1/numpy_base.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from tqdm import tqdm\n",
    "import random\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "df=pd.read_csv('data/train.tsv', sep='\\t')\n",
    "df_test=pd.read_csv('data/test.tsv', sep='\\t')\n",
    "# df=df.head(1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "def BOW1_tokenize(data):\n",
    "    feature_set=set()\n",
    "    for row in data:\n",
    "        token_list=row.split()\n",
    "        for token in token_list:\n",
    "            feature_set.add(token)\n",
    "    return feature_set\n",
    "\n",
    "    \n",
    "# Bag of Word 特征提取，基于单词的特征模型。\n",
    "def BOW1_creat_feature(data,feature_set):\n",
    "    feature_size=len(feature_set)\n",
    "    feature_list=np.zeros((data.shape[0],feature_size)).astype('int16')\n",
    "    feature_map=dict(zip(feature_set,range(feature_size)))\n",
    "    for index in range(data.shape[0]):\n",
    "        token_list=data[index].split()\n",
    "        for token in token_list:\n",
    "            if token in feature_set:\n",
    "                feature_index=feature_map[token]\n",
    "                feature_list[index,feature_index]+=1\n",
    "    return feature_list\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "def NGram_tokenize(data,ngram=2):\n",
    "    # ngram特征集合\n",
    "    feature_set=set()\n",
    "    for row in data:\n",
    "        token_list=row.split()\n",
    "        # 对每个词提取ngram特征\n",
    "        for token in token_list:\n",
    "            for i in range(len(token)-ngram):\n",
    "                feature_set.add(token[i:i+ngram])\n",
    "    return feature_set\n",
    "\n",
    "    \n",
    "def NGram_creat_feature(data,feature_set,ngram=2):\n",
    "\n",
    "    # NGram 特征提取。\n",
    "    \n",
    "    feature_size=len(feature_set)\n",
    "    feature_list=np.zeros((data.shape[0],feature_size)).astype('int16')\n",
    "    feature_map=dict(zip(feature_set,range(feature_size)))\n",
    "    for index in range(data.shape[0]):\n",
    "        token_list=data[index].split()\n",
    "        for token in token_list:\n",
    "            for i in range(len(token)-ngram):\n",
    "                gram=token[i:i+ngram]\n",
    "                if gram in feature_set:\n",
    "                    feature_index=feature_map[gram]\n",
    "                    feature_list[index,feature_index]+=1\n",
    "    return feature_list\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "def CreateFeature(ngram=2,analyzer='word'):\n",
    "    if analyzer=='word':\n",
    "        feature_set=BOW1_tokenize(data=df['Phrase'].to_numpy())\n",
    "        X=BOW1_creat_feature(data=df['Phrase'].to_numpy(),feature_set=feature_set)\n",
    "        test_X=BOW1_creat_feature(data=df_test['Phrase'].to_numpy(),feature_set=feature_set)\n",
    "    if analyzer=='char':\n",
    "        feature_set=NGram_tokenize(data=df['Phrase'].to_numpy(),ngram=ngram)\n",
    "        X=NGram_creat_feature(data=df['Phrase'].to_numpy(),feature_set=feature_set,ngram=ngram)\n",
    "        test_X=NGram_creat_feature(data=df_test['Phrase'].to_numpy(),feature_set=feature_set,ngram=ngram)\n",
    "    return feature_set,X,test_X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_test_split(X,y,test_rate):\n",
    "    X_size=X.shape[0]\n",
    "    train_size=int((1-test_rate)*X_size)\n",
    "    index = [i for i in range(X_size)] \n",
    "    # print(train_size)\n",
    "    random.shuffle(index)\n",
    "    train_X = X[index][0:train_size]\n",
    "    train_y = y[index][0:train_size]\n",
    "    val_X = X[index][train_size+1:-1]\n",
    "    val_y = y[index][train_size+1:-1]\n",
    "    # print(train_X.shape,val_X.shape,train_y.shape,val_y.shape)\n",
    "    return train_X,val_X,train_y,val_y\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "# def train(batchsize=32,lr=1e0,epoch_number=100):\n",
    "#     iter_number=train_X.shape[0]//batchsize\n",
    "#     iter_remain=train_X.shape[0]%batchsize\n",
    "#     weight=np.zeros((train_X.shape[1],class_number))\n",
    "#     # 不同初始值的影响\n",
    "#     # weight=np.random.normal(0,1,[train_X.shape[1],class_number])\n",
    "#     loss_list=[]\n",
    "#     for i in range(epoch_number):\n",
    "#         loss=0\n",
    "#         for j in tqdm(range(iter_number)):\n",
    "#             train_data=train_X[j*batchsize:j*batchsize+batchsize]\n",
    "#             train_label=train_y[j*batchsize:j*batchsize+batchsize]\n",
    "#             weight_temp=np.zeros((train_X.shape[1],class_number))\n",
    "#             for z in range(batchsize):\n",
    "#                 y=train_label[z]               \n",
    "#                 y_hat=np.exp(train_data[z].dot(weight))/np.sum(np.exp(train_data[z].dot(weight)))\n",
    "#                 loss+= (-1/train_X.shape[0])*np.sum(np.multiply(y,np.log10(y_hat)))\n",
    "#                 weight_temp+=(lr/batchsize)*np.outer(train_data[z],(y-y_hat))\n",
    "#             weight+=weight_temp\n",
    "#         print(loss)\n",
    "#         loss_list.append(loss)\n",
    "        \n",
    "#     return loss_list,weight\n",
    "\n",
    "# 矩阵乘法加速\n",
    "def train_faster(train_X,train_y,val_X,val_y,batchsize=32,lr=1e0,epoch_number=100):\n",
    "    iter_number=train_X.shape[0]//batchsize\n",
    "    iter_remain=train_X.shape[0]%batchsize\n",
    "    weight=np.zeros((train_X.shape[1],class_number))\n",
    "    # 不同初始值的影响\n",
    "    # weight=np.random.normal(0,1,[train_X.shape[1],class_number])\n",
    "    train_loss_list=[]\n",
    "    test_loss_list=[]\n",
    "    for i in range(epoch_number):\n",
    "        train_loss=0\n",
    "        test_loss=0\n",
    "        for j in tqdm(range(iter_number)):\n",
    "            train_data=train_X[j*batchsize:j*batchsize+batchsize]\n",
    "            y_train=train_y[j*batchsize:j*batchsize+batchsize]\n",
    "            y=np.exp(train_data.dot(weight))   \n",
    "            y_hat=np.divide(y.T,np.sum(y,axis=1)).T\n",
    "            train_loss+= (-1/train_X.shape[0])*np.sum(np.multiply(y_train,np.log10(y_hat)))\n",
    "            # 每个batch权重更新一次\n",
    "            weight+=(lr/batchsize)*train_data.T.dot(y_train-y_hat)\n",
    "       \n",
    "        y=np.exp(val_X.dot(weight))   \n",
    "        y_hat=np.divide(y.T,np.sum(y,axis=1)).T\n",
    "        test_loss= (-1/val_X.shape[0])*np.sum(np.multiply(val_y,np.log10(y_hat)))\n",
    "        # print('train_loss:',train_loss,\" test_loss:\",test_loss)\n",
    "        train_loss_list.append(train_loss)\n",
    "        test_loss_list.append(test_loss)\n",
    "    \n",
    "    return train_loss_list,test_loss_list,weight\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [],
   "source": [
    "def drawpic(train_loss_list=[],test_loss_list=[],epoch_number=10,path='1.jpg'):\n",
    "    # make data\n",
    "    x = [i for i in range(epoch_number)]\n",
    "    # plot\n",
    "    fig, ax = plt.subplots()\n",
    "    ax.plot(x, train_loss_list, linewidth=2.0)\n",
    "    ax.plot(x, test_loss_list, linewidth=2.0)\n",
    "    plt.savefig(path)\n",
    "    plt.show()\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [],
   "source": [
