spanish_joke_classifier_tweet_only_LM_Transformer.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Use the tweets for the updated vocab, test classifer on humor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/farzin/haha_2019\r\n"
     ]
    }
   ],
   "source": [
    "!pwd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "%reload_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"]=\"0\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from fastai.text import *\n",
    "from fastai.callbacks import SaveModelCallback\n",
    "import sentencepiece as spm #https://github.com/google/sentencepiece"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Python code Examples (testing out tokenizer)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Reading the texts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "PATH = Path('./data/rnn/')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "all_texts_df = pd.read_csv('./data/haha_2019_train.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>text</th>\n",
       "      <th>is_humor</th>\n",
       "      <th>votes_no</th>\n",
       "      <th>votes_1</th>\n",
       "      <th>votes_2</th>\n",
       "      <th>votes_3</th>\n",
       "      <th>votes_4</th>\n",
       "      <th>votes_5</th>\n",
       "      <th>funniness_average</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>705196579758583809</td>\n",
       "      <td>Niveles de retraso mental: \\r\\n\\r\\n— Bajo.\\r\\n...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>678040651817213952</td>\n",
       "      <td>—Vamos Luke desenfunda tu sable, demuestra tu ...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>546750892213829633</td>\n",
       "      <td>- ¿Te ofrezco algo?, ¿Agua, café, mi corazón, ...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>965807211292364801</td>\n",
       "      <td>No se porqué me hago la cabeza deooos</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>638403841839484928</td>\n",
       "      <td>Quisiera saber que hago durante la siesta de l...</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   id                                               text  \\\n",
       "0  705196579758583809  Niveles de retraso mental: \\r\\n\\r\\n— Bajo.\\r\\n...   \n",
       "1  678040651817213952  —Vamos Luke desenfunda tu sable, demuestra tu ...   \n",
       "2  546750892213829633  - ¿Te ofrezco algo?, ¿Agua, café, mi corazón, ...   \n",
       "3  965807211292364801              No se porqué me hago la cabeza deooos   \n",
       "4  638403841839484928  Quisiera saber que hago durante la siesta de l...   \n",
       "\n",
       "   is_humor  votes_no  votes_1  votes_2  votes_3  votes_4  votes_5  \\\n",
       "0         1         1        2        2        0        0        0   \n",
       "1         1         1        3        0        1        0        0   \n",
       "2         1         0        2        1        0        1        1   \n",
       "3         0         3        0        0        0        0        0   \n",
       "4         0         4        0        1        0        0        0   \n",
       "\n",
       "   funniness_average  \n",
       "0                1.5  \n",
       "1                1.5  \n",
       "2                2.6  \n",
       "3                NaN  \n",
       "4                NaN  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_texts_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### apply pre-rules to text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "## New Line char for replacement in text\n",
    "NL = 'xxnl'\n",
    "defaults.text_spec_tok.append(NL) #add a New Line special char\n",
    "\n",
    "def sub_nl(t:str) -> str:\n",
    "    \"Replaces \\n by xxnl\"\n",
    "    return t.replace(\"\\r\\n\",\"\\n\").replace(\"\\n\",NL+\" \")\n",
    "\n",
    "# def sub_br(t:str) -> str:\n",
