
            "epochs 965/1000\n",
            "train_loss 5.158213138580322\n",
            "train_acc 0.00027312016167203325\n",
            "valid_loss 153.6848602294922\n",
            "val_acc 0.07512124751374358\n",
            "epochs 966/1000\n",
            "train_loss 5.153614521026611\n",
            "train_acc 0.0002729864799430988\n",
            "valid_loss 153.6743927001953\n",
            "val_acc 0.07513256166421342\n",
            "epochs 967/1000\n",
            "train_loss 5.149006366729736\n",
            "train_acc 0.0002728529395356488\n",
            "valid_loss 153.66397094726562\n",
            "val_acc 0.0751439020554244\n",
            "epochs 968/1000\n",
            "train_loss 5.1444091796875\n",
            "train_acc 0.00027272006319925334\n",
            "valid_loss 153.6535186767578\n",
            "val_acc 0.07515522084722762\n",
            "epochs 969/1000\n",
            "train_loss 5.139836311340332\n",
            "train_acc 0.0002725879407640045\n",
            "valid_loss 153.64315795898438\n",
            "val_acc 0.0751666127507633\n",
            "epochs 970/1000\n",
            "train_loss 5.135295867919922\n",
            "train_acc 0.0002724557315769534\n",
            "valid_loss 153.63304138183594\n",
            "val_acc 0.07517813225786085\n",
            "epochs 971/1000\n",
            "train_loss 5.130738735198975\n",
            "train_acc 0.000272322684815243\n",
            "valid_loss 153.62290954589844\n",
            "val_acc 0.07518963618276757\n",
            "epochs 972/1000\n",
            "train_loss 5.126119136810303\n",
            "train_acc 0.0002721896528852924\n",
            "valid_loss 153.612548828125\n",
            "val_acc 0.0752009737327171\n",
            "epochs 973/1000\n",
            "train_loss 5.121567726135254\n",
            "train_acc 0.0002720603140635201\n",
            "valid_loss 153.60202026367188\n",
            "val_acc 0.0752121863202774\n",
            "epochs 974/1000\n",
            "train_loss 5.117081642150879\n",
            "train_acc 0.0002719333175404105\n",
            "valid_loss 153.5915069580078\n",
            "val_acc 0.07522341051735566\n",
            "epochs 975/1000\n",
            "train_loss 5.112611770629883\n",
            "train_acc 0.00027180682473744336\n",
            "valid_loss 153.5810546875\n",
            "val_acc 0.07523469895095332\n",
            "epochs 976/1000\n",
            "train_loss 5.1081366539001465\n",
            "train_acc 0.0002716803935002715\n",
            "valid_loss 153.5706329345703\n",
            "val_acc 0.0752459975083184\n",
            "epochs 977/1000\n",
            "train_loss 5.103667736053467\n",
            "train_acc 0.00027155434005320643\n",
            "valid_loss 153.5602264404297\n",
            "val_acc 0.07525734199516591\n",
            "epochs 978/1000\n",
            "train_loss 5.099208354949951\n",
            "train_acc 0.0002714285689692658\n",
            "valid_loss 153.54991149902344\n",
            "val_acc 0.07526872796879616\n",
            "epochs 979/1000\n",
            "train_loss 5.09475564956665\n",
            "train_acc 0.000271303010007474\n",
            "valid_loss 153.5395965576172\n",
            "val_acc 0.0752801338028803\n",
            "epochs 980/1000\n",
            "train_loss 5.090249538421631\n",
            "train_acc 0.0002711772330467983\n",
            "valid_loss 153.5290985107422\n",
            "val_acc 0.07529137637611712\n",
            "epochs 981/1000\n",
            "train_loss 5.085837364196777\n",
            "train_acc 0.0002710545654386246\n",
            "valid_loss 153.51866149902344\n",
            "val_acc 0.07530265711102402\n",
            "epochs 982/1000\n",
            "train_loss 5.081438064575195\n",
            "train_acc 0.00027093209959835245\n",
            "valid_loss 153.50827026367188\n",
            "val_acc 0.07531395442056237\n",
            "epochs 983/1000\n",
            "train_loss 5.077063083648682\n",
            "train_acc 0.0002708104931606123\n",
            "valid_loss 153.4978790283203\n",
            "val_acc 0.0753252562826674\n",
            "epochs 984/1000\n",
            "train_loss 5.072713375091553\n",
            "train_acc 0.00027068948447077465\n",
            "valid_loss 153.48760986328125\n",
            "val_acc 0.07533661372581264\n",
            "epochs 985/1000\n",
            "train_loss 5.06839656829834\n",
            "train_acc 0.0002705686056287959\n",
            "valid_loss 153.4774932861328\n",
            "val_acc 0.07534804614042515\n",
            "epochs 986/1000\n",