    "class_number=5\n",
    "y=df['Sentiment'].to_numpy()\n",
    "y_onehot=np.zeros((y.shape[0],class_number)).astype('int16')\n",
    "for i in range(y.shape[0]):\n",
    "    y_onehot[i,y[i]]+=1\n",
    "\n",
    "\n",
    "\n",
    "parameter_list=[\n",
    "    {'ngram':2,'analyzer':'word','batchsize':32,\"lr\":1e0,'epoch_number':20,'test_rate':0.2},\n",
    "    {'ngram':2,'analyzer':'char','batchsize':32,\"lr\":1e0,'epoch_number':20,'test_rate':0.2},\n",
    "    {'ngram':3,'analyzer':'char','batchsize':32,\"lr\":1e0,'epoch_number':20,'test_rate':0.2},\n",
    "    {'ngram':2,'analyzer':'word','batchsize':8,\"lr\":1e0,'epoch_number':20,'test_rate':0.2},\n",
    "    {'ngram':2,'analyzer':'word','batchsize':128,\"lr\":1e0,'epoch_number':20,'test_rate':0.2},\n",
    "    {'ngram':2,'analyzer':'word','batchsize':32,\"lr\":1e1,'epoch_number':20,'test_rate':0.2},\n",
    "    {'ngram':2,'analyzer':'word','batchsize':32,\"lr\":1e0,'epoch_number':10,'test_rate':0.2},\n",
    "\n",
    "]\n",
    "\n",
    "def flow(parameter_list):\n",
    "    for parameter_dict in parameter_list:\n",
    "        \n",
    "        ngram=parameter_dict['ngram']\n",
    "        analyzer=parameter_dict['analyzer']\n",
    "        batchsize=parameter_dict['batchsize']\n",
    "        lr=parameter_dict['lr']\n",
    "        epoch_number=parameter_dict['epoch_number']\n",
    "        test_rate=parameter_dict['test_rate']\n",
    "\n",
    "        \n",
    "        feature_set,X,test_X=CreateFeature(ngram=ngram,analyzer=analyzer)\n",
    "\n",
    "        train_X,val_X,train_y,val_y=train_test_split(X,y_onehot,test_rate=test_rate)\n",
    "        train_loss_list,test_loss_list,weight=train_faster(train_X=train_X,train_y=train_y,val_X=val_X,val_y=val_y,batchsize=batchsize,lr=lr,epoch_number=epoch_number)\n",
    "        \n",
    "\n",
    "        y_temp=np.exp(val_X.dot(weight))   \n",
    "        y_temp=np.divide(y_temp.T,np.sum(y_temp,axis=1)).T\n",
    "        y_predict=np.array([np.argmax(i) for i in y_temp])\n",
    "        y_val=np.array([np.argmax(i) for i in val_y])\n",
    "\n",
    "        acc=np.sum(y_predict.astype('int')==y_val.astype('int'))/y_val.shape[0]\n",
    "        parameter_dict['acc']='%.4f' % acc\n",
    "        print('acc:',acc)\n",
    "\n",
    "        parameter_dict['best_train_loss']='%.4f' % np.min(train_loss_list) \n",
    "        parameter_dict['best_test_loss']='%.4f' % np.min(test_loss_list)\n",
    "        \n",
    "        print('train_loss:',parameter_dict['best_train_loss'],\" test_loss:\",parameter_dict['best_test_loss'])\n",
    "      \n",
    "        path='pic/ngram_{}-analyzer_{}-lr_{}-batchsize_{}-epoch_number_{}-test_rate_{}.jpg'.format(str(ngram),str(analyzer),str(lr),str(batchsize),str(epoch_number),str(test_rate))\n",
    "        parameter_dict['path']=path\n",
    "        print(path)\n",
    "        drawpic(train_loss_list=train_loss_list,test_loss_list=test_loss_list,epoch_number=epoch_number,path=path)\n",
    "    return parameter_list\n",
    "    \n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 25/25 [00:00<00:00, 6265.02it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8356.52it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12535.28it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8353.86it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8353.86it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8358.52it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 6266.14it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12524.80it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8352.53it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8349.86it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12580.40it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12521.81it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8375.88it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12349.26it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12349.26it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12530.78it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8357.18it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12535.28it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12342.00it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8266.92it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "acc: 0.494949494949495\n",
      "train_loss: 0.3815  test_loss: 0.5423\n",
      "pic/ngram_2-analyzer_word-lr_1.0-batchsize_32-epoch_number_20-test_rate_0.2.jpg\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 25/25 [00:00<00:00, 12307.23it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 25103.57it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12547.28it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8356.52it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12527.79it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12524.80it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 6267.26it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8352.53it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12523.30it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12397.45it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8373.20it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12536.78it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12532.28it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12533.78it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8391.29it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8355.85it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12532.28it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8358.52it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12532.28it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12523.30it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "acc: 0.494949494949495\n",
      "train_loss: 0.4033  test_loss: 0.5376\n",
      "pic/ngram_2-analyzer_char-lr_1.0-batchsize_32-epoch_number_20-test_rate_0.2.jpg\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 25/25 [00:00<00:00, 1460.27it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 6611.87it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8359.85it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 5012.31it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 6283.04it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 2779.38it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 6271.01it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8355.19it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12536.78it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8355.85it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 9875.46it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8126.61it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 6180.82it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 5837.42it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 9542.05it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 7423.02it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 9910.93it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 5349.88it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 10946.61it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 5826.72it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "acc: 0.5303030303030303\n",