    "#     \"Replaces the <br /> by \\n\"\n",
    "#     re_br = re.compile(r'<\\s*br\\s*/?>', re.IGNORECASE)\n",
    "#     return re_br.sub(\"\\n\", t)\n",
    "\n",
    "def spec_add_spaces(t:str) -> str:\n",
    "    \"Add spaces between special characters\"\n",
    "    return re.sub(r'([/#?!@,])', r' \\1 ', t)\n",
    "\n",
    "def rm_useless_spaces(t:str) -> str:\n",
    "    \"Remove multiple spaces\"\n",
    "    return re.sub(' {2,}', ' ', t)\n",
    "\n",
    "def replace_rep(t:str) -> str:\n",
    "    \"Replace repetitions at the character level\"\n",
    "    def _replace_rep(m:Collection[str]) -> str:\n",
    "        c,cc = m.groups()\n",
    "        return f' {TK_REP} {len(cc)+1} {c} '\n",
    "    re_rep = re.compile(r'(\\S)(\\1{3,})')\n",
    "    return re_rep.sub(_replace_rep, t)\n",
    "    \n",
    "def replace_wrep(t:str) -> str:\n",
    "    \"Replace word repetitions\"\n",
    "    def _replace_wrep(m:Collection[str]) -> str:\n",
    "        c,cc = m.groups()\n",
    "        return f' {TK_WREP} {len(cc.split())+1} {c} '\n",
    "    re_wrep = re.compile(r'(\\b\\w+\\W+)(\\1{3,})')\n",
    "    return re_wrep.sub(_replace_wrep, t)\n",
    "\n",
    "def deal_caps(t:str) -> str:\n",
    "    \"Replace words in all caps\"\n",
    "    res = []\n",
    "    for s in re.findall(r'\\w+|\\W+', t):\n",
    "        res += ([f' {TK_UP} ',s.lower()] if (s.isupper() and (len(s)>2)) else [s.lower()])\n",
    "    return ''.join(res)\n",
    "\n",
    "def fixup(x:str) -> str:\n",
    "    \"List of replacements from html strings\"\n",
    "    re1 = re.compile(r'  +')\n",
    "    x = x.replace('#39;', \"'\").replace('amp;', '&').replace('#146;', \"'\").replace(\n",
    "        'nbsp;', ' ').replace('#36;', '$').replace('\\\\n', \"\\n\").replace('quot;', \"'\").replace(\n",
    "        '<br />', \"\\n\").replace('\\\\\"', '\"').replace('<unk>',UNK).replace(' @.@ ','.').replace(\n",
    "        ' @-@ ','-').replace('\\\\', ' \\\\ ')\n",
    "    return re1.sub(' ', html.unescape(x))\n",
    "\n",
    "default_rules = [fixup, replace_rep, replace_wrep, deal_caps, spec_add_spaces, \n",
    "                 rm_useless_spaces, sub_nl]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Default Rules:\n",
      " ['fixup', 'replace_rep', 'replace_wrep', 'deal_caps', 'spec_add_spaces', 'rm_useless_spaces', 'sub_nl'] \n",
      "\n",
      "\n",
      "<function fixup at 0x7f4f3d6922f0>\n",
      "<function replace_rep at 0x7f4f3d6920d0>\n",
      "<function replace_wrep at 0x7f4f3d6921e0>\n",
      "<function deal_caps at 0x7f4f3d692268>\n",
      "<function spec_add_spaces at 0x7f4f3d6e7e18>\n",
      "<function rm_useless_spaces at 0x7f4f3d692158>\n",
      "<function sub_nl at 0x7f4f3d6e7bf8>\n"
     ]
    }
   ],
   "source": [
    "## apply the rules\n",
    "raw_text = all_texts_df.loc[:,'text']\n",
    "\n",
    "print(\"Default Rules:\\n\",[x.__name__ for x in default_rules],\"\\n\\n\")\n",
    "\n",
    "for rule in default_rules:\n",
    "    print(rule)\n",
    "    raw_text = raw_text.apply(lambda x: rule(str(x)))\n",
    "    \n",
    "all_texts_df['new_text'] = 'xxbos ' + raw_text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    xxbos niveles de retraso mental: xxnl xxnl — b...\n",
       "1    xxbos —vamos luke desenfunda tu sable , demues...\n",
       "2    xxbos - ¿te ofrezco algo ? , ¿agua , café , mi...\n",
       "3          xxbos no se porqué me hago la cabeza deooos\n",
       "4    xxbos quisiera saber que hago durante la siest...\n",
       "Name: new_text, dtype: object"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_texts_df['new_text'].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(24000, 11)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_texts_df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Clean up the imbalanced data to balance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of positive samples: 9253\n",