            "train_loss 5.064111232757568\n",
            "train_acc 0.0002704474752065905\n",
            "valid_loss 153.46751403808594\n",
            "val_acc 0.07535958440203103\n",
            "epochs 987/1000\n",
            "train_loss 5.059842586517334\n",
            "train_acc 0.00027032559060339385\n",
            "valid_loss 153.4578094482422\n",
            "val_acc 0.0753712449062732\n",
            "epochs 988/1000\n",
            "train_loss 5.055549144744873\n",
            "train_acc 0.00027020274977146\n",
            "valid_loss 153.44810485839844\n",
            "val_acc 0.07538293858160613\n",
            "epochs 989/1000\n",
            "train_loss 5.051263332366943\n",
            "train_acc 0.00027007996882625425\n",
            "valid_loss 153.43846130371094\n",
            "val_acc 0.07539468106260756\n",
            "epochs 990/1000\n",
            "train_loss 5.046957969665527\n",
            "train_acc 0.00026995709021530975\n",
            "valid_loss 153.42881774902344\n",
            "val_acc 0.0754064251872478\n",
            "epochs 991/1000\n",
            "train_loss 5.042665481567383\n",
            "train_acc 0.0002698348569258367\n",
            "valid_loss 153.41917419433594\n",
            "val_acc 0.07541819021425908\n",
            "epochs 992/1000\n",
            "train_loss 5.038366794586182\n",
            "train_acc 0.0002697129202715587\n",
            "valid_loss 153.40951538085938\n",
            "val_acc 0.07542995906551368\n",
            "epochs 993/1000\n",
            "train_loss 5.034084320068359\n",
            "train_acc 0.00026959181755396107\n",
            "valid_loss 153.39987182617188\n",
            "val_acc 0.07544175795337651\n",
            "epochs 994/1000\n",
            "train_loss 5.029820919036865\n",
            "train_acc 0.00026947118469531863\n",
            "valid_loss 153.39031982421875\n",
            "val_acc 0.07545360264266492\n",
            "epochs 995/1000\n",
            "train_loss 5.025559425354004\n",
            "train_acc 0.00026935068700158107\n",
            "valid_loss 153.3807830810547\n",
            "val_acc 0.07546548226382584\n",
            "epochs 996/1000\n",
            "train_loss 5.021311283111572\n",
            "train_acc 0.0002692304587981198\n",
            "valid_loss 153.37130737304688\n",
            "val_acc 0.07547737804998178\n",
            "epochs 997/1000\n",
            "train_loss 5.017064571380615\n",
            "train_acc 0.0002691104950477333\n",
            "valid_loss 153.36183166503906\n",
            "val_acc 0.07548928282340057\n",
            "epochs 998/1000\n",
            "train_loss 5.012827396392822\n",
            "train_acc 0.00026899105236784986\n",
            "valid_loss 153.35240173339844\n",
            "val_acc 0.07550123052642448\n",
            "epochs 999/1000\n",
            "train_loss 5.0085954666137695\n",
            "train_acc 0.00026887188785342284\n",
            "valid_loss 153.34300231933594\n",
            "val_acc 0.07551321279461262\n",
            "epochs 1000/1000\n",
            "train_loss 5.004364967346191\n",
            "train_acc 0.0002687529192301624\n",
            "valid_loss 153.3336639404297\n",
            "val_acc 0.07552524267448461\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 318
        },
        "id": "ZJxw3xX7Xljq",
        "outputId": "ac1b9397-1449-4d0e-b88f-108c4bd8d46e"
      },
      "source": [
        "plt.plot(train_losses, label='train loss')\r\n",
        "plt.plot(valid_losses, label='validation loss')\r\n",
        "\r\n",
        "plt.title('Training history')\r\n",
        "plt.ylabel('Loss')\r\n",
        "plt.xlabel('Epoch')\r\n",
        "plt.legend()\r\n",
        "# plt.ylim([0.000185, 0.0002]);"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f79009e9710>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 304
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tgbxORegfNR0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 318
        },
        "outputId": "cbc6debd-25af-4cac-d8af-e210704dd40c"
      },
      "source": [
        "plt.plot(train_accs, label='train acc')\r\n",
        "plt.plot(val_accs, label='validation acc')\r\n",
        "\r\n",
        "plt.title('Training history')\r\n",
        "plt.ylabel('Acc')\r\n",
        "plt.xlabel('Epoch')\r\n",
        "plt.legend()\r\n",
        "# plt.ylim([0.000185, 0.0002]);"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f7901306518>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 307
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}