      "train_loss: 0.3559  test_loss: 0.5807\n",
      "pic/ngram_3-analyzer_char-lr_1.0-batchsize_32-epoch_number_20-test_rate_0.2.jpg\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 100/100 [00:00<00:00, 20056.92it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 16537.75it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 16703.72it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 12546.90it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 16711.71it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 16711.71it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 14354.22it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 20054.05it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 16706.38it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 20120.43it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 12502.77it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 14323.83it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 17683.31it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 14327.74it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 19051.16it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 20055.01it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 19499.32it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 16711.71it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 19986.20it/s]\n",
      "100%|██████████| 100/100 [00:00<00:00, 20055.97it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "acc: 0.5707070707070707\n",
      "train_loss: 0.3200  test_loss: 0.5200\n",
      "pic/ngram_2-analyzer_word-lr_1.0-batchsize_8-epoch_number_20-test_rate_0.2.jpg\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 6/6 [00:00<00:00, 2005.72it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 3008.83it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 3005.59it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 3004.87it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 3008.11it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 3008.11it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 3008.83it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 3008.47it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 2961.73it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 2004.29it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 3005.59it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 2005.09it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 3008.11it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 3008.11it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 3007.75it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 3008.83it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 3008.47it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 3004.16it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 6009.03it/s]\n",
      "100%|██████████| 6/6 [00:00<00:00, 6013.34it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "acc: 0.5\n",
      "train_loss: 0.4699  test_loss: 0.5582\n",
      "pic/ngram_2-analyzer_word-lr_1.0-batchsize_128-epoch_number_20-test_rate_0.2.jpg\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAD7CAYAAABjVUMJAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAAA0j0lEQVR4nO3deXxU9b3/8dcnO1lYshISwhr2ncgigiCioChii6JWbdVSbWldai2391evt8u93lat2rorblVxV4oCUpVNRAj7vgcIhCzsSYBsn98f5wSGmISBLJNkPs/HYx4zc873zHzmMMw753vO+R5RVYwxxvifAF8XYIwxxjcsAIwxxk9ZABhjjJ+yADDGGD9lAWCMMX7KAsAYY/yUVwEgImNFZIuIbBeRaVW0GSkiq0Vkg4gs8JieISLr3HnpHtOjRWSeiGxz71vV/OMYY4zxlpzrPAARCQS2AmOATGA5cJOqbvRo0xJYAoxV1T0iEq+qOe68DCBNVfMqvO5fgEOq+qgbKq1U9be19smMMcZUK8iLNoOA7aq6E0BEZgATgI0ebW4GPlLVPQDlP/7nMAEY6T5+HZgPVBsAsbGx2r59ey9e2hhjTLkVK1bkqWpcxeneBEASsNfjeSYwuEKbLkCwiMwHooCnVPUNd54CX4iIAi+o6ovu9ARVzQJQ1SwRia/szUVkCjAFICUlhfT09MqaGWOMqYKI7K5sujcBIJVMq9hvFAQMBEYDzYBvRWSpqm4FhqnqfvcHfp6IbFbVhd4W7gbGiwBpaWk2boUxxtQSb3YCZwJtPZ4nA/sraTNHVQvcvv6FQF8AVd3v3ucAH+N0KQFki0gigHvvTbeRMcaYWuJNACwHUkWkg4iEAJOBmRXafAoMF5EgEQnH6SLaJCIRIhIFICIRwBXAeneZmcDt7uPb3dcwxhhTT87ZBaSqJSIyFZgLBALTVXWDiNztzn9eVTeJyBxgLVAGvKyq60WkI/CxiJS/19uqOsd96UeB90TkTmAPMKm2P5wxxpiqnfMw0IYkLS1NbSewMcacHxFZoappFafbmcDGGOOn/CYAGtOWjjHG1IcmHwDFpWX8/pP1jHpsPieLS31djjHGNBhNPgCCAwNYuecwGQcLmb/FjjQ1xphyTT4AACb2TwLg41X7fFyJMcY0HH4RANf0bYMIfL05l6OFxb4uxxhjGgS/CICE5mEM6xRLUWkZn6/P8nU5xhjTIPhFAABcZ91AxhhzFv8IgJzNjC/4kNCgAJbtOkTm4UJfV2SMMT7X9APgxBGYfiVhXz3Mb9tuAGDmmopj2RljjP9p+gHQrCWM/A8Abs95jK6yh49X7rMTw4wxfq/pBwDA4J9Bn8kElp7g5dC/kZ1zgI1Zx3xdlTHG+JR/BIAIXPMktO5DW7J5KvgZZq7ae87FjDGmKfOPAAAIbgY3/pOS0FaMClxDwoonKC2zbiBjjP/ynwAAaNWOwBtepZQA7ij7kG0L3vF1RcYY4zP+FQCAdBrF4nZTAWi/6NeQu8XHFRljjG/4XQAAJF/9EP8qHUJYWSFlM26Bk7ZD2Bjjf/wyADrFR/FG/G/YVNaWgIPb4OO7oazM12UZY0y98ssAABjXvxM/K36AgoBI2PIZLHrc1yUZY0y98tsAuKZvG/ZJa3556hcoAl//GbZ+4euyjDGm3ngVACIyVkS2iMh2EZlWRZuRIrJaRDaIyAJ3WlsR+VpENrnT7/Vo/4iI7HOXWS0iV9XOR/JOXFQowzrH8lVpX9ak/gJQ+PAuOLijPsswxhifOWcAiEgg8AwwDugB3CQiPSq0aQk8C1yrqj2BSe6sEuDXqtodGAL8osKyf1PVfu7t8xp/mvM0sX8bAP58dBx0Gw+njsKMW+BUfn2XYowx9c6bLYBBwHZV3amqRcAMYEKFNjcDH6nqHgBVzXHvs1R1pfv4OLAJSKqt4mvqih6taRYcyPI9R8m89AmI7QK5m2DmVLCxgowxTZw3AZAEeI6bkMn3f8S7AK1EZL6IrBCR2yq+iIi0B/oD33lMnioia0Vkuoi0quzNRWSKiKSLSHpubq4X5XovIjSIK3omAPDJpmMw+W0IiYINH8OSp2v1vYwxpqHxJgCkkmkV/zwOAgYCVwNXAr8XkS6nX0AkEvgQuE9Vyw+6fw7oBPQDsoBKD8NR1RdVNU1V0+Li4rwo9/x4XihGYzrD9S86M/79COz4qtbfzxhjGgpvAiATaOvxPBmoOKB+JjBHVQtUNQ9YCPQFEJFgnB//t1T1o/IFVDVbVUtVtQx4Caerqd4N7xxLTEQIO3ILWL/vGHS7Ci79LWgZfHAHHN7ti7KMMabOeRMAy4FUEekgIiHAZGBmhTafAsNFJEhEwoHBwCYREeAVYJOqPuG5gIgkejydCKy/0A9RE0GBAVzT19kZ/Mlq93KRl06D1CvhxGF49xYosiuIGWOannMGgKqWAFOBuTg7cd9T1Q0icreI3O222QTMAdYCy4CXVXU9MAy4FbisksM9/yIi60RkLTAKuL+2P5y3yruBZq7ZT0lpGQQEOF1B0R3hwDqYdZ/tFDbGNDnSmK6MlZaWpunp6bX+uqrKqMfmk3GwkDfuGMSILu6+hpxN8NJoKC6Asf8HQ+6u9fc2xpi6JiIrVDWt4nS/PRPYk4ic3go43Q0EEN8drnvGeTz3P2Dxk7YlYIxpMiwAXNf1cwJg7voDFBaVnJnRcyJc9ntnp/C//wveucnZN2CMMY2cBYCrfWwE/dq2pKColHkbs8+eOeJBuGkGhLWArbPhhRGwb6VvCjXGmFpiAeBhotsN9Onqike5Al3Hwc8WQpv+cGQPTL8Slr1kXULGmEbLAsDD+D6JBAYIC7bmcjD/1PcbtGoPd8yFi34KpUXw+YPw4Z1w6ni912qMMTVlAeAhJjKUEamxlJYpn63LqrxRUChc/Rj8cDqERML6D+HFUZC9sX6LNcaYGrIAqMBzaIhq9foBTJkP8T3g4DZ46TJY/XbdF2iMMbXEAqCCK3q0JiIkkFV7jpCRV1B949hUuOtL6HszlJyAT+6BT6dC8Yn6KdYYY2rAAqCCZiGBXNmzNVDFzuCKQsLhumfh2r9DUBisehNeHmMXljHGNHgWAJXwPCnMqzOlRWDAbXDXv53hI7LXwQuXwsZP67hSY4y5cBYAlbi4UwxxUaHsyitgTeZR7xds3dvZL9D9Wig6Du/dBrOnQUlRndVqjDEXygKgEkGBAVzTxx0h9Fw7gysKawE3vOGMHRQQBN89B69dBUf2nntZY4ypRxYAVSg/Kexfa/ZTXFp2fguLOAPH/WQONE+GzOXw/CXw7bO2NWCMaTAsAKrQK6k5neIiOFhQxOLteRf2Im0vgrsXOdcWOHnEGVDumYucS07aGcTGGB+zAKiCiJweIO7T8+0G8hQeDTe/64wlFNsFDmfA+z+GV8bAnqW1UqsxxlwIC4BqTCgfIXRDNgWnSs7RuhoizlhC93wL4/8GEfFOt9D0K2HGLZC3vZYqNsYY71kAVCMlJpyB7VpxoriULzYeqPkLBgZB2h3wq5XOdYeDw2HzLHhmEHz2a8jPrfl7GGOMlywAzuH0OQGrvDgpzFuhUTDqd/DLlc75Aygsfxme7g8L/2rXIDbG1AsLgHMY3zuRoABh0bZc9hys5R/m5onOGcR3fwOpVzjnDnz1J/j7QFj1FpSV1u77GWOMBwuAc2gVEcL4PomUKdz37irnovG1LaEH3PI+3DYTWveB4/vh0587F57Z/mXtv58xxuBlAIjIWBHZIiLbRWRaFW1GishqEdkgIgvOtayIRIvIPBHZ5t63qvnHqRv/dU1PWjcPY+WeIzz9VR3usO14KUxZABNfhBZtIXs9/PN6eHMiHFhXd+9rjPFL5wwAEQkEngHGAT2Am0SkR4U2LYFngWtVtScwyYtlpwFfqmoq8KX7vEFqFRHCEzf2RQT+8dU2lu06VHdvFhAAfW+Eqelw+X