      "Number of negative samples: 14747\n",
      "Total samples: 24000\n"
     ]
    }
   ],
   "source": [
    "print('Number of positive samples:', (all_texts_df.is_humor == 1).sum())\n",
    "print('Number of negative samples:',  (all_texts_df.is_humor == 0).sum())\n",
    "print('Total samples:', len(all_texts_df))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "rnd_seed = 20190313\n",
    "np.random.seed(rnd_seed)\n",
    "\n",
    "idx = np.random.permutation(len(all_texts_df))\n",
    "test_cut = int(0.15 * len(idx))\n",
    "valid_cut = int(0.15 * len(idx-test_cut))\n",
    "\n",
    "df_train_all = all_texts_df.iloc[idx[:-(valid_cut+test_cut)],:]\n",
    "df_valid     = all_texts_df.iloc[idx[-(valid_cut+test_cut):-test_cut],:]\n",
    "df_test      = all_texts_df.iloc[idx[-test_cut:],:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test prevalence(n = 3600): 0.3938888888888889\n",
      "Valid prevalence(n = 3600): 0.3852777777777778\n",
      "Train all prevalence(n = 16800): 0.3838095238095238\n",
      "all samples (n = 24000)\n"
     ]
    }
   ],
   "source": [
    "print('Test prevalence(n = %d):'%len(df_test),df_test.is_humor.sum()/ len(df_test))\n",
    "print('Valid prevalence(n = %d):'%len(df_valid),df_valid.is_humor.sum()/ len(df_valid))\n",
    "print('Train all prevalence(n = %d):'%len(df_train_all), df_train_all.is_humor.sum()/ len(df_train_all))\n",
    "print('all samples (n = %d)'%len(all_texts_df))\n",
    "assert len(all_texts_df) == (len(df_test)+len(df_valid)+len(df_train_all)),'math didnt work'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((6448, 11), (10352, 11))"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## oversample the less-frequent occuring cases\n",
    "# split the training data into positive and negative\n",
    "rows_pos = df_train_all.is_humor == 1\n",
    "df_train_pos = df_train_all.loc[rows_pos]\n",
    "df_train_neg = df_train_all.loc[~rows_pos]\n",
    "df_train_pos.shape,df_train_neg.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train prevalence (n = 20704): 0.5\n"
     ]
    }
   ],
   "source": [
    "df_resample_pos = df_train_pos.sample(n=len(df_train_neg),replace=True,\n",
    "                                      axis=0,random_state=rnd_seed).reset_index(drop=True)\n",
    "df_train = pd.concat([df_resample_pos,df_train_neg],axis=0) #randomized again in DataBunch?\n",
    "print('Train prevalence (n = %d):'%len(df_train), df_train.is_humor.sum()/ len(df_train))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### sp tokenizer from training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_prefix = './all_tweets_es_0509'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "## itos from m.vocab file: just read directly and populate the dictionary\n",
    "itos = [] #{}\n",
    "with open(f'{model_prefix}.vocab','r') as f:\n",
    "    for line_num,line in enumerate(f):\n",
    "#         itos[line_num] = line.split(\"\\t\")[0]\n",
    "        itos.append(line.split(\"\\t\")[0])\n",
    "        \n",
    "class SPTokenizer(BaseTokenizer):\n",
    "    \"Wrapper around a SentncePiece tokenizer to make it a `BaseTokenizer`.\"\n",
    "    def __init__(self, model_prefix:str):\n",
    "        self.tok = spm.SentencePieceProcessor()\n",
    "        self.tok.load(f'{model_prefix}.model')\n",
    "\n",
    "    def tokenizer(self, t:str) -> List[str]:\n",
    "#         return self.tok.EncodeAsIds(t)  ## tokenize + numericalize. We have them broken into two parts,just return tokens\n",
    "        return self.tok.EncodeAsPieces(t)\n",
    "        \n",
    "class CustomTokenizer():\n",
    "    '''Wrapper for SentencePiece toeknizer to fit into Fast.ai V1'''\n",
    "    def __init__(self,tok_func:Callable,model_prefix:str, pre_rules:ListRules=None):\n",
    "        self.tok_func,self.model_prefix = tok_func,model_prefix\n",