9DaAvY8ZVzItmbE2HrF1BWB1shxhi/480WwCBgu6ruVNUiYAYwoUKbm4GPVHUPgKrmeLHsBOB19/HrwHUX/CnqwcWdYrnn0k5OV9CMVRwtLK7bNwwOg0vug1+tgsH3QFAzJwjengT/SIPvXrArkRljasSbAEgCPAeyyXSneeoCtBKR+SKyQkRu82LZBFXNAnDv4yt7cxGZIiLpIpKem+vbwyTvH9OFvskt2H/0JL/7ZJ13I4XWVEQMjHsUHtjobBE0T4ZDO2D2Q/B4d2ewuUM7674OY0yT400ASCXTKv7yBQEDgauBK4Hfi0gXL5etlqq+qKppqpoWFxd3PovWuuDAAJ6a3J+IkEA+W5vF+ysy6+/Nw6OdLYJ71ziDzaVc7Bw19N1z8PQAePtG2PG1DTFhjPGaNwGQCbT1eJ4MVDwoPhOYo6oFqpoHLAT6nmPZbBFJBHDvc2gE2sdG8IcJvQB4ZOYGdubm128BgUHQYwLcMRt+thD63QKBwbB1Drx5HTw7BNKn27kExphz8iYAlgOpItJBREKAycDMCm0+BYaLSJCIhAODgU3nWHYmcLv7+Hb3NRqF6wckcW3fNhQWlXLvjNUUlfhop2xiX+dqZPdvhFH/DyJbQ+5mmHU/PNEd5j0MR/b4pjZjTIN3zgBQ1RJgKjAX50f9PVXdICJ3i8jdbptNwBxgLbAMeFlV11e1rPvSjwJjRGQbMMZ93iiICH+a2IvkVs1Yt+8oj3+xxbcFRcbBpb+B+9bBD16BpDRn9NFvnoKn+sK7t0LGYjt6yBhzFqmXHZm1JC0tTdPT031dxmkrdh/mhhe+pbRM+eedg7kkNdbXJZ2RmQ7fPe8MPV3mDmTXIgV6/xD63ADx3X1bnzGm3ojIClVN+950C4CaefrLbTwxbyvxUaHMvnc4MZGhvi7pbMeyIP0VWP02HPMY1jqhN/SZBL1+CC0qHtRljGlKLADqSGmZctOLS1mWcYjLu8fz0m1piFR28JOPlZXB7m9g3XvOxepPll/rWKD9Jc5WQfdroVlLX1ZpjKkDFgB1aN+RE4x9ciHHT5bwxwk9uXVoe1+XVL2SU7DtC1j7HmydC6WnnOmBIdDlSuh9gzM4XXCYb+s0xtQKC4A6Nmvtfqa+vYrQoABmTr2Erq2jfF2Sd04cgU3/crYMdi3i9GkaoS2gx7XOlkG7S5whKowxjZIFQD146IM1vJeeSdeEKD6dOoyw4EBfl3R+ju2HdR84YeA5+FxUG+h1PXQbD20HQUAj+1zG+DkLgHpQcKqE8X9fzK68An58cXseubanr0u6cDmbnSBY9/7Z5xI0i3a6ibqMhc6jnYvbGGMaNAuAerIu8yjXP/cNxaXK9B+ncVm3BF+XVDOqsPc72DgTtnwOh3edmRcQDB2GQ5dx0HUstEzxXZ3GmCpZANSjFxbs4H9nbyY6IoQ59w4nvnkT2ZmqCnlbnSDYMgcyl4F6nFyW0Au6jnMCoU1/229gTANhAVCPysqUW6d/xzfbDzI8NZbXfzKIgIAGeGhoTRXkOUcTbZntDFVd5DEuUmSC21U0DjqOhJBwn5VpjL+zAKhn2cdOMvbJhRwuLOb/Xd2du4Z39HVJdavkFGQscrYMtsyGYx4jpQaFQYdLodMo5z6+OzTEcyWMaaIsAHzg3xuzueuNdIIDhY9/PoxeSS18XVL9UHUuZ7lltnPbv/Ls+RHx0GGEcwnMjiNt34ExdcwCwEce/nQ9b3y7m45xEXz882G0aBbs65Lq3/EDzsXtdy2AnQsg/8DZ81t1cMKgg3uLiPFNncY0URYAPnKyuJQJ//iGLdnHGZDSkjfuHExkaJCvy/Kd8h3JO+c7YZCxGE4dPbtNQu8zWwcpQyE00heVGtNkWAD4UObhQm58YSn7jpxgSMdoXv3xIJqF2MlUAJSWQNYa2DXfCYQ9S88MTQEQEATJFzlbBilDIDnNzj0w5jxZAPjY7oMFTHr+W3KOn2J4aiwv355GaJCFwPcUn3DOO9i5wOky2r/q7ENNJcA53DRlCLQd7Gwh2GimxlTLAqAB2J5znBtfWMrBgiIu757Acz8aQHCgHStfrRNHnFFMMxY7WwcH1p65vkG5Fm3dMHBDIaGnDVdhjAcLgAZiU9YxbnppKUcKi7m6TyJP3diPIAsB7xUVwL6VThjsXQp7l39/H0JIlNNVlDIUUgY7V0iz/QjGj1kANCBrM49wy0vfcfxUCdf3T+KxSX2b5oli9aGsDHI3uYHwnXN/ZPfZbSQQWveCNgMgaYBzlnJcdwj0453xxq9YADQwK3Yf4tZXllFYVMpNg1L4n4m9GuaFZBqjY1lnwmDvUshaC1p6dpugZpDYxwmD8mCI7mTDV5gmqUYBICJjgaeAQJwLvj9aYf5I4FOgfKSwj1T1DyLSFXjXo2lH4GFVfVJEHgF+CuS6836nqp9XV0dTCgCAJTvy+MmryzlVUsZPhrXn4fE9LATqQlGBszN5/yqn+2j/Sjic8f12oc0hsa8TCkkDnGBomWJnLZtG74IDQEQCga3AGCATWA7cpKobPdqMBB5U1fHneJ19wGBV3e0GQL6qPubth2hqAQAwf0sOU95YQVFpGfeM7MRDV3a1EKgPhYfcUFgJ+1c7wXB8//fbhce4Wwn9oXVv5wikVh1sS8E0KlUFgDedoIOA7aq6032hGcAEYGO1S33faGCHqu4+Z0s/MrJrPP+4uT8/f2slz83fQbPgQH41OtXXZTV94dHO9Qw6jz4z7fiBs7cS9q2EwoOw/d/OrVxIpHOkUUIvZ99CQm9I6AEhEfX/OYypAW8CIAnY6/E8ExhcSbuhIrIG2I+zNbChwvzJwDsVpk0VkduAdODXqnq44ouKyBRgCkBKStMcM+aKnq352439uHfGKp6Yt5Ww4ACmjOjk67L8T1RrZzjrruOc56rOxXDKtxKy18OB9c5QFnu/c26nCcR0OhMKrfs4j5u3sS4k02B50wU0CbhSVe9yn98KDFLVX3q0aQ6UqWq+iFwFPKWqqR7zQ3CCoaeqZrvTEoA8nIvQ/hFIVNU7qqulKXYBefpgRSYPvr8GgP++tie3X9zetwWZyhXkOZfMPLDuTCjkbfn++QkAzVq5odAb4ro5I6HGdYUwPxkY0DQINekCygTaejxPxvkxP01Vj3k8/lxEnhWRWFXNcyePA1aW//i77U4/FpGXgFlefZIm7IcDkzlVUsp/frye/5q5gbDgAG68qGlu9TRqEbHO0NadRp2ZVnIKcrd4hIJ7f+KwM0x2xqKzXyOqDcR3cw5HjetqwWB8wpsAWA6kikgHnJ24k4GbPRuISGsgW1VVRAYBAcBBjyY3UaH7R0QSVTXLfToRWH9hH6FpuWVwO04Wl/HHWRuZ9tE6QoMCua6/DXXQ4AWFOoeVJvY5M00Vju13giB7vXOd5dxNkLfN2eF8fL9zIR1Pp4Ohm20xmDp3zgBQ1RIRmQrMxTkMdLqqbhCRu935zwM/BO4RkRLgBDBZ3b4lEQnHOYLoZxVe+i8i0g+nCyijkvl+685LOnCyuJS/zt3Cr99fQ2hQAON6J/q6LHO+RJxxilokOVdHK1dW6hyGmrvZuXkVDIkQ0xliUyEm1X3cGVq2s2EvzAWzE8EasCe+2MLTX20nOFB44daBjf8C86Z61QVDycnKlwkMcQ5LjU11dkLHpLqPOzuHsNoOaIOdCdwoqSr/8/kmXlq0i5DAAB67oS/X9m3j67JMfSsrdYa3OLjDCYOD2+HgNsjbXvm5C+XCWnpsNXSG6A5OWER3cHZOG79hAdBIqSp/mLWRV7/JAOBXo1O5//JUO1nMOE7lw6EdTijkucFQ/rjoeNXLhbWE6I5nh0KrDs60qNa25dDEWAA0YqrKq99k8KfPNlKmML5PIo9N6ktYsPX9miqoQn62GwZuKBzOgEO74PAuKC6setmgZtCq/ffDoVU7Z+jt4LD6+hSmllgANAFfb8nhl2+vIv9UCX3btuSlWwcS39z+M5rzpAr5OXBopxMG5aFQfl94sPrlIxOcMZJatnPvU848b5FsAdEAWQA0EVsOHOfO15eTefgEiS3CePn2NHq2sUMETS06efT7oXBol3NW9NHM74+sWlFk67ODoXzLoUVb54goGzKj3lkANCF5+af42ZsrWLH7MOEhgTw1uT9jetgRQqYelJbA8SwnDM667XZuR/edOyCatYLmyc7WQosk5765x+OoRAgMrp/P4ycsAJqYk8Wl/MdH6/h41T5EYNrYbkwZ0dF2DhvfKi1xjkyqGBCHd8OxTOfEuNKi6l9DApytiBZJ0NwNhfJgaN7GuY9qbSFxHiwAmiBV5dn5O/jr3C0ATBqYzJ8n9iYkyIYqNg1UWRkU5sHRvc7WwtFMOObelz8+fgDn/NDqCETEQfNE5+zps+7dW/NE52gn+6PIAqApm70ui/vfW83J4jIGdYjm+R8NJDoixNdlGXNhSoqcbqaKwXAsy9m6OJYFBTmgZed+reBwZ2shqg1EJThbFlEJzo7syARnXmSC0y3VhIPCAqCJW5d5lLveWE72sVOkRIcz/cdpdI6P8nVZxtSN0hLnMNfjB86Ewvfus6Ao37vXCww5EwqRCZWHRWS8s9URFFq3n60OWAD4gQNHT3LXG8tZv+8YUWFBPHPzAEZ0ifN1Wcb4zsljZ0LieLZzLYf8HGdaeYDk58Cpo96/ZmgLiIyDiHj33vOxGxLlj0Mj6+6znQcLAD9RWFTCA++uYc6GAwQGCI9c04Nbh7b3dVnGNGxFhU4glN/Kw+J0aGRDfi4U5J77KCdPweHO8OER8c59eCxExLj3ns/daSERddIVZQHgR8rKlMfnbeGZr3cAcPvQdvx+fA+CAm3nsDE1UlYGJ484Ww0FOU4g5Oc6j/NznIsFFeScmVbVIH5VCQqrOiTaDoH2wy6o7JpcEMY0MgEBwm+u7EanuEimfbiO17/dzc68Ap6a3N92DhtTEwEBzvWkw6OBbtW3VXX2QeS7QVF40AmIwjwoOOjeV3hectI9XDbz+6837N4LDoCqWAA0YdcPSCYlOpwpb65g0bY8xj65kMdv6MvwVNsvYEydE4HQKOcW48U1vlWhqMAjECqERPsRtV+idQE1fZmHC7n/3dUszzgMwF2XdOA3Y7sSGmSDyRnjD6rqArJOYT+Q3CqcGVOG8usxXQgMEF5evIuJzyxhe041wwUbY5o8CwA/ERgg/HJ0Kh/cPZSU6HA2Zh3j6qcX8+bS3TSmrUBjTO2xAPAz/VNa8fm9w/nBgGROlZTx+0/W89M30jmYf8rXpRlj6pkFgB+KDA3i8Rv68veb+hMVFsS/N+Uw9qlFLNya6+vSjDH1yKsAEJGxIrJFRLaLyLRK5o8UkaMistq9PewxL0NE1rnT0z2mR4vIPBHZ5t7bRUrr2TV92zDnvhEMah9N7vFT3DZ9GX/410ZOlZzHiS7GmEbrnAEgIoHAM8A4oAdwk4j0qKTpIlXt597+UGHeKHe6517oacCXqpoKfOk+N/UsqWUz3pkyhN9c2ZXAAGH6N7uY8I9v2JZtO4iNaeq82QIYBGxX1Z2qWgTMACbUwntPAF53H78OXFcLr2kuQGCA8ItRnfnwnotpFxPO5gPHGf/3xbz5bYbtIDamCfMmAJKAvR7PM91pFQ0VkTUiMltEenpMV+ALEVkhIlM8pieoahaAex9f2ZuLyBQRSReR9Nxc66OuS/3atuSzXw1n0kB3B/GnG7jr9XTybAexMU2SNwFQ2chEFf8sXAm0U9W+wN+BTzzmDVPVAThdSL8QkfM6nU1VX1TVNFVNi4uzM1jrWmRoEH+d1Jdnbh5A87Agvtycw9gnFzF/S46vSzPG1DJvAiATaOvxPBnY79lAVY+par77+HMgWERi3ef73fsc4GOcLiWAbBFJBHDv7RemAbm6TyJz7hvB4A7R5OWf4sevLufX763hUME5LudnjGk0vAmA5UCqiHQQkRBgMjDTs4GItBb3YrQiMsh93YMiEiEiUe70COAKYL272Ezgdvfx7cCnNf0wpna1admMt386hIfGdiUkKIAPV2Yy+vH5vJ++1/YNGNMEnDMAVLUEmArMBTYB76nqBhG5W0Tudpv9EFgvImuAp4HJ6vxCJACL3enLgM9UdY67zKPAGBHZBoxxn5sGJjBA+PnIzsy9bwTDOsdwuLCY33ywlskvLmV7jpdXWzLGNEg2GJzxmqryyep9/HHWJg4VFBEcKNwzsjM/H9mJsGAbWM6YhsoGgzM1JiJM7J/MV7++lMkXtaW4VHn6y22Me2oRS7bn+bo8Y8x5sgAw561leAiP/qAP7/1sKJ3jI9mVV8DNL3/HA++utjGFjGlELADMBRvUIZrPfzWcB6/oQkhQAB+t2sfoJxbw7vI9lJU1nq5FY/yVBYCpkZCgAKZelsoX941geGosRwqL+e2H65j84lIbTsKYBs4CwNSK9rERvHHHIJ6a3I/YyBCWZRziqqcX8djcLZwstsHljGmILABMrRERJvRL4ssHRnLToBSKS5V/fL2dK59cyKJtNoyHMQ2NBYCpdS3Cg/nf63vzwd1D6ZIQye6Dhdz6yjJ++kY6O3Pt3AFjGgoLAFNn0tpHM+uXw3lobFfCQwKZtzGbK/62kEdmbrAhJYxpACwATJ0KCQrg5yM7M//BkUy+qC1lqry2JINL//o1LyzYYRefMcaH7ExgU682HzjGnz/bxKJtzoljya2aMW1cN67unYg7nJQxppbZmcCmQejWujlv3jmY135yEV0SIsk8fIKpb6/i+ueWsGL3YV+XZ4xfsS0A4zMlpWW8l57JE/O2kJfv7BO4uncivx3bjZSYcB9XZ0zTUdUWgAWA8bn8UyU8P38HLy3ayamSMkICA7j94nZMvSyVFs2CfV2eMY2eBYBp8PYfOcFjc7fw0ap9ALQMD+be0an8aEg7ggOtt9KYC2UBYBqNdZlH+dNnG/lu1yEAOsRGMG1cN67okWA7io25ABYAplFRVeZtzObR2ZvZmVcAQN/kFtw/pguXdomzIDDmPFgAmEapuLSMt5bu5h9fbz+9o3hgu1Y8MKYLF3eKsSAwxgsWAKZRKywq4c1vd/P8gh0cLiwGnOGoHxjThSEdY3xcnTENmwWAaRLyT5Xw+pIMXly4k6MnnCC4uFMMD4zpQlr7aB9XZ0zDZAFgmpRjJ4t57ZsMXlq0k+MnSwAY0SWO+y9PpX9KKx9XZ0zDUqMzgUVkrIhsEZHtIjKtkvkjReSoiKx2bw+709uKyNcisklENojIvR7LPCIi+zyWuaomH9D4l+ZhwfxqdCqLH7qMX13WmcjQIBZuzWXis0u447XlrMs86usSjWnwzrkFICKBwFZgDJAJLAduUtWNHm1GAg+q6vgKyyYCiaq6UkSigBXAdaq6UUQeAfJV9TFvi7UtAFOVwwVFvLRoJ68tyaCwyBlgbkyPBO67PJWebVr4uDpjfKsmWwCDgO2qulNVi4AZwARv3lRVs1R1pfv4OLAJSPK+bGO80yoihIfGdmPRQ6OYMqIjYcEBzNuYzdVPL+aef65gywG7PKUxFXkTAEnAXo/nmVT+Iz5URNaIyGwR6Vlxpoi0B/oD33lMnioia0VkuohU2nErIlNEJF1E0nNz7apSpnoxkaH87qruLHxoFHcM60BIUACz1x/gyicXMuWNdFbusQHnjCnnTQBUdqB1xX6jlUA7Ve0L/B345KwXEIkEPgTuU9Vj7uTngE5APyALeLyyN1fVF1U1TVXT4uLivCjXGIiPCuPha3qw6KFR3Da0HSFBAXyxMZvrn13CjS98y/wtOTSmAyCMqQveBEAm0NbjeTKw37OBqh5T1Xz38edAsIjEAohIMM6P/1uq+pHHMtmqWqqqZcBLOF1NxtSqhOZh/GFCLxb/dhQ/H9mJqNAgvtt1iB+/upyrnl7MzDX7KSkt83WZxviENwGwHEgVkQ4iEgJMBmZ6NhCR1uKekikig9zXPehOewXYpKpPVFgm0ePpRGD9hX8MY6oXHxXGQ2O78c1/XMa0cd2IiwplU9YxfvXOKi57fAH/XLqbk8V2dTLjX7w6D8A9RPNJIBCYrqp/FpG7AVT1eRGZCtwDlAAngAdUdYmIXAIsAtYB5X9m/U5VPxeRN3G6fxTIAH6mqlnV1WFHAZnacrK4lI9W7uPFhTvIOFgIQGxkCD8Z1oEfDWlnw1CbJsVOBDOmEqVlypz1B3huwXbW73N2T0WGBnHLkBTuHNaB+OZhPq7QmJqzADCmGqrK4u15PDd/B0t2HAQgJDCAHwxMYsqITnSIjfBxhcZcOAsAY7y0Zu8Rnl+wgzkbDqAKIjCuV2vuvKQjA1Ja2gikptGxADDmPO3IzeelhTv5cGUmxaXO/5O+yS34ybAOXNU7kZAgu0qZaRwsAIy5QAeOnuSNbzN4e9kejrhDUcdHhXLrkHbcPDiFmMhQH1doTPUsAIypoRNFpXyyeh+vfrOLrdn5AIQEBTChbxt+MqwDPdo093GFxlTOAsCYWqKqLNlxkOmLd/HVlhzK/wsN7hDNHZd04PLuCQQG2H4C03BYABhTBzLyCnhtSQbvp++lwB2FtG10M24f2p4bLmpL8zA7n8D4ngWAMXXo+Mli3k/P5LUlGew55JxYFh4SyKSBydx+cXs6xkX6uELjzywAjKkHpWXKV5tzePWbXafPJwAY1TWOW4e249Iu8dY9ZOqdBYAx9WzzgWO89k0GH6/ax6kSZySUpJbNuGlQW264qC3xUXaWsakfFgDG+MihgiLeT9/L28v2sNsddygoQLiiZwK3DG7H0I4xBNhWgalDFgDG+FhZmfLNjjzeWrqHeZuyKS1z/u91iI3g5kEp/HBgMq0iQnxcpWmKLACMaUCyj53k3eV7eWfZHrKOngSccwqu7p3ILYNTGNiulQ05YWqNBYAxDVBJaRlfb8nlre92s2Br7ulzCromRHHLkBSu659kh5KaGrMAMKaB23uokHeW7eG99L3k5RcB0Cw4kAn92nDL4Hb0Tm7h4wpNY2UBYEwjUVRSxhcbD/DW0j18u/PMoaTdE5tzQ1oy1/VLsn0F5rxYABjTCG3PyeedZXv4cGXm6YHoQgIDGNMjgR+mJTMiNc7OKzDnZAFgTCN2qqSULzfl8F76XhZuzcU9gIjWzcP4wcAkJg1sS3u7aI2pggWAMU3EgaMn+XBlJu+n7z19PWOAQe2jmZSWzFW9E4kIDfJhhaahqVEAiMhY4Cmci8K/rKqPVpg/EvgU2OVO+khV/1DdsiISDbwLtMe5KPwNqnq4ujosAIw5Q1VZnnGY99L38tnaLE4UO4PRRYQEMr5PGyalJdvhpAaoQQCISCCwFRgDZALLgZtUdaNHm5HAg6o63ttlReQvwCFVfVREpgGtVPW31dViAWBM5fJPlfD52izeS99L+u4zf0d1jI1gUlpbrh+QRIJd4N5vVRUA3mwnDgK2q+pO94VmABOAjdUude5lJwAj3XavA/OBagPAGFO5yNAgbrjIGWNoR24+76dn8tHKTHbmFfB/czbz17mbuSQ1jon923BFj9bWRWQA7wIgCdjr8TwTGFxJu6EisgbYj7M1sOEcyyaoahaAqmaJSHxlby4iU4ApACkpKV6Ua4x/6xQXybRx3Xjwii4s3JbLe8sz+XJzNgu35rJway7hIeu5smdrruufxLBOMQQF2rWN/ZU3AVBZB2LFfqOVQDtVzReRq4BPgFQvl62Wqr4IvAhOF9D5LGuMPwsKDOCybglc1i2BwwVFfLYui49X7WPF7sN8vGofH6/aR1xUKNf2bcPE/kn0bNPc9hf4GW8CIBNo6/E8Geev/NNU9ZjH489F5FkRiT3Hstkikuj+9Z8I5FzIBzDGnFuriBB+NKQdPxrSjj0HC/lktRMAu/IKeGXxLl5ZvIvO8ZFM7J/EhH5tSG4V7uuSTT3wZidwEM6O3NHAPpwduTe7XTzlbVoD2aqqIjII+ABoh3PkT6XLishfgYMeO4GjVfWh6mqxncDG1B5VZU3mUT5ZtY9/rdnPwYKi0/MGdYhmYv8kruqVSItwG4uosavpYaBXAU/i/KBPV9U/i8jdAKr6vIhMBe4BSoATwAOquqSqZd3pMcB7QAqwB5ikqoeqq8MCwJi6UVxaxuJteXy0ah9fbDhw+gI2IYEBjO4ez3X9kxjZNY7QoEAfV2ouhJ0IZozxyvGTxczdkM0nq/bxzY680yOURoUGMaZnAtf0bcMlnWMJtp3HjYYFgDHmvB04epKZa/bxyar9bMw6vauPluHBjOvVmvF92jCkY4yNR9TAWQAYY2pkZ24+s9ZmMWvtfrZm55+eHhsZwlW9Exnfpw1p7VrZ5S0bIAsAY0yt2XLgOLPW7mfW2ix25RWcnt66eRhX90lkfJ9E+rVtaYeVNhAWAMaYWqeqbNh/jH+t3c+sNVnsO3Li9LzkVs0Y36cN4/sk2jkGPmYBYIypU6rKqr1HmLUmi8/W7Sf72KnT8zrGRjC2V2vG9UqkV5KFQX2zADDG1JuyMmV5xiFmrc3i83VZZ51jkNyqGWN7tmZc79b0b2v7DOqDBYAxxidKSsv4btchZq/PYu6GbHKPn9kySGgeypU9WzO2V2sGtY+2cYnqiAWAMcbnysqUFXsOM3vdAeZuOHDWPoPoiBCu6JHA2F6tubhTLCFBFga1xQLAGNOgqCrr9h1l9voDzFl/4KyjiZqHBXF5dycMRnSJIyzYzkCuCQsAY0yDpapsyT7O7HVOGGzJPn56XnhIIKO6xXNFjwRGdo2nRTMbm+h8WQAYYxqNnbn5p7cM1u07enp6UIAwpGMMY3okMKZHAm1aNvNhlY2HBYAxplHae6iQuRsOMG9jNsszDlHm8ZPVs01zrujRmjE9EuieGGWHl1bBAsAY0+gdLijiq805zNuYzcJtuRQWlZ6el9SyGWN6JHBFjwQu6hBtg9V5sAAwxjQpJ4tLWbIjj3kbs5m3MYe8/DOHlzYPC+KybvGM6dGaS7vGEenn10C2ADDGNFllZcrqzCN8sSGbeRsPsCP3zBFFIYEBDO0Uw+Xd4xnVLd4vr3ZmAWCM8Rs7c/PdLYNsVuw5jOfPXNeEKC7rHs9l3eLp37alX5x8ZgFgjPFLefmn+GpzDl9vzmHh1lwKPPYbtAwP5tIucVzWLZ5Lu8TRMjzEh5XWHQsAY4zfKyopY3nGIb7clMNXm7PJOFh4el6AQFq7aEZ1i2d093hS4yObzFFFFgDGGFPBztx8vtqcw1ebc1i26xAlHseYJrdqxmXdnK6iIR1jGvXZyBYAxhhTjWMni1m8LY8vN+Uwf0vOWSOYNgsOZGinGEZ2jePSLnG0i4nwYaXnr0YBICJjgaeAQOBlVX20inYXAUuBG1X1AxHpCrzr0aQj8LCqPikijwA/BXLdeb9T1c+rq8MCwBhTH8rKlDWZR/h6cw5fbs5hw/5jZ81vHxPOpV3iGNnV2TpoFtKwtw4uOABEJBDYCowBMoHlwE2qurGSdvOAk8B0Vf2gkvn7gMGqutsNgHxVfczbD2EBYIzxhexjJ1mwNZcFW3NZtDWXYydLTs8LCQpgcIdoNxDi6BTX8PYdVBUA3pwdMQjYrqo73ReaAUwANlZo90vgQ+CiKl5nNLBDVXd7XbUxxjQACc3DuCGtLTektaWktIw1mUdYsMUJhLX7jrJoWx6LtuXxp882kdSyGSPcMLi4UwxRYQ138DpvAiAJ2OvxPBMY7NlARJKAicBlVB0Ak4F3KkybKiK3AenAr1X1cMWFRGQKMAUgJSXFi3KNMabuBAUGMLBdNAPbRfPAFV05mH+KRdvyWLA1l4Vbc9l35ATvLNvDO8v2EBQgDGzXiku7xjEiNY4eic0b1BXQvOkCmgRcqap3uc9vBQap6i892rwPPK6qS0XkNWCWZxeQiIQA+4GeqprtTksA8gAF/ggkquod1dViXUDGmIasrExZv//o6a2DlXsOnzV4XXRECMM6xzK8cyyXpMbW22imNekCygTaejxPxvkx95QGzHD7vWKBq0SkRFU/ceePA1aW//gDeD4WkZeAWV7UYowxDVZAgNAnuSV9klvyy9GpHC0sZvH2PBZszWHxtjz2Hz3Jv9bs519rnJ/QTnERDE+NY3hqLIM7xtT7mEXebAEE4ewEHo2zE3c5cLOqbqii/Wt8fwtgBjBXVV/1mJaoqlnu4/txdg5Prq4W2wIwxjRWqsrOvAIWbc1l8fY8vt1x8KyzkoMChAEprbgkNZbhqbH0SW5JYC11F9X0MNCrgCdxDgOdrqp/FpG73Q/1fIW2r+ERACISjrMPoaOqHvVo9ybQD6cLKAP4WXkgVMUCwBjTVBSXlrFqzxEWb8tl0fY81uw9clZ3UfOwIC7uFMvwLrEM7xxHSsyFD2JnJ4IZY0wDdrSwmG935p0+omjPocKz5t88OIX/mdj7gl67JvsAjDHG1LEW4cGM7ZXI2F6JAOw5WMii7bks2prHkh15dE9sXuvvaQFgjDENUEpMOLfEtOOWwe0oKS07a5yi2mIBYIwxDVxQYABBdTDaRNO/EoIxxphKWQAYY4yfsgAwxhg/ZQFgjDF+ygLAGGP8lAWAMcb4KQsAY4zxU41qKAgRyQUu9IIysTjDTzdUVl/NWH01Y/XVXEOusZ2qxlWc2KgCoCZEJL2ysTAaCquvZqy+mrH6aq4x1FiRdQEZY4yfsgAwxhg/5U8B8KKvCzgHq69mrL6asfpqrjHUeBa/2QdgjDHmbP60BWCMMcaDBYAxxvipJhcAIjJWRLaIyHYRmVbJfBGRp935a0VkQD3W1lZEvhaRTSKyQUTuraTNSBE5KiKr3dvD9VWf+/4ZIrLOfe/vXX/Tx+uvq8d6WS0ix0Tkvgpt6nX9ich0EckRkfUe06JFZJ6IbHPvW1WxbLXf1Tqs768istn99/tYRFpWsWy134U6rO8REdnn8W94VRXL+mr9vetRW4aIrK5i2TpffzWmqk3mhnPR+h1ARyAEWAP0qNDmKmA2IMAQ4Lt6rC8RGOA+jgK2VlLfSGCWD9dhBhBbzXyfrb9K/q0P4Jzg4rP1B4wABgDrPab9BZjmPp4G/F8V9Vf7Xa3D+q4AgtzH/1dZfd58F+qwvkeAB7349/fJ+qsw/3HgYV+tv5remtoWwCBgu6ruVNUiYAYwoUKbCcAb6lgKtBSRxPooTlWzVHWl+/g4sAlIqo/3rkU+W38VjAZ2qOqFnhleK1R1IXCowuQJwOvu49eB6ypZ1Jvvap3Up6pfqGqJ+3QpkFzb7+utKtafN3y2/sqJiAA3AO/U9vvWl6YWAEnAXo/nmXz/B9abNnVORNoD/YHvKpk9VETWiMhsEelZv5WhwBciskJEplQyv0GsP2AyVf/H8+X6A0hQ1SxwQh+Ir6RNQ1mPd+Bs0VXmXN+FujTV7aKaXkUXWkNYf8OBbFXdVsV8X64/rzS1AJBKplU8ztWbNnVKRCKBD4H7VPVYhdkrcbo1+gJ/Bz6pz9qAYao6ABgH/EJERlSY3xDWXwhwLfB+JbN9vf681RDW438CJcBbVTQ513ehrjwHdAL6AVk43SwV+Xz9ATdR/V//vlp/XmtqAZAJtPV4ngzsv4A2dUZEgnF+/N9S1Y8qzlfVY6qa7z7+HAgWkdj6qk9V97v3OcDHOJvanny6/lzjgJWqml1xhq/Xnyu7vFvMvc+ppI2vv4e3A+OBW9TtsK7Ii+9CnVDVbFUtVdUy4KUq3tfX6y8IuB54t6o2vlp/56OpBcByIFVEOrh/JU4GZlZoMxO4zT2aZQhwtHxzva65fYavAJtU9Ykq2rR22yEig3D+jQ7WU30RIhJV/hhnZ+H6Cs18tv48VPmXly/Xn4eZwO3u49uBTytp4813tU6IyFjgt8C1qlpYRRtvvgt1VZ/nPqWJVbyvz9af63Jgs6pmVjbTl+vvvPh6L3Rt33COUtmKc4TAf7rT7gbudh8L8Iw7fx2QVo+1XYKzmboWWO3erqpQ31RgA85RDUuBi+uxvo7u+65xa2hQ6899/3CcH/QWHtN8tv5wgigLKMb5q/ROIAb4Etjm3ke7bdsAn1f3Xa2n+rbj9J+Xfwefr1hfVd+FeqrvTfe7tRbnRz2xIa0/d/pr5d85j7b1vv5qerOhIIwxxk81tS4gY4wxXrIAMMYYP2UBYIwxfsoCwBhj/JQFgDHG+CkLAGOM8VMWAMYY46f+PwjILWo3kbgQAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 25/25 [00:00<00:00, 12302.90it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 6532.37it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12654.79it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8351.86it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12347.81it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8271.48it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8361.85it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8370.53it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8383.24it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8357.18it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 6267.26it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8355.19it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8357.18it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 6281.91it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12630.40it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8353.86it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 6266.89it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12533.78it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8360.52it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12533.78it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "acc: 0.494949494949495\n",