    "        self.pre_rules  = ifnone(pre_rules,  defaults.text_pre_rules )\n",
    "        \n",
    "    def __repr__(self) -> str:\n",
    "        res = f'Tokenizer {self.tok_func.__name__} using `{self.model_prefix}` model with the following rules:\\n'\n",
    "        for rule in self.pre_rules: res += f' - {rule.__name__}\\n'\n",
    "        return res        \n",
    "\n",
    "    def process_text(self, t:str,tok:BaseTokenizer) -> List[str]:\n",
    "        \"Processe one text `t` with tokenizer `tok`.\"\n",
    "        for rule in self.pre_rules: t = rule(t)  \n",
    "        toks = tok.tokenizer(t)\n",
    "    \n",
    "        return toks \n",
    "    \n",
    "    def _process_all_1(self,texts:Collection[str]) -> List[List[str]]:\n",
    "        'Process a list of `texts` in one process'\n",
    "        tok = self.tok_func(self.model_prefix)\n",
    "        return [self.process_text(t,tok) for t in texts]\n",
    "                                                                     \n",
    "    def process_all(self, texts:Collection[str]) -> List[List[str]]: \n",
    "        \"Process a list of `texts`.\"                                 \n",
    "        return self._process_all_1(texts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['→', '😫', '🤨', '😲', '🏟', '📖', '🎸', '🏿', '😮', '🎬']"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "itos[-10:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "mycust_tok = CustomTokenizer(SPTokenizer,model_prefix,pre_rules=default_rules)\n",
    "sp_vocab = Vocab(itos)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = TextClasDataBunch.from_df(PATH,df_train,df_valid,df_test,\n",
    "                               tokenizer=mycust_tok, vocab=sp_vocab,\n",
    "                               text_cols='new_text', label_cols='is_humor')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>text</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>▁ xxbos ▁ xxbos ▁❤ .- ~ * ( ▁? ▁el ▁di vo 00 ▁d 3 ▁tw 1 tter ▁! ▁! ▁! ▁? ▁? ▁~ k 3 ▁di 4 vl 0 zz ▁ki 3 r 3 z ▁k 3 ▁p 4 rt 3 ▁d 3 ▁no 00 ▁n 0 ▁3 nt 13 nd 3 z ▁b 3 the ▁k 0 n ▁ki 3 n ▁t 3 ▁d 3 ▁l 4</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>▁ xxbos ▁ xxbos ▁❤ ️ -( el ▁di vo 00 ▁d 3 ▁tw 1 tter ▁! ▁! ▁? ▁i 0 ▁zo i ▁tu ▁ma 3 zt lo ▁? ▁ki 3 n ▁z up 0 ▁3 nz 3 ñ 4 l th 3 ▁? ▁d 3 ▁tu ▁k cuer p 0 ▁io o ▁k 0 noz ko ▁az t 4 ▁la ▁maz ▁in ty ma ▁p 4 r te ▁?</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>▁ xxbos ▁ xxbos ▁—¡ idio ta ▁! ▁ xxnl ▁—¡ h ist érica ▁! ▁ xxnl ▁— que ▁tarado . xxnl ▁— tar ada ▁tú . xxnl ▁— te ▁odio . xxnl ▁— yo ▁también . xxnl ▁— ▁¡ ▁ xxup ▁jode te ▁! ▁ xxnl ▁— no ▁me ▁hables . xxnl ▁—¡ ok ▁! ▁ xxnl ▁—¡ ok ▁! ▁ xxnl ▁— te ▁amo . xxnl ▁— yo</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>▁ xxbos ▁ xxbos ▁¡ ya ▁llegó ▁santa ▁! ▁ xxnl ▁ xxunk _ xxunk xxnl ▁( ・ xxunk ・ ) xxnl ▁_ | ▁ xxunk / ( ___ ▁ xxnl ▁/ ▁l ▁( __ __ / xxnl ▁ xxnl ▁ah ▁no ▁, ▁es ▁mi ▁papá ▁bien ▁pedo . xxnl ▁ xxnl ▁&lt; xxunk / xxunk - xxunk ___ xxnl ▁/ &lt; _ ▁/ ▁ xxrep ▁4 ▁_ ▁/ ”</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>▁ xxbos ▁ xxbos ▁¡ ya ▁llegó ▁santa ▁! ▁ xxnl ▁ xxunk _ xxunk xxnl ▁( ・ xxunk ・ ) xxnl ▁_ | ▁ xxunk / ( ___ ▁ xxnl ▁/ ▁l ▁( __ __ / xxnl ▁ xxnl ▁ah ▁no ▁, ▁es ▁mi ▁papá ▁bien ▁pedo . xxnl ▁ xxnl ▁&lt; xxunk / xxunk - xxunk ___ xxnl ▁/ &lt; _ ▁/ ▁ xxrep ▁4 ▁_ ▁/ ”</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>▁ xxbos ▁ xxbos ▁( ▁ xxup ▁o _ o ) ▁( xxunk _ ▁ xxunk ) ▁( ▁ xxup ▁o _ o ) ▁( xxunk _ ▁ xxunk ) ▁( ▁ xxup ▁o _ o ) ▁( xxunk _ ▁ xxunk ) ▁( ▁ xxup ▁o _ o ) ▁( xxunk _ ▁ xxunk ) ▁( xxunk _ ▁ xxunk ). ▁yo ▁con ▁el ▁móvil ▁en ▁clase .</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>▁ xxbos ▁ xxbos ▁— cuel ga ▁tú . xxnl ▁— no ▁, ▁cuelga ▁tú . ▁ xxnl ▁— no ▁, ▁tú . xxnl ▁— que ▁t ... xxnl ▁—¡ ▁ xxup ▁pinche ▁ xxup ▁tonta ▁me ▁ xxup ▁colgó ▁! ▁ xxnl ▁—¡ te ▁ xxup ▁escuché ▁ xxup ▁pendejo ▁! ▁ xxnl ▁— ▁ xxup ▁jajaj á ▁¡ ay ▁amor ▁! ▁es ▁broma ▁, ▁te ▁amo .</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>▁ xxbos ▁ xxbos ▁ xxup ▁tengo ▁un ▁ xxup ▁gran ▁ xxup ▁anuncio ▁, ▁ xxup ▁una ▁ xxup ▁noticia ▁ xxup ▁que ▁ xxup ▁cambiará ▁la ▁ xxup ▁historia ▁de ▁ xxup ▁twitter ▁, ▁ xxup ▁que ▁ xxup ▁devolver á ▁el ▁ xxup ▁sentido ▁a ▁ xxup ▁nuestra ▁ xxup ▁vida . xxnl ▁ xxnl ▁ xxup ▁pendientes ▁ xxup ▁hoy ▁en ▁la ▁ xxup ▁noche</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>▁ xxbos ▁ xxbos ▁— mamá ▁, ▁¿ puedo ▁salir ▁? ▁— no . ▁—¿ me ▁das ▁un ▁calce tín ▁? ▁— ok ▁, ▁toma . ▁— el ▁ xxup ▁amo ▁le ▁ha ▁ xxup ▁dado ▁a ▁ xxup ▁dob by ▁un ▁ xxup ▁calce tín ... ▁¡ ▁ xxup ▁dob by ▁es ▁ xxup ▁libre ▁! ▁# ▁ xxup ▁jajaja ▁# ▁ xxup ▁hp quo tes ▁# ▁xd</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>▁ xxbos ▁ xxbos ▁—¿ maestro ▁entonces ▁ya ▁aprobé ▁? ▁ xxnl ▁ xxnl ▁( • ▁- ▁• ) ▁ xxnl ▁ xxunk ▁ xxunk ▁ xxunk ▁__ ( • ▁- ▁• ) ▁ xxnl ▁_ xxunk ▁_ xxunk ▁ xxunk ▁ xxunk ▁ xxnl ▁ xxnl ▁—¡ hija ▁, ▁yo ▁soy ▁el ▁conser je ▁! ▁ xxnl ▁ xxnl ▁—¡ ay ▁, ▁por ▁no ▁poner ▁atención ▁!</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data.show_batch(10)  #less xxunk, but still a fair amount..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Check `xxunk` fraction on humor data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((498,), (20704,))"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pct_xxunk = pd.Series([sum(y==0)/len(y) for y in [x.data for x in data.train_dl.x]])\n",
    "pct_xxunk[pct_xxunk>0].shape, pct_xxunk.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.024053323029366306"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "498/20704."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pd.Series([sum(y==0)/len(y) for y in [x.data for x in data.train_dl.x]]).plot()\n",
    "_ = plt.title('Pct UNK in each tweet')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### build and train model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "config = tfmer_clas_config.copy()\n",
    "config['n_layers'] = 8\n",
    "config['ctx_len'] = 128\n",
    "config['d_model'] = 512\n",
    "config['d_inner'] = 2048\n",
    "config['mask'] = True  \n",
    "\n",
    "learn = text_classifier_learner(data, Transformer, drop_mult=0.5,pretrained=False,\n",
    "                               config=config)\n",
    "learn.load_encoder('twitter_es_enc_Tfm_0516')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n"
     ]
    }
   ],
   "source": [
    "lr_find(learn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEKCAYAAADjDHn2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzt3Xl8XHW9//HXJ5M9adKmTffSjS60hbY0tAVECggWFxCuYMGLAhe4yqKI4kW9Lhd+Llevy2W5ClwVVKByqyJiWWUVCzSF0n1NoU3XdEmbff38/piTdhqSJrQ5mZnk/Xw85sGcM98z5/Nl0vnM93y/5/s1d0dERORIUuIdgIiIJD4lCxER6ZCShYiIdEjJQkREOqRkISIiHVKyEBGRDilZiIhIh5QsRESkQ0oWIiLSodR4B9BVBgwY4KNGjYp3GCIiSWXJkiW73b2wo3I9JlmMGjWK4uLieIchIpJUzOzdzpTTZSgREemQkoWIiHRIyUJERDqkZCEiIh1SshARkQ4pWYiISIeULEREpENKFiIiSewPS0p55I3NoZ9HyUJEJIk9/MZm/vTW1tDPo2QhIpLENu2uYmxhTujnUbIQEUlS5dX17K2qZ8yA3NDPFWqyMLO5ZrbWzDaY2W1tvP5TM1saPNaZWXmr1/PMrNTM7g4zThGRZLSxrAqA0QPCb1mENpGgmUWAe4BzgVJgsZk97u6rWsq4+5diyt8ETG/1NncAL4cVo4hIMtu0O5osxiT5ZaiZwAZ3L3H3emA+cOERyl8GPNKyYWYzgEHAMyHGKCKStErKKklNMUYUZId+rjCTxTBgS8x2abDvPcxsJDAaeD7YTgF+DHzlSCcws+vMrNjMisvKyrokaBGRZFFSVsVxBdmkRcLvfk6UDu55wAJ3bwq2rwcWunvpkQ5y9/vcvcjdiwoLO1y7Q0SkR9m0u6pbLkFBuIsfbQVGxGwPD/a1ZR5wQ8z2qcAZZnY9kAukm1mlu7+nk1xEpDdqanY27anizAnd80M5zGSxGBhnZqOJJol5wOWtC5nZRKAfsKhln7t/Oub1K4EiJQoRkUO2lddQ39jcLSOhIMTLUO7eCNwIPA2sBh5195VmdruZXRBTdB4w3909rFhERHqakpaRUN2ULEJdg9vdFwILW+37Vqvt73TwHg8AD3RxaCIiSa2krBKA0d3UZ5EoHdwiIvI+bNpdRZ+MVApzM7rlfEoWIiJJqKQsOhLKzLrlfEoWIiJJqKSssts6t0HJQkQk6dTUN7Ftfy1jCsOfQLCFkoWISJLpzjmhWihZiIgkmZLdwUgoXYYSEZH2bOrGqclbKFmIiCSZkt1VDM3PJDs91FvlDqNkISKSZErKKrvtZrwWShYiIknE3SnZXdUtS6nGUrIQEUkiuyvrqaht7NaRUKBkISKSVA7OCdWNndugZCEiklRa7rEY24035IGShYhIUinZXUV6agpD+2Z163mVLEREkkhJWSWj+mcTSemeCQRbKFmIiCQJd2fdzu6dQLCFkoWISJJ4Zf1uNu+t5qwJA7v93EoWIiJJ4n9e3MCgvAwuOnlYt59byUJEJAm8uXkfr5Xs5dozxpCRGun28ytZiIgkgf95YSN9s9O4bOZxcTm/koWISIJbu6OC51bv5MrTRpGT0X2TB8ZSshARSXC/eGkj2ekRrjxtVNxiULIQEUlgW/ZW8/jb27h85nH0zU6PWxxKFiIiCezelzcSMeOaM8bENQ4lCxGRBPWPDbt5tLiUf5oxjMH5mXGNRclCRCQBvbSujKseWMzo/jl85bwJ8Q6H+HSri4hIu55btZPrH3qT4wfm8rtrZlGQE7++ihZqWYiIJJAnl2/nc79bwglD+vDItbMTIlGAWhYiIglj855qbnzkLaaN6MuvrzqFvMy0eId0UKgtCzOba2ZrzWyDmd3Wxus/NbOlwWOdmZUH+6eZ2SIzW2lmy8zsU2HGKSKSCN7ZU0VTs/NvcycmVKKAEFsWZhYB7gHOBUqBxWb2uLuvainj7l+KKX8TMD3YrAY+4+7rzWwosMTMnnb38rDiFRGJt6q6RgBy43SX9pGE2bKYCWxw9xJ3rwfmAxceofxlwCMA7r7O3dcHz7cBu4DCEGMVEYm7yl6aLIYBW2K2S4N972FmI4HRwPNtvDYTSAc2tvHadWZWbGbFZWVlXRK0iEi8tLQscjK6f1bZjiTKaKh5wAJ3b4rdaWZDgN8CV7l7c+uD3P0+dy9y96LCQjU8RCS5VdVHvwLjNVngkYSZLLYCI2K2hwf72jKP4BJUCzPLA/4KfMPdXwslQhGRBFJZ10hqipGRmii/4w8JM6LFwDgzG21m6UQTwuOtC5nZRKAfsChmXzrwJ+A37r4gxBhFRBJGVV0jORmpmFm8Q3mP0JKFuzcCNwJPA6uBR919pZndbmYXxBSdB8x3d4/ZdynwQeDKmKG108KKVUQkEVTWNSZk5zaEfFOeuy8EFrba961W299p47jfAb8LMzYRkUQTbVkkXuc2JE4Ht4hIr1dV15SQndugZCEikjAS+TKUkoWISIKoqmskJ13JQkREjqBlNFQiUrIQEUkQ0ctQ6uAWEZF2uDtV9ergFhGRI6hrbKap2ZUsRESkfYk84ywoWYiIJIRDM84qWYiISDsOtSzUwS0iIu2oqkvc6clByUJEJCHoMpSIiHRIHdwiItIhtSxERKRDB1sWmhtKRETac6iDW6OhRESkHVX1jWSkppAaScyv5cSMSkSkl0nktSxAyUJEJCEk8vTkoGQhIpIQEnlJVVCyEBFJCFUJvJYFKFmIiCSEqnpdhhIRkQ5Uqs9CREQ6UlXXmLA35IGShYhIQlAHt4iIHFF0/W11cIuIyBFU1zfhnriTCIKShYhI3CX6jLMQcrIws7lmttbMNpjZbW28/lMzWxo81plZecxrnzWz9cHjs2HGKSIST4m+lgVAaJGZWQS4BzgXKAUWm9nj7r6qpYy7fymm/E3A9OB5AfBtoAhwYElw7L6w4hURiZdEX1IVQkwWwExgg7uXAJjZfOBCYFU75S8jmiAAPgw86+57g2OfBeYCj4QYb+iamp21OypY8u5e3tpcTmZ6hElD8pg8NI+Jg/PISk/czi0RCU/lwctQifsdEGayGAZsidkuBWa1VdDMRgKjgeePcOywEGI8ZrsO1PLm5nJKdleSkRohJz1CdkYq6RFjd2U9O/bXsn1/LdvKa1i+df/BP4rCPhnUNTTx8OubAUgxGD+oDzNG9qNoVD+KRhYwvF8WZtapONydirpG0iMpZKYd+oOrb2xmWWk5/9i4h0Ub91Be08DgvAwG52cyOC+LQXkZDMjNoH9uOgNyMyjsk3HY8SISvqrefBnqfZoHLHD3pvdzkJldB1wHcNxxx4UR12Gamp01Ow6weNNeit/dx1uby9laXnPEYyIpxsA+GQzKy+TCaUMPSwQApftqWLntAKu27eetLeX8eek2HgoSSF5mKgNyM+ibnUZBTjp5mWmHJY+m5mbKKuvYXh5NSDUN0f99mWkp9MuOlt+8t5qahibMYNKQPIbmZ7KzopblW/ezu7K+zXhPGdWPsycO5OyJgxhbmNPphCUiR6eqPvE7uMOMbCswImZ7eLCvLfOAG1odO6fVsS+2Psjd7wPuAygqKvKjD/VwDU3N7Nhfy5a91WzZV82WvTWs2LafJe/soyL4BTA4L5MZI/tx1emjmH5cPyYO7kNjU3SsdHV9E3WNTQzIjf5qj6S0/2U7oiCbEQXZzJ0yGDj8UtW6nZXsq65nX3U928prWVNbgcfUMiUFCnMzOGFIHmdPHMigvEzqm5opr66nvLqBfdUNzBpTwGljBzB7TAF9s9MPO3ddYxO7K+vZU1nH7so6dlfWs2l3FS+s2cX3Fq7hewvXcFxBNlOG5XH8wD6MG5jL8QNzSU9Nob6xmfrGZhqamqltaKamoYnahqZownLIzoiQk55KVnr0v5lp0RZPZlqEtIhRUdtIRW0jB2obqKxtpLG5mcZmp6nZaXanb3Y6Q/IzGZKXRV5WqhKW9Gi9uoMbWAyMM7PRRL/85wGXty5kZhOBfsCimN1PA98zs37B9nnA18IIcl9VPbcueJs9VfXsq6pnb1U9B2obDysTSTFGD8jhY1OHMnN0P04ZVcDwftltvl9+dtoxxRNJMSYNzWPS0Lxjep/OyEiNMKxvFsP6Zh22/9/mTmRreQ3Pr9nFK+vKWLXtAE+t2EFzl6Xj9ycrLUL/3GhLKT8rjbysVNJTIzQ0NtPY3Ex9k5MeiX5GYwpzGVuYy3EF2aRFjBSLPlIjltC/2qR3S4ahs6FF5u6NZnYj0S/+CPArd19pZrcDxe7+eFB0HjDf/dBvZnffa2Z3EE04ALe3dHZ3tfTUFLaV11KQk87wftn0z0mnb3YaQ/IzGdEv+qt/SH5mwi51GJZhfbO4YvZIrpg9EoDahiZKyqoo2V1JU7OTHkkhPTWFtKCPJCstQlZ69Lk71DQ0UVXXSE19E1X10VZHy6OhycnNTCUvM5U+mWnkZqSSnppCJMWIpES/3PdWtfT31LB9fy37quo5UNvA/poGNu2uor6x+eD5UyMp1NY38cr63dQ1Nrdbp1H9s5k1uj+zxhQwc3QBw/p2vk9IJEyVwWio7ATuLzT3OP1c7GJFRUVeXFwc7zAkjpqbna3lNWwsq6R0X83BS1rNHk12b23exxub9h5sOWakplDYJ9qpX5ibwejCHKYN78vUEX0Zkp+pRCLd5o4nVjH/jc2svH1ut5/bzJa4e1FH5RK3zSPyPqWk2ME+oPa0DFIofmcfW8trKKuoo6yijnf2VPHC2l00NEV/PA3IzaBoZD/OmljInAnR/iCRsCT6kqqgZCG9TCTFmDw0n8lD89/zWl1jE6u3V7CstJylm8tZVLKHp1buAGDy0DzOmjCQsyYWMm1EvyMOWhB5vyrrGhO6cxuULEQOykiNMG1EX6aN6MtnTo3eu7J2ZwXPr9nFi2vK+PlLG7n7hQ30zU7jzPGFnHPCIOZOHkx6au/qz5Ku12NaFmY2Fih19zozmwOcBPzG3cuPfKRI8jIzJg6O3l1//Zzj2V/dwCsbynh+zS5eWlvGn5duY0RBFl84exwXTR/W6wZBSNeJrmWRuJ3b0PmWxR+AIjM7nuh9DX8GHgY+ElZgIokmPzuNj500lI+dNJTmZueldWX8+Nm13LpgGT9/cSM3nXM8EwblkZ5qpAWjxQb1ySRFl6ykA5V1jQztm9j9Yp1NFs3BUNiLgLvc/S4zeyvMwEQSWUqKcdbEgcyZUMgzq3byk2fW8aXfv/2ecicMyePWD4/nrAkDNbpK2lVV30MuQwENZnYZ8Fng48G+Y7v7TKQHMDM+PHkw554wiNc37WV/TQMNTdE72/dVN/DgP97h6geKOWVUP746dyKnjCqId8iSgHpMnwVwFfA54Lvuvim4K/u34YUlklxSUoxTx/Z/z/4rZo/k98VbuOtv67nkF4sYmp8ZvRExM5XcjFQKctIZ1jeL4f2yGN4vm/GDchmoYbq9To8ZDRWsQfEFgGAKjj7u/p9hBibSE6SnpnDF7JF88uThPPT6u6zeXkFlXQOVdY2UV9ezYVcl2/fXHJxKJS1i/PtHJ/GZU0fqslUv0RjMr5aT3gOShZm9CFwQlF8C7DKzV939lhBjE+kxstIjXHPGmDZfa5m4snRfDf/7Sgnffnwlb7yzlx9cfCJ9MnW1t6erqm9Z+CixR0N1dqxfvrsfAC4mOmR2FvCh8MIS6T3SIimMKMjm1LH9uf8zRdx2/kSeWrGDC+5+ldXbD8Q7PAlZMqxlAZ1PFqlmNgS4FHgixHhEerWUFONzZ47l4WtmUVXXyCfueZUX1+6Kd1gSomSYcRY6nyxuJzp77EZ3X2xmY4D14YUl0rvNGtOfhV88g7GFuVz32yW8vK4s3iFJSJJhLQvoZLJw9/9z95Pc/fPBdom7/1O4oYn0bgNyM3jomlmMLczl2t8U88p6JYyeqKqupc+iByQLMxtuZn8ys13B4w9mNjzs4ER6u3456Tx0zSxGD8jhmgeLeXXD7niHJF2s8uBlqJ7Rwf1r4HFgaPD4S7BPREJWkJPOw9fOZvSAHP7lwcX8Y6MSRk/S0zq4C9391+7eGDweAApDjEtEYhQELYzjCrL5lweKWfxOKAtHShxU1fesDu49ZvbPZhYJHv8M7AkzMBE5XP/cDH53zSyG5Gdy1a8Xs3SLJn3uCXpUBzdwNdFhszuA7cAngStDiklE2jGwTyYPXzubgpx0PvPL11mxdX+8Q5JjVFXXSCTFyEjwdVE6OxrqXXe/wN0L3X2gu38C0GgokTgYnJ/Jw9fOok9mGlf88nWeX7OTppb5QlrZWFbJ5j3V3RyhvB9VdU3kpEcSfnqXY2n33AL8rKsCEZHOG94vm4evncXl97/O1Q8UM7BPBhdNH8bFJ0cHKS5cvp0nV2xn3c5KUlOM7198IpcUjYhz1NKWZJhEEI4tWSR2GhTp4Ub2z+H5r5zJC2t2sWDJVn75903c+3IJACkGM0cX8B8XTOa51Tu5dcEy3t1TzS3njn/PYkzunvC/anuyZJieHI4tWbTd7hWRbpORGmHulCHMnTKEPZV1LFy+nUhKCudOGkRhnwwALp91HN98bAV3v7CBd/dW86NPnkTpvmqeXrmTZ1btZP3OCr54zjiuPWOMVvWLg8qekCzMrIK2k4IBWaFEJCJHpX9uBlecOuo9+9MiKXz/4hMZNSCHHzy5hhfX7KIiGIEzdURfZozsx/efXMPrm/by40um0i8nvZsj792qesJlKHfv012BiEh4zKITFI7qn8Nflm1j9pj+nHvCIAbnZ+Lu/GbRu3z3r6v56J2vcNflJzNjZL94h9xrVNU1HWwFJrLEHqslIl1q7pTB3HP5yVwxeySD86Mr8pkZnz1tFAs+fyqRiPGpexfxxLJtcY6090iWy1BKFiICwEnD+/LETWcwbURfvvzo27rpr5tU1SfHZSglCxE5KD8rjXuvmMHAvAyuebCYreU18Q6px0uW0VBKFiJymP65Gfzys6dQ19DENQ8WH5zoTrpeXWMTDU2uloWZzTWztWa2wcxua6fMpWa2ysxWmtnDMft/GOxbbWZ3mgaCi3Sb8YP6cPenT2btjgN8cf7Sdu8Ql2NzcC2L9MSenhxCTBZmFgHuAc4HJgGXmdmkVmXGAV8DTnf3ycDNwf7TgNOBk4ApwCnAmWHFKiLvdeb4Qr4T3NT3gydXxzucHilZllSFY7spryMzgQ3uXgJgZvOBC4FVMWWuBe5x930A7t6y2LADmUA60Xs60oCdIcYqIm34zKmjKCmr4v5XNjGyfw7/PHtkvEPqUZJlxlkI9zLUMGBLzHZpsC/WeGC8mb1qZq+Z2VwAd18EvEB0htvtwNPurp82InHwzY9N4uyJA/n24yt5SWuBd6lkalnEu4M7FRgHzAEuA+43s75mdjxwAjCcaII528zOaH2wmV1nZsVmVlxWpj9ikTBEUoy7LpvOhEF9uOGhN1mz40C8Q+oxKpUsANgKxE5zOTzYF6sUeNzdG9x9E7COaPK4CHjN3SvdvRJ4Eji19Qnc/T53L3L3osJCLdwnEpacjFR+eWURORkRrv71YnYdqI13SD1CSwd3b78MtRgYZ2ajzSwdmEd0He9YjxFtVWBmA4helioBNgNnmlmqmaUR7dzWZSiROBqSn8UvP3sK5TUNXPObYmrqm+IdUtI7dBmqF4+GcvdG4EbgaaJf9I+6+0ozu93MLgiKPU10ydZVRPsobnX3PcACYCOwHHgbeNvd/xJWrCLSOVOG5XPnvOks37qfWxe8jbuG1B6Llgkd+2SkxTmSjoXa9nH3hcDCVvu+FfPciS6idEurMk3Av4YZm4gcnQ9NGsS/zZ3ID55cw7iBffjih8bFO6SkVVkbjIbKTPzLUIkfoYgknH/94BjW7azgp8+t4/iBuXz0pCHxDikpVdY1kJ0eIZIE64jEezSUiCQhs+hSrTNG9uPL/7eU5aX74x1SUkqWJVVByUJEjlJGaoR7r5hB/5wMPve7JVTXaw6p9+tAbWNSXIICJQsROQYDcjP46aemsbW8hp+/uDHe4SSdytpG+qhlISK9wczRBVw4bSj3vlzCu3uq4h1OUqmsa6RPZuKPhAIlCxHpAl//yAmkpRh3PLGq48JyUGWt+ixEpBcZlJfJTeeM47nVu3hh7a6ODxAg6OBWn4WI9CZXnz6aMQNyuP0vq6hr1N3dnXGgtkEtCxHpXdJTU/j2BZPZtLuKX/59U7zDSXjuHvRZKFmISC9z5vhCzp00iDv/tp4nl2+PdzgJrbq+CXeULESkd/ruRVOYODiPzz/0Jt9buJrGpuZ4h5SQDi18pNFQItILDeyTye//dTZXzB7JfS+X8On/fZ1dFZrSvLWKJJoXCpQsRCQEGakR7vjEFH5y6VTeLi3nY3f+nRVbNSVIrIraBgDdlCcicvHJw/nT9aeTFklh3n2v8Y8Nu+MdUsJouQylPgsREeCEIXn84fOnMaxvFlf+ejEL1fENJNf05KBkISLdYHB+Jo/+66mcNDyfGx5+k9++9m68Q4q7ioMd3EoWIiIH5Wen8dt/mcU5EwfyzcdW8NSKHfEOKa5aOriTYZU8ULIQkW6UlR7hF/88g+H9snj4jc3xDieuWi5DJcP626BkISLdLDWSwiemDePv68vYdaD3DqltWSUvNZIcX8PJEaWI9CgXnTyMZofH394W71DiJplWyQMlCxGJg7GFuUwd0Zc/vLk13qHETUUSrZIHShYiEicXTx/G6u0HWLPjQLxDiYuKJFolD5QsRCROPj51KKkpxp96aesimdayACULEYmTgpx05kwo5LGlW2lq9niH0+2i628nx7BZULIQkTi6aPpwdh6oY9HGPfEOpdupZSEi0knnnDCQPpmp/PGt0niH0u0qkmiVPFCyEJE4ykyL8NETh/DUih1U1zfGO5xuk2yr5IGShYjE2UXTh1Fd38TTK3vP9B/V9U00e/LMCwVKFiISZ6eMKmBU/2x++fdNuPeOju5D05OrgxsAM5trZmvNbIOZ3dZOmUvNbJWZrTSzh2P2H2dmz5jZ6uD1UWHGKiLxkZJi3HDW8azYeoBnV+2MdzjdItlWyYMQk4WZRYB7gPOBScBlZjapVZlxwNeA0919MnBzzMu/AX7k7icAM4FdYcUqIvF10fRhjOqfzc+eW98rWhcHWxa6DAVEv+A3uHuJu9cD84ELW5W5FrjH3fcBuPsugCCppLr7s8H+SnevDjFWEYmj1EgKN509jlXbD/BML2hdtCypqpZF1DBgS8x2abAv1nhgvJm9amavmdncmP3lZvZHM3vLzH4UtFREpIe6cNpQRg/I4WfPrae5h9+k1zI9uUZDdV4qMA6YA1wG3G9mfYP9ZwBfAU4BxgBXtj7YzK4zs2IzKy4rK+uumEUkBNHWxfGs3n6AZ1b17JFRybZKHoSbLLYCI2K2hwf7YpUCj7t7g7tvAtYRTR6lwNLgElYj8BhwcusTuPt97l7k7kWFhYWhVEJEus8FU4cyphe0LiqTbJU8CDdZLAbGmdloM0sH5gGPtyrzGNFWBWY2gOjlp5Lg2L5m1pIBzgZWhRiriCSA1EgKXzhnHGt2VPBUD77voqWDO1lWyYMQk0XQIrgReBpYDTzq7ivN7HYzuyAo9jSwx8xWAS8At7r7HndvInoJ6m9mthww4P6wYhWRxPHxqUM5fmAu/7ZgGY+91X0z0u6qqGVXRfes3FdR20BWWvKskgfRvoHQuPtCYGGrfd+Kee7ALcGj9bHPAieFGZ+IJJ5IivHAVadw8/yl3Pz7pbywdhd3fGIKeV18A9uy0nKeWbmTFdv2s3LbAcoq6sjPSuO5W86ksE9Gl56rtWSb6gPi38EtIvIew/tlM/+62dxy7nieWLad83/2Covf2dsl793Y1MzPnlvHJ+55lZ+/tJEd+2s5Y9wAbv3wBKrrG/n+wtVdcp4jSbZV8iDkloWIyNFq6b/4wLgB3Dx/KZ+6dxE3nHU8XzhnHGlHeflmW3kNN89fyhvv7OXi6cP4zoWTD2ux1NQ3cfcLG7ikaASnju3fVVV5j8q65FolD9SyEJEEd/Jx/Vj4xTO4+OTh3PX8Bj75i0W8s7uq08fXNTaxdkcFj7yxmfP/+xVWbtvPTy6dyk8+Ne09l7ZuPPt4RhRk8c0/r6C+sfmI71vb0MTPnlvH+p0V77tOalmIiIQgNyOV/7pkKnMmFPL1Py7nI3e+wn9cMJlLika0Wb6mvolv/nkFS97dx7t7qmgZhXvS8HzunDedUQNy2jwuMy3C7RdM4aoHFnP/KyXccNbxbZZranZunr+Up1bu4KHXN/PHz5/GiILsTtensraRAbmdL58I1LIQkaTxsZOG8tTNH+Sk4fncumAZTy7f3ma5Hzy5mgVLShk/KJcbzjqe/543jSdu+gB/uv70dhNFi7MmDmTu5MHc9fx6tux97yxD7s6/P7acp1bu4JoPjKauoYnP/uoN9lTWdboe0Q7u5LnHApQsRCTJDO2bxW+unsXUEX356oJlbN5z+Bf6C2t38eCid7n69NHce0URXz5vAhdOG8aUYflEUqxT5/jWxyeRYsY3HlvB7lZJ4L+eWcsjb2zhhrPG8u8fm8SvrjyFreU1XP1gcacXcEq2VfJAyUJEklB6agp3XzYdM7jxkTepa2wCYG9VPV9dsIzxg3L56twJR/3+Q/tm8dUPT+DldWXM/O5zXHrvIn71903c9bf13PPCRi6bOYKvnBd9/6JRBdx9+cksLy3n8797k4amI/d1JOMqeaBkISJJakRBNj+6ZCrLSvfz/YVrcHdu+8My9lc38LNPTScz7djujr7y9NEs/MIZ3Hj2OPZXN3D7E6v48bPrmDt5MP/vEydidqiVcu6kQXzvohN5aV0Zd/1t/RHfNxlXyQN1cItIEvvw5MFcdfoofv3qO+yvaeCZVTv5+kcmMmloXpe8/6SheUwamsct546npKySt0vLOX/KkDYvZ82beRx/Xb6dvy7fzi3ntd+qaZnqI9lGQ6llISJJ7Wvnn8DU4fn86a2tnDqmP9d8YEwo5xlTmMtF04cfscUyZ8JANpZVtdkx3qKiNvnrKaymAAALB0lEQVSWVAUlCxFJcumpKdx9+clcWjScn3xqKimd7MQOw5wJ0blPX1zX/pIJybhKHihZiEgPMKIgmx9+cipD8rPiGseYATmMKMjipbXtrwJdmYTrb4OShYhIlzEz5owfyD827jk4Qqu1g0uqqmUhItJ7zZlQSHV9E4s37Wvz9WRcJQ+ULEREutSpY/uTHknhhXYuRbVchurqKdfDpmQhItKFstNTmTWmgBfbSxZJuEoeKFmIiHS5M8cXtjuEtrKuMelWyQMlCxGRLjdnwkCg7SG0FbUNSTcSCpQsRES63NjCHIb3a3sIbUVt8s0LBUoWIiJdzsw4c3xhm0Nok3GVPFCyEBEJxZwJA9scQluZhKvkgZKFiEgoTguG0LYeFVVZ15h091iAkoWISChyMlKZObqA51bvxN0P7q+obSQ3I7nusQAlCxGR0MydMph39lSzenvFwX0VtQ3q4BYRkUPmThlMisHCYK3wZF0lD5QsRERCMyA3g9lj+rNw+XbcnZqG5FwlD5QsRERC9ZETh1Cyu4o1OyqSdnpyULIQEQlV7KWoA7XJOeMsKFmIiIRqQG4Gs0b356/Ltx9cyyLZZpyFkJOFmc01s7VmtsHMbmunzKVmtsrMVprZw61eyzOzUjO7O8w4RUTC9JGThlBSVsWSd6M36OkyVAwziwD3AOcDk4DLzGxSqzLjgK8Bp7v7ZODmVm9zB/ByWDGKiHSHuZOjl6IeLd4C6DJUazOBDe5e4u71wHzgwlZlrgXucfd9AO5+8FZHM5sBDAKeCTFGEZHQFfbJYOboAtbtrASULFobBmyJ2S4N9sUaD4w3s1fN7DUzmwtgZinAj4GvHOkEZnadmRWbWXFZ2XunAhYRSRQfPXHIwee6z+L9SwXGAXOAy4D7zawvcD2w0N1Lj3Swu9/n7kXuXlRYWBh6sCIiR+vDUwZjFn2ejC2LMCPeCoyI2R4e7ItVCrzu7g3AJjNbRzR5nAqcYWbXA7lAuplVunubneQiIoluYJ9MZo0uYOmW8qRbJQ/CTRaLgXFmNppokpgHXN6qzGNEWxS/NrMBRC9Llbj7p1sKmNmVQJEShYgku1s/PJG3t5THO4yjElqycPdGM7sReBqIAL9y95VmdjtQ7O6PB6+dZ2argCbgVnffE1ZMIiLxNGNkP2aM7BfvMI6KxU6dm8yKioq8uLg43mGIiCQVM1vi7kUdlUu+C2ciItLtlCxERKRDShYiItIhJQsREemQkoWIiHRIyUJERDqkZCEiIh3qMfdZmFkZ8G6r3fnA/ve5r6PnA4DdxxBqW+fvbJnO7k+W+rT3murTM+oTux27/1jq1N31ab0dj/ocqVxX1Kevu3c8uZ6799gHcN/73dfRc6J3n3dpTJ0t09n9yVKfzn5Gqk9y1qdVPWLLHHWdurs+R/hcuq0+RyrXVfXpzKOnX4b6y1Hs68zzY9GZ92mvTGf3J0t92ntN9ekZ9YndTtb6tN6OR32OVK6r6tOhHnMZqruYWbF34tb4ZKH6JLaeVh/oeXXqafVpT09vWYThvngH0MVUn8TW0+oDPa9OPa0+bVLLQkREOqSWhYiIdKhXJwsz+5WZ7TKzFUdx7AwzW25mG8zsTrOWBRPBzG4yszVmttLMfti1UR8xpi6vj5l9x8y2mtnS4PGRro+83ZhC+XyC179sZh4sutUtQvp87jCzZcFn84yZDe36yNuNKYz6/Cj4t7PMzP4ULLPcLUKqzyXB90CzmSV3v8bRDvnqCQ/gg8DJwIqjOPYNYDZgwJPA+cH+s4DngIxge2CS1+c7wFd6yucTvDaC6MJb7wIDkrk+QF5MmS8Av0jy+pwHpAbP/xP4zySvzwnABOBFoit+dktdwnj06paFu78M7I3dZ2ZjzewpM1tiZq+Y2cTWx5nZEKL/SF/z6F/Eb4BPBC9/HviBu9cF59gVbi0OCak+cRNifX4KfBXo1g67MOrj7gdiiubQjXUKqT7PuHtjUPQ1YHi4tTgkpPqsdve13RF/2Hp1smjHfcBN7j4D+ArwP22UGQaUxmyXBvsguo74GWb2upm9ZGanhBptx461PgA3BpcFfmVm