      "train_loss: 0.2863  test_loss: 0.6994\n",
      "pic/ngram_2-analyzer_word-lr_10.0-batchsize_32-epoch_number_20-test_rate_0.2.jpg\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 25/25 [00:00<00:00, 12343.45it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8351.19it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 6268.76it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12349.26it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8359.85it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8353.86it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8351.19it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 12532.28it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8351.86it/s]\n",
      "100%|██████████| 25/25 [00:00<00:00, 8363.18it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "acc: 0.5707070707070707\n",
      "train_loss: 0.4530  test_loss: 0.5318\n",
      "pic/ngram_2-analyzer_word-lr_1.0-batchsize_32-epoch_number_10-test_rate_0.2.jpg\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAD6CAYAAACoCZCsAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAAAyDElEQVR4nO3deXxU9bnH8c+TFUJYAiQQQhb2VQENIKIlSEVArkgXi1L11rYUl9ZWrVt7e71tvaWttrW27nq1LYraqtiWVVYVWcIeCEsIEJKQhS0bhGzP/eMcYAgJmZCESTLP+/XKa2bOnHPmmbw0X87y+z2iqhhjjPE/Ab4uwBhjjG9YABhjjJ+yADDGGD9lAWCMMX7KAsAYY/yUBYAxxvgprwJARCaJyG4RSRORx2tZJ0lEtojIDhFZ5bH8gIhsd99L9ljeWUSWishe9zGi4V/HGGOMt6SucQAiEgjsAW4EMoENwO2qutNjnU7AGmCSqmaISJSq5rnvHQASVfVItf3+BjimqnPcUIlQ1ccuVkvXrl01ISGhft/QGGP83MaNG4+oamT15UFebDsKSFPVdAARmQdMA3Z6rHMH8IGqZgCc+eNfh2lAkvv8LWAlcNEASEhIIDk5+WKrGGOMqUZEDta03JtTQDHAIY/Xme4yT/2BCBFZKSIbReQuj/cUWOIun+WxvJuqHgZwH6O8qMUYY0wj8eYIQGpYVv28URBwNTABaAt8ISJrVXUPMFZVs0UkClgqIrtUdbW3BbqhMQsgLi7O282MMcbUwZsjgEwg1uN1TyC7hnUWqWqJe65/NTAMQFWz3cc84EOcU0oAuSISDeA+1njaSFVfUdVEVU2MjLzgFJYxxphL5E0AbAD6iUgvEQkBZgAfV1tnPnC9iASJSBgwGkgVkXYi0h5ARNoBE4EUd5uPgbvd53e7+zDGGHOZ1HkKSFUrROQBYDEQCLyhqjtEZLb7/kuqmioii4BtQBXwmqqmiEhv4EMROfNZb6vqInfXc4D3ROTbQAbw9cb+csYYY2pX522gzUliYqLaXUDGGFM/IrJRVROrL/ebkcAVlVW+LsEYY5qVVh8ApeWVzFm4i0nPfcrpikpfl2OMMc1Gqw+A4MAAlu/KJS2vmL+sqXEshDHG+KVWHwCBAcITkwcB8PzyvZw4Webjiowxpnlo9QEAkDQgkmv7dKGwtII/LU/zdTnGGNMs+EUAiAhPTnGOAv7yxUEOHTvp44qMMcb3/CIAAIbGdGT6iBjKKqv4zeLdvi7HGGN8zm8CAODhif0JCQrgn1uz2XrohK/LMcYYn/KrAOgZEca3xiYA8PSCVFrSIDhjjGlsfhUAAPcl9SUiLJj1+4/xSao3bQuMMaZ18rsA6Ng2mO/f0A+AOQtTbYSwMcZv+V0AAHzzmnjiu4SxL7+Ed5MP1b2BMca0Qn4ZACFBATx600AAfr90L8WnK3xckTHGXH5+GQAAU67ozvDYThwpPs0rq9N9XY4xxlx2fhsAIsJPbnYGh726Op3cwlIfV2SMMZeX3wYAwMiEzkwc3I1T5ZX8fukeX5djjDGXlV8HAMBjkwcSGCC8l3yIPblFvi7HGGMuG68CQEQmichuEUkTkcdrWSdJRLaIyA4RWeUuixWRFSKS6i5/0GP9p0Qky91mi4hMaZyvVD99IsO5Y1QcVQq/WpDqixKMMcYn6gwAEQkE/gxMBgYDt4vI4GrrdAJeAG5R1SGc6+9bATysqoOAa4D7q237e1Ud7v4saPC3uUQPfrkf4aFBrNidz5q0I74qwxhjLitvjgBGAWmqmq6qZcA8YFq1de4APlDVDABVzXMfD6vqJvd5EZAKxDRW8Y2la3gos8f1BpwpIqqqbIoIY0zr500AxACeo6UyufCPeH8gQkRWishGEbmr+k5EJAEYAazzWPyAiGwTkTdEJKJ+pTeub1/Xm+4d2rAju5D5W7N8WYoxxlwW3gSA1LCs+j+Rg4CrgZuBm4D/EpH+Z3cgEg78A/ihqha6i18E+gDDgcPAszV+uMgsEUkWkeT8/Hwvyr00bUMCeWiiU/Izi/dQWm79g40xrZs3AZAJxHq87glk17DOIlUtUdUjwGpgGICIBOP88Z+rqh+c2UBVc1W1UlWrgFdxTjVdQFVfUdVEVU2MjIz09ntdkq9e1ZOB3duTdeIUb6050KSfZYwxvuZNAGwA+olILxEJAWYAH1dbZz5wvYgEiUgYMBpIFREBXgdSVfV3nhuISLTHy+lAyqV+icYSGCA84XYO+9OKNI6XWP9gY0zrVWcAqGoF8ACwGOci7nuqukNEZovIbHedVGARsA1YD7ymqinAWOBO4IYabvf8jYhsF5FtwHjgR4395S7Fl/p15bq+XSkqreB56x9sjGnFpCU1RUlMTNTk5OQm/5wd2QVMff4zggKETx4aR3yXdk3+mcYY01REZKOqJlZf7vcjgWsypIfTP7i8Uq1/sDGm1bIAqMUjEwcQGhTAv7cdZnPGcV+XY4wxjc4CoBY9OrXlnut6AfC/1j/YGNMKWQBcxL1JfejcLoQNB46zZGeur8sxxphGZQFwER3aBPODG/oC8OuFuyi3/sHGmFbEAqAOd4yOJ6FLGOlHSpi3PsPX5RhjTKOxAKhDSFAAj01y+gf/4RPrH2yMaT0sALwwaWh3ro6P4GhJGS+v2ufrcowxplFYAHhBRHhyinMU8Oqn6eQUWP9gY0zLZwHgpavjOzNpSHdKy6v43VIbHGaMafksAOrhsckDCQoQ3t+Yya6cwro3MMaYZswCoB56dW3HzNFxqMKvFuzydTnGGNMgFgD19IMJ/WgfGsSqPfl8ttf6BxtjWi4LgHrqEh7K7KQ+gDNFhPUPNsa0VBYAl+Db1/UiumMbdh4u5MPN1j/YGNMyWQBcgjbBgTw8cQAAzy7Zbf2DjTEtkgXAJZo+IoZB0R3ILijljc/3+7ocY4ypN68CQEQmichuEUkTkcdrWSfJbfm4Q0RW1bWtiHQWkaUistd9jGj417l8AgPODQ57ccU+jln/YGNMC1NnAIhIIPBnYDIwGLhdRAZXW6cT8AJwi6oOAb7uxbaPA8tUtR+wzH3dNCrKoAnm87++XyRf6h9J0ekK/rhsb6Pv3xhjmpI3RwCjgDRVTVfVMmAeMK3aOncAH6hqBoCq5nmx7TTgLff5W8Ctl/wt6vLJf8NfpsGRxv8j/cTkgYjA39Ye5MCRkkbfvzHGNBVvAiAGOOTxOtNd5qk/ECEiK0Vko4jc5cW23VT1MID7GFXf4r1yugi2vw/7V8ELY2DZL6DsZKPtflB0B756VU8qqpTfLLbBYcaYlsObAJAallU/nxIEXA3cDNwE/JeI9Pdy24t/uMgsEUkWkeT8/Pz6bOoIbQ/3rYMRd0JVOXz6DLwwGnYvqv++avHwxP60CQ5gwfYcNh60/sHGmJbBmwDIBGI9XvcEsmtYZ5GqlqjqEWA1MKyObXNFJBrAfcyjBqr6iqomqmpiZGSkF+XWoF0XmPYnuGcJdBsKJzLgnW/AvJnO8waK7tiWb1v/YGNMC+NNAGwA+olILxEJAWYAH1dbZz5wvYgEiUgYMBpIrWPbj4G73ed3u/toWnGjYdYquOlXEBIOu/4Ffx4Nn/3euVDcALPH9aFLuxA2HjzO4h05jVSwMcY0nToDQFUrgAeAxTh/1N9T1R0iMltEZrvrpAKLgG3AeuA1VU2pbVt313OAG0VkL3Cj+7rpBQbBmPvggQ0w5CtQfhI+eQpeug72f3rJu23fJpgHv9wPgF8v2m39g40xzZ60pNMViYmJmpyc3Lg73bcc/v0IHHM7fV35DZj4Swiv/zXp8soqbvr9atKPlPA/twzh7msTGrdWY4y5BCKyUVUTqy+3kcB9boB718D4n0BgKGx7F55PhPWvQlX9pngIDgzgUbd/8HPL9lJYWt4UFRtjTKOwAAAIbgPjHoX710LfG+F0ASx4BF69AbI21mtXNw3pRmJ8BMdKynhppfUPNsY0XxYAnjr3hpnvw21/hQ4xcHgLvDoB/vUQnPLu9k4R4cmbBwHw+mf7OVxwqgkLNsaYS2cBUJ0IDL4F7l8P1/4AAgIh+XXntNCWd7yaUuKquAhuviKa0xVVPLtkz2Uo2hhj6s8CoDah4TDxF/C9TyHuWjh5BD6aDW/eDHmpdW7+6KQBBAcK/9iUyc5s6x9sjGl+LADq0m0wfGsB3PoShHWFg587t4wu/RmcLq51s/gu7Zg5Ot7pH7yw7sAwxpjLzQLAGyIw/HZn7EDiPc7dQZ8/5wwiS/1nraeFzvQP/nTvEVbvuYRpLIwxpglZANRHWGeY+nv4zjKIHgaFmfDuN+Ht2+DYhU1hOrcL4b7xfQFniohK6x9sjGlGLAAuRc+r4bsrYMozENoB9i6BF66BVb+FitPnrfqtsQn06NiGXTlFfLAp00cFG2PMhSwALlVAIIz6LjyQ7IweriiFFb90ppzet/zsam2CA3nkpjP9g/dwqsz6BxtjmgcLgIZq3w2+8grc/U/o2t+ZUuKv0+H9b0HhYQBuHR7DkB4dyCm0/sHGmObDAqCx9PoSzP4cJvw3BLWFHR/An0bC2hcJ0EqenOIMDvvjsr0s35Xr42KNMcYCoHEFhcD1D8H962DAzVBWBIseh1eSGBuazl1j4jldUcWsv2xk/pYsX1drjPFzFgBNISIebn8bbp8HHeMgdzu8fiP/w8v86NouVFQpP3x3C39be9DXlRpj/JgFQFMaMNk5Grj+YQgIRjb/hQe3T2dhwjxGsIeffrSdP69Isw5ixhifsH4Al0v+Hlj8BKR9cnbR3qoY5lUm0X7UN3nwljGI1NRC2RhjGqa2fgAWAJfb0X2w+a+weS6UOG2QyzSQ1E7jGDr1+wT2SYIAOzAzxjSeBjWEEZFJIrJbRNJE5PEa3k8SkQIR2eL+/MxdPsBj2RYRKRSRH7rvPSUiWR7vTWngd2wZuvSBLz8FD+2Eb8zlSI8kAqliWMFyAudOR58bBqt+AwV2kdgY07TqPAIQkUBgD07f3kycRu+3q+pOj3WSgEdUdWod+8kCRqvqQRF5CihW1We8LbZVHAHUYHPKDj57/zlu1eXEBrhzBkmA05zmqrug/00QGOzbIo0xLVZtRwBBXmw7CkhT1XR3R/OAacDOi251oQnAPlW1W1+qGTF0CMERv2X662sZWLqZe9t/zrXlXyB7F8PexRDeDYbfASPudI4gjDGmEXhzCigGOOTxOtNdVt0YEdkqIgtFZEgN788A3qm27AER2SYib4hIhHclt05DYzry7r1jSW8/kpkFs7m9/ZsUj/sf6DoAinPhs9/D81fBm1Nh23tQXurrko0xLZw3AVDTrSnVzxttAuJVdRjwPPDReTsQCQFuAd73WPwi0AcYDhwGnq3xw0VmiUiyiCTn57fuKZX7RIbz/r3X0rtrO9bmClOTh5F5xwq4ZwkMnwnBYXDgU/jgu/DsAFjwKOSk+LpsY0wL5U0AZAKxHq97AtmeK6hqoaoWu88XAMEi0tVjlcnAJlXN9dgmV1UrVbUKeBXnVNMFVPUVVU1U1cTIyEivvlRLFtOpLe/NHsOQHh04cPQkX3tpLWltBsOtL8DDu5zpqKOHQ+kJWP8yvDTWaV6/8U04XeTj6o0xLYk3AbAB6Ccivdx/yc8APvZcQUS6i3sTu4iMcvd71GOV26l2+kdEoj1eTgfsn7KuruGhvDPrGkYmRJBTWMptL69le2YBtOnoNKT53ir43moY+V0I7QhZG+GfD8IzA2D+/XBovVe9i40x/s2rcQDuLZp/AAKBN1T1aRGZDaCqL4nIA8C9QAVwCnhIVde424bhXEPoraoFHvv8K87pHwUOAN9T1cMXq6O13gVUm1Nlldw7dyMrd+cTHhrEa3cnck3vLuevVH4Kds6HTX9x2lWeETnIuYNo2AynkY0xxm/ZQLAWqqyiioff38o/t2YTGhTACzOvYsKgbjWvfGSvM8hsy9tQ4l4vCQyBQf/hhEHCl2yQmTF+yAKgBausUn42P4W56zIIChCevW0Y04bXdCOWq6IM9ixyjgrSPuHsNftO8XDVnc4F5Q49LkvtxhjfswBo4VSV3y7ezQsr9yECP582lDuvia97wxOHYMtc2Pw3KHDv5pUASLjeGWDWbyJ06es0vjfGtEoWAK3ES6v2MWfhLgB+fNMA7kvq490kclWVkL7COSrY9W+oqjj3XkQvNwxuhPjrILhNE1VvjPEFC4BW5J31GTz54XZUYdaXevPE5IH1m0n05DFIW+aMMk77BE4dP/decBj0TnKODPpNhI4XOdVkjGkRLABamX9ty+ZH726hvFL5RmIs//uVKwgMuITTOFWVkJnshMGeJU7zGk/dhp4Lg54jIdCb2UOMMc2JBUArtHJ3HrP/tpHS8iqmXNGd339jOKFBgQ3baWE27F3ihEH6SigvOfdem07Q98vO6aI+E6Bdl9r2YoxpRiwAWqkNB45xz5sbKCqt4Pp+XXn5zqsJC2mkf6VXnHbGFuxZ4hwhHEs/954EQEwi9HePDrpfaReSjWmmLABasR3ZBdz1+nqOlpRxdXwEb9w9ko5hTTB99NF9sMedofTA51BVfu699tHOReR+E51rCKHtG//zjTGXxAKglUvPL+bO19eTdeIUA7u35y/fHkVU+ya8m+d0EaSvcsJg71Io8hjEHRAMCWOhn3ubade+TVeHMaZOFgB+IPvEKb75+jrS80tI6BLGX789mtjOYU3/waqQs/3cheTMDZw3YWzn3k4Y9J8I8WMhKLTpazLGnGUB4CeOFp/m7v9bT0pWId07tOFv3xlF36jLfDqm5CjsW+acLkr7xJm59Izgds4pov4TnY5ndpupMU3OAsCPFJaW8523klm//xgRYcG8dc8oruzZyTfFVFZAVrJ77WAJ5Fab9LVzH4gdDbGjnMfIgTZfkTGNzALAz5SWV3Lf3E0s35VHeGgQr96VyJg+zeC2zYJM55rB3iXONQTP20zBmd46duS5UIi52i4oG9NAFgB+qLyyioff28rHW7MJCQrghTuu4suDa5lJ1Bcqy50jgkPr4dA6yFgHhZnnryMB0G2IGwhuKHSKt1tOjakHCwA/VVWl/OzjFP62NoPAAOHZrw/j1hHN+Lx7QaYbCG4o5Gw7f94igPDu504ZxY6G6CvtwrIxF2EB4Mc8ZxIF+Pm0Idw1JsG3RXmr7CRkb4ZDa8+FgufcRQCBodBjhBMKcddAz1EQ3vrbhxrjLQsAw8ur9vErdybR+5L68KMb+xMc2MIuuKrC0TQnCA6tc0Ihf9eF63XuXe3i8iC7uGz8VoMCQEQmAc/htIR8TVXnVHs/CZgP7HcXfaCqP3ffOwAUAZVAxZkiRKQz8C6QgNMS8jZVrfZPu/NZADTcPHcm0SqFK2I68rvbhtGvWwu/yHrymDOh3ZlQyNoI5SfPXye0gzOZ3ZlQ6JloF5eN37jkABCRQGAPcCOQidMk/nZV3emxThLwiKpOrWH7A0Ciqh6ptvw3wDFVnSMijwMRqvrYxWqxAGgcX+w7yiPvbyXrxClCggL48cQB3HNdr0ubTbQ5qqzwuLjsnjo60wznDAmAqCHuEcIoZy6jrv0gsAmm0DDGxxoSAGOAp1T1Jvf1EwCq+iuPdZKofwDsBpJU9bCIRAMrVXXAxWqxAGg8RaXl/OJfO3kv2bnrZlSvzjz79WGXZ+SwLxRkQabHxeXDWy+8uBwQ7IxD6DbE42cohEfZXUemRWtIAHwNmKSq33Ff3wmMVtUHPNZJAv6Bc4SQjRMGO9z39gPHceYGeFlVX3GXn1DVTh77OK6qERerxQKg8X2yM5fHP9jOkeLTtAsJ5KdTBzNjZGz9Gsy0ROWn3IvL65zTR7k74Pj+mtcN63ouDLoNdp5HDoTgtpe3ZmMuUUMC4OvATdUCYJSqft9jnQ5AlaoWi8gU4DlV7ee+10NVs0UkClgKfF9VV3sbACIyC5gFEBcXd/XBgwfr/eXNxR0rKeOnH21nwfYcAMYPiOTXX72SqA5+1hrydBHk7XJOH+XuOPdzuuDCdSXA6aXseaTQbQh0jLWjBdPsNOkpoBq2OUDNp32eAopV9Rk7BdS8qCofb83mvz5KobC0gk5hwfxi2lD+Y1gPX5fmW6rO2ITcHecHw9E00MoL1w/tcOEppKhBdsHZ+FRDAiAI5yLwBCAL5yLwHWdO8bjrdAdyVVVFZBTwdyAeCAMCVLVIRNrhHAH8XFUXichvgaMeF4E7q+qjF6vFAqDp5RSU8ug/trF6Tz4AU6+M5hfThhLRLsTHlTUz5aVwZLfHkUIK5KTAySM1rx+RcO4o4UwwRCRAQAM7uBnjhYbeBjoF+APObaBvqOrTIjIbQFVfEpEHgHuBCuAU8JCqrhGR3sCH7m6CgLdV9Wl3n12A94A4IAP4uqoeu1gdFgCXh6ry9voMnv53KifLKolsH8pvvnol4wdG+bq05q84r9oppBTI3w2VZReuGxzmHB1EDT53pNClr9Ncx8YsmEZkA8FMvR08WsLD720l+aAzPOP2UbH85ObBhIdaY/h6qSx3Thl5XlfI3XHhvEdnBLWFzr2cwWxnfrr0cR7b97BwMPVmAWAuSWWV8tqn6Ty7ZA9llVXEdm7LM18bxujezWBm0Zbu1HHI3elxpLDL6btckl/7NkFtIKKXGwi9nOm0zwSEhYOphQWAaZDdOUU89N4WdmQXIgLfHtuLR24aQJtgO4fd6EoLnSA4lg7H9sGx/U4/5mP7vAuHzr2hy5mjBzcgOsRYOPgxCwDTYGUVVTy/fC8vrNxHZZXSLyqc3902nCt6dvR1af6j1nBIh5K82rcLDPU4Yuh17pRS5z4WDn7AAsA0mi2HTvDQe1tIzy8hKEB44Ia+3D++b8ubWK61KS10BrOdCYQzP0f3eRkO7lFDRAJ07OkEQ8ee0DbCxja0cBYAplGdKqvkN4t38X+fHwDgyp7OxHKXvf+w8c7povMD4dh+9wgiHYpzL75tUFund/OZQOgQ477ueW55mw6X53uYS2IBYJrEmrQj/Pjv285OLPfoTQO4Z2wvAlrLxHL+4HTRuUA4ug9OZEBhljN/UmEWnC6sex+hHTyCoZagsKkzfMYCwDSZwtJyfvHPnby/0bmtcXSvzjzTmieW8zelhR6BkHkuGAoynZ/CLKgorXs/bTtfeOTQMfbc8w49bDbWJmIBYJrc0p25PPHBNo4Ul9EuJJCf/cdgbkv0g4nl/J2q05Ohejh4hkbhYagqr2NHAuHdzgVC+2ins1u7KGdG1nZR514H+9k8VQ1kAWAui6PFp/npRyksTHEmlrthYBRzvnoFUe3tf1i/VlXlXIiu6SjiTFAU54BWebe/0A7QLtINhkgnOM4+rxYWIXYkagFgLhtVZf6WbH4235lYLiIsmF/eegU3Xxnt69JMc1ZZDkU554KhONeZWqMk33ksznWel+Rf2MvhYkLCq4VFlBMYNYVFaHjTfT8fsgAwl131ieVuGdaDn08bQqcwm1jONEBVFZSecMMh7/yQKMmD4vzzH2uah6k2wWEXBkNYF+dW2Lad3cdqP0HN/79nCwDjE6rK3HXOxHKnyivp1iGUX3/1SpIG2MRy5jJQhdKCWgIi78Jl3lzMri4k3A2DTh7BUENYhFVbFhTa6F+3NhYAxqeqTyx3x+g4fjJlEO1sYjnTXKg6t8SeFxZ5cOoEnDrmzN3k+XPSXVZTXwhvBId5BEWni4dF2wjnonhY50v6KAsA43OVVcqrn6bzO3diubjOYfzy1qF8qX+kr0sz5tKcCY2zwVAtKE4evzA4zqxTn+sYANfcB5Nq7cN1UbUFgP3zy1w2gQHC7HF9SBoQyUPvbmXn4ULuemM9X+ofyZNTBjKwu40mNS2MiDMKuk0HiIj3fjtVKCu+8GjiYj+d4hq/fDsCML5QVlHF65/t54UVaRSdriBA4GtX9+ThiQPo5m+9iI1pYnYKyDRLx0rK+OOyvfxt7UEqqpS2wYF89/pezBrXxxrPGNNIagsAr6ZvFJFJIrJbRNLc/r3V308SkQIR2eL+/MxdHisiK0QkVUR2iMiDHts8JSJZHttMacgXNC1T53YhPHXLEJY+NI5JQ7pzqrySPy5PI+m3K3l7XQYVlV4ODDLG1Js3TeEDcZrC3whk4jSFv11Vd3qskwQ8oqpTq20bDUSr6iYRaQ9sBG5V1Z0i8hRQrKrPeFusHQG0fskHjvHLf6ey5dAJAPpGhfPklIGMHxBlU0oYc4kacgQwCkhT1XRVLQPmAdO8+VBVPayqm9znRUAqEON92cbfJCZ05sP7ruVPd4wgtnNb0vKKuefNZGa+to6UrAJfl2dMq+JNAMQAhzxeZ1LzH/ExIrJVRBaKyJDqb4pIAjACWOex+AER2SYib4hIRD3qNq2YiDD1yh588tA4fnrzIDq2DWbNvqNMff4zHnp3C1knTvm6RGNaBW8CoKbj7urnjTYB8ao6DHge+Oi8HYiEA/8AfqiqZyYXfxHoAwwHDgPP1vjhIrNEJFlEkvPzL9IP1bQ6oUGBfOf63qz+8Xi+e30vQgID+GBzFuOfWcmvF+2isLSu2SWNMRfjTQBkArEer3sC2Z4rqGqhqha7zxcAwSLSFUBEgnH++M9V1Q88tslV1UpVrQJexTnVdAFVfUVVE1U1MTLSBgz5o45hwfzk5sEse3gc/zGsB2UVVby4ch9Jv13JW2sOUG4Xio25JN4EwAagn4j0EpEQYAbwsecKItJd3Ct0IjLK3e9Rd9nrQKqq/q7aNp5TQ04HUi79axh/ENs5jOdvH8GH913LyIQIjpWU8d8f72Di71ezeEcOLemWZmOaA6/GAbi3aP4BCATeUNWnRWQ2gKq+JCIPAPcCFcAp4CFVXSMi1wGfAtuBM/9Me1JVF4jIX3FO/yhwAPieqh6+WB12F5A5Q1VZsjOXOQt3sf9ICQAjEyJ4csogRsTZ5SRjPNlAMNMqlVdW8fa6DJ5btpdjJc60v1OvjOaxSQOtJaUxLgsA06oVlpbz4sp9vPHZfk5XVBESGMBdY+J54Ia+1n/A+D0LAOMXsk6c4tnFu/lgcxYAHdsG8/0b+nLnmHhCgwJ9XJ0xvmEBYPxKSlYBT/87lS/SjwIQ1zmMRycN4OYrom1EsfE7FgDG76gqK3bn8asFu9ibVwzAiLhO/GTKIBITLq2xhjEtkQWA8VsVlVW8l5zJ75bu4UjxaQAmDenOY5MH0qtrOx9XZ0zTswAwfq/4dAWvrE7nldX7KC2vIihA+OY18fxgQj86t7MLxab1sgAwxpVTUMrvlu7m/Y2ZqEL70CDuvjaB/xybQNfwy9eo25jLxQLAmGpSDxfyq4W7WL3HmWMqNCiAb4yM5bvX97YxBKZVsQAwphbJB47x0qp9fJKaBzi9i6deGc3scX0YFG19ik3LZwFgTB125xTx8qp9zN+aTWWV8//F+AGR3JvUl5EJEXb7qGmxLACM8VLm8ZO89ul+5m3IoLTcmcLqqrhO3JvUlwkDowgIsCAwLYsFgDH1dKykjDfXHOAvXxzgxEmn90C/qHC+N64P04b3IDjQq5baxvicBYAxl6jkdAXzNhzitU/TOVxQCkCPjm34zvW9mTEqlrCQIB9XaMzFWQAY00BlFVXM35LFy6vTSXNHFncKC+Y/r03g7jEJRNhYAtNMWQAY00iqqpRPUnN5YeU+thw6AUDb4EBmjIrlO9f3JqZTW98WaEw1FgDGNDJVZd3+Y7y4ch+r3LEEQQHCtOExzB7Xm37d2vu4QmMcFgDGNKEd2QW8vCqdf23Lxr2DlC8P6sa9SX24Ot46lBnfqi0AvLqNQUQmichuEUkTkcdreD9JRApEZIv787O6thWRziKyVET2uo/2f4lpsYb06Mgfbx/BykfG881r4ggJCuCT1Fy++uIabnv5C1bsyrOexabZqfMIQEQCgT3AjUAmTpP421V1p8c6ScAjqjrV221F5DfAMVWd4wZDhKo+drFa7AjAtBT5Rad5c81+/vLFQYpKKwAY2L099yb14eYrogmyW0jNZdSQI4BRQJqqpqtqGTAPmObl515s22nAW+7zt4BbvdynMc1eZPtQfnzTQNY8fgNPTB5IVPtQduUU8eC8LSQ9s5K/fnGA0vJKX5dp/Jw3ARADHPJ4nekuq26MiGwVkYUiMsSLbbup6mEA9zGqXpUb0wK0bxPM98b14dPHxjPnK1fQq2s7Mo+f4r/m72DsnOX8afleCtxBZsZcbt4EQE3j3qufN9oExKvqMOB54KN6bHvxDxeZJSLJIpKcn59fn02NaTZCgwKZMSqOTx4axwszr+KKmI4cLSnjmSV7uHbOMv53QSo57iAzYy4XbwIgE4j1eN0TyPZcQVULVbXYfb4ACBaRrnVsmysi0QDuY15NH66qr6hqoqomRkZGelGuMc1XYIAw5YpoPn5gLHO/M5rr+nalpKySV1anc92vl3P/25tYs++IXTA2l4U3Y9g3AP1EpBeQBcwA7vBcQUS6A7mqqiIyCidYjgInLrLtx8DdwBz3cX6Dv40xLYSIMLZvV8b27cq2zBO8tGofi1Jy+Pe2w/x722F6R7Zj5uh4vnpVDJ3CbISxaRpejQMQkSnAH4BA4A1VfVpEZgOo6ksi8gBwL1ABnAIeUtU1tW3rLu8CvAfEARnA11X12MXqsLuATGuWfeIU8zYcYt76DPKKnN7FoUEBTL2yBzOviWNEbCebktpcEhsIZkwLUV5ZxbLUPOauO8ine4+cXT44ugMzr4lj2vAYwkNtAjrjPQsAY1qgA0dKeGd9Bu8lH+K4e7dQu5BAbh0Rw8zR8QzuYR3LTN0sAIxpwU5XVLIoJYe5azNYf+DcmdKr4joxc3Q8N18ZTZvgQB9WaJozCwBjWok9uUXMXXuQDzZlUXTaGWXcsW0wX7u6J3eMjqNPZLiPKzTNjQWAMa3MybIK/rk1m7nrMtiWWXB2+bV9ujBzdDw3Du5GSJBNOWEsAIxp1bZlnmDu2gzmb80628e4a3goM0bGMmNULD0jwnxcofElCwBj/EDBqXI+2pzF39YeZK/btUwExg+IYuboOJIGRBFoTe39jgWAMX5EVdlw4Dhz1x1k4fYcyiqdo4KYTm2ZMTKWb4yMJapDGx9XaS4XCwBj/NTR4tP8fWMmb6/P4ODRk4DTuWzikG7MHB3PmN5dCLCjglbNAsAYP1dVpXyWdoS56w7ySWoelW7rsl5d23HHqDi+dnVPa2zfSlkAGGPOyiko5d0Nh3hnfQY5hc4spCFBAUy9IprbR8eRGB9h0060IhYAxpgLVFRWsXxXHnPXZbB6bz5n/hzEdm7L9OEx3Doiht42rqDFswAwxlxUxtGTvLMhgw82ZZJbePrs8mGxnfjKiBimXhlNl/BQH1ZoLpUFgDHGK5VVytr0o3y4OYuF2w9TUua0rgwKEMb1j+TWETHcOLibTT3RglgAGGPq7VRZJUtTc/lwUyar9x45e+E4PDSIyUO7M31EDNfYXUTNngWAMaZBjhSf5p9bs/locxZbPaaeiO7YhmnDY5g+IoYB3dv7sEJTGwsAY0yjScsrZv6WLD7cnEXm8VNnlw+K7sBXRsRwy/AedLOBZs2GBYAxptFVVSkbM47z4eYs/rU1m8JSZ3bSAIGxfbty6/AYJg3tTjtrYONTDQoAEZkEPIfT1vE1VZ1Ty3ojgbXAN1T17yIyAHjXY5XewM9U9Q8i8hTwXSDffe9Jt6F8rSwAjGm+TldUsmJXHh9uzmL5rjzKK52/LW2DA5k4pBvTR8RwXd+uBAXaDKWX2yUHgIgEAnuAG4FMnCbxt6vqzhrWWwqU4vT+/XsN72cBo1X1oBsAxar6jLdfwgLAmJbhxMky/r39MB9tzmLDgeNnl3cND+WWYT2YPiKGoTEdbLDZZVJbAHhzXDYKSFPVdHdH84BpwM5q630f+Acwspb9TAD2qepBr6s2xrRIncJCmDk6npmj48k4epKPtmTx0eYs0o+U8Mbn+3nj8/30jQpn+ogYpg3vYdNV+4g3x2IxwCGP15nusrNEJAaYDrx0kf3MAN6ptuwBEdkmIm+ISIQXtRhjWpi4LmH8YEI/lj08jo/uH8t/XptAl3YhpOUV89vFu7nu1yu47eUvmLc+g4JT5b4u1694cwro68BNqvod9/WdwChV/b7HOu8Dz6rqWhF5E/iX5ykgEQkBsoEhqprrLusGHAEU+AUQrar31PD5s4BZAHFxcVcfPGgHEMa0dOWVVXy6N58PNmWxdGcupyuc6apDggKYMDCK6SNiSBoQZR3NGklDrgGMAZ5S1Zvc108AqOqvPNbZD5w5mdcVOAnMUtWP3PenAfer6sRaPiMBJzSGXqwWuwZgTOtTVFrOopQcPtycxRfpR8/OR9S+TRA3DurGpKHd+VL/SBt53AANCYAgnIvAE3Au4m4A7lDVHbWs/yYXHgHMAxar6v95LItW1cPu8x/hXByecbFaLACMad0OF5xi/hZnsNmunKKzy9uFBDJ+YBSTh0YzfmAkYSF2W2l9NPQ20CnAH3BuA31DVZ8WkdkAqvpStXXfxCMARCQM5xpCb1Ut8Fjvr8BwnFNAB4DvnQmE2lgAGOM/9uUXsyglh4Uph0nJKjy7PDQogKQBkUweGs0Ng6Lo0CbYh1W2DDYQzBjTYh06dpJFKTksSDnM5owTZ5eHBAZwXb+uTBranRsHdbOGNrWwADDGtAqHC06xOCWHhSk5rD9w7Ow1g8AAYUzvLky+ojsTB3cnsr1NXX2GBYAxptXJLzrNkp05LErJYc2+o2dnKxWBkQmdmTy0O5OGdie6Y1sfV+pbFgDGmFbteEkZS1NzWZSSw2d7j1BWWXX2vRFxnZg8tDuTh0YT29n/Bp1ZABhj/EZhaTnLU/NYmHKYlbvzz44zABga04HJQ6OZNLQ7ffyk3aUFgDHGL50sq2Dl7nwWbD/Mil15ZzucAQzo1p5JQ7sz+YruDOjWvtXOTWQBYIzxe6XllXy69wgLUw7zyc7cs9NXA/Tq2o5JQ7szZWh0q5uozgLAGGM8lFVUsWbfERal5LBkZy7HSsrOvtczoi2ThnRn8hXRjIjt1OJbXloAGGNMLSoqq1i//xgLU3JYtCOH/KLTZ9/rGh7K+AGRTBgUxXX9Iglvgc1tLACMMcYLVVXKpozjLNiew+IdOWSdONfyMjhQuKZ3F24YGMUNA6OI79LOh5V6zwLAGGPqSVXZnVvEstQ8VuzKY1PGcao8/mT2jQo/GwZXx0cQ3Ey7nVkAGGNMAx0rKWPVnjyWpeaxak8+RR4XkTu0CWLcgCgmDIxiXP/IZjUthQWAMcY0ovLKKpIPHGfF7jyWpeayL7/k7HsBAlfFRXDDoCgmDOxG/27hPr2ryALAGGOa0IEjJSzflcfyXXms23+U8spzf1tjOrVlwqAoxg+MYkzvLpe9t4EFgDHGXCZFpeV8nnbEuXawO48jxeduMW0bHMjYvl2ZMMi5dtCtQ5smr8cCwBhjfKCqStmWVcDy1FyW7cpjR3bhee8PjenADQOiuGFQN66M6dgkYw4sAIwxphnIKSh1rxvk8VlaPqXl5+Yp6hoewvgBUY0+5sACwBhjmpnS8kq+SD/Kil1OIFQfczC6lzPmYMKgho05aGhLyEnAczgtIV9T1Tm1rDcSWAt8w6Ml5AGgCKgEKs4UISKdgXeBBJyWkLep6vGL1WEBYIxprVSVPbnFLNuVy/LUC8cczBwdx9PTr7ikfdcWAHUeX4hIIPBn4EYgE9ggIh+r6s4a1vs1sLiG3YxX1SPVlj0OLFPVOSLyuPv6Ma++jTHGtDIiwoDu7RnQvT33JfXleEkZq/bks2xXHit35zG4R4dG/0xvTjCNAtJUNd0tch4wDdhZbb3vA/8ARnr52dOAJPf5W8BKLACMMQaAiHYh3DoihltHxFBeWXW221lj8mbccgxwyON1prvsLBGJAaYDL9WwvQJLRGSjiMzyWN5NVQ8DuI9R9SncGGP8RXBgQJOMHfDmCKCme5KqR9EfgMdUtbKG0W5jVTVbRKKApSKyS1VXe1ugGxqzAOLi4rzdzBhjTB28OQLIBGI9XvcEsqutkwjMcy/4fg14QURuBVDVbPcxD/gQ55QSQK6IRAO4j3k1fbiqvqKqiaqaGBkZ6c13MsYY4wVvAmAD0E9EeolICDAD+NhzBVXtpaoJqpoA/B24T1U/EpF2ItIeQETaAROBFHezj4G73ed3A/Mb/G2MMcZ4rc5TQKpaISIP4NzdEwi8oao7RGS2+35N5/3P6AZ86J4WCgLeVtVF7ntzgPdE5NtABvD1S/8axhhj6ssGghljTCtX2ziA5tm9wBhjTJOzADDGGD/Vok4BiUg+cPASN+8KVB+N7M/s93GO/S7OZ7+P87WG30e8ql5wG2WLCoCGEJHkms6B+Sv7fZxjv4vz2e/jfK3592GngIwxxk9ZABhjjJ/ypwB4xdcFNDP2+zjHfhfns9/H+Vrt78NvrgEYY4w5nz8dARhjjPHgFwEgIpNEZLeIpLnNZ/ySiMSKyAoRSRWRHSLyoK9rag5EJFBENovIv3xdi6+JSCcR+buI7HL/Oxnj65p8RUR+5P5/kiIi74hIG1/X1NhafQB4dDSbDAwGbheRwb6tymcqgIdVdRBwDXC/H/8uPD0IpPq6iGbiOWCRqg4EhuGnvxe3x8kPgERVHYozD9oM31bV+Fp9AODR0UxVy4AzHc38jqoeVtVN7vMinP+5Yy6+VesmIj2Bm4HXfF2Lr4lIB+BLwOsAqlqmqid8WpRvBQFtRSQICOPCafBbPH8IgDo7mvkjEUkARgDrfFyKr/0BeBSo8nEdzUFvIB/4P/eU2GvuNO5+R1WzgGdwZio+DBSo6hLfVtX4/CEAvOlo5ldEJBynf/MPVbXQ1/X4iohMBfJUdaOva2kmgoCrgBdVdQRQAvjlNTMRicA5U9AL6AG0E5Fv+raqxucPAeBNRzO/ISLBOH/856rqB76ux8fGAre4nezmATeIyN98W5JPZQKZqnrmqPDvOIHgj74M7FfVfFUtBz4ArvVxTY3OHwKgzo5m/kKczjyvA6mq+jtf1+NrqvqEqvZ0O9nNAJaraqv7V563VDUHOCQiA9xFE4CdPizJlzKAa0QkzP3/ZgKt8IK4N03hW7TaOpr5uCxfGQvcCWwXkS3usidVdYHvSjLNzPeBue4/ltKBb/m4Hp9Q1XUi8ndgE87dc5tphSOCbSSwMcb4KX84BWSMMaYGFgDGGOOnLACMMcZPWQAYY4yfsgAwxhg/ZQFgjDF+ygLAGGP8lAWAMcb4qf8HSwDUn3U/pkgAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "parameter_list=flow(parameter_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'ngram': 2, 'analyzer': 'word', 'batchsize': 32, 'lr': 1.0, 'epoch_number': 20, 'test_rate': 0.2, 'acc': '0.4949', 'best_train_loss': '0.3815', 'best_test_loss': '0.5423', 'path': 'pic/ngram_2-analyzer_word-lr_1.0-batchsize_32-epoch_number_20-test_rate_0.2.jpg'}\n",
      "{'ngram': 2, 'analyzer': 'char', 'batchsize': 32, 'lr': 1.0, 'epoch_number': 20, 'test_rate': 0.2, 'acc': '0.4949', 'best_train_loss': '0.4033', 'best_test_loss': '0.5376', 'path': 'pic/ngram_2-analyzer_char-lr_1.0-batchsize_32-epoch_number_20-test_rate_0.2.jpg'}\n",
      "{'ngram': 3, 'analyzer': 'char', 'batchsize': 32, 'lr': 1.0, 'epoch_number': 20, 'test_rate': 0.2, 'acc': '0.5303', 'best_train_loss': '0.3559', 'best_test_loss': '0.5807', 'path': 'pic/ngram_3-analyzer_char-lr_1.0-batchsize_32-epoch_number_20-test_rate_0.2.jpg'}\n",
      "{'ngram': 2, 'analyzer': 'word', 'batchsize': 8, 'lr': 1.0, 'epoch_number': 20, 'test_rate': 0.2, 'acc': '0.5707', 'best_train_loss': '0.3200', 'best_test_loss': '0.5200', 'path': 'pic/ngram_2-analyzer_word-lr_1.0-batchsize_8-epoch_number_20-test_rate_0.2.jpg'}\n",
      "{'ngram': 2, 'analyzer': 'word', 'batchsize': 128, 'lr': 1.0, 'epoch_number': 20, 'test_rate': 0.2, 'acc': '0.5000', 'best_train_loss': '0.4699', 'best_test_loss': '0.5582', 'path': 'pic/ngram_2-analyzer_word-lr_1.0-batchsize_128-epoch_number_20-test_rate_0.2.jpg'}\n",
      "{'ngram': 2, 'analyzer': 'word', 'batchsize': 32, 'lr': 10.0, 'epoch_number': 20, 'test_rate': 0.2, 'acc': '0.4949', 'best_train_loss': '0.2863', 'best_test_loss': '0.6994', 'path': 'pic/ngram_2-analyzer_word-lr_10.0-batchsize_32-epoch_number_20-test_rate_0.2.jpg'}\n",
      "{'ngram': 2, 'analyzer': 'word', 'batchsize': 32, 'lr': 1.0, 'epoch_number': 10, 'test_rate': 0.2, 'acc': '0.5707', 'best_train_loss': '0.4530', 'best_test_loss': '0.5318', 'path': 'pic/ngram_2-analyzer_word-lr_1.0-batchsize_32-epoch_number_10-test_rate_0.2.jpg'}\n"
     ]
    }
   ],
   "source": [
    "for i in parameter_list:\n",
    "    print(i)"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "b3ba2566441a7c06988d0923437866b63cedc61552a5af99d1f4fb67d367b25f"
  },
  "kernelspec": {
   "display_name": "Python 3.9.7 64-bit ('base': conda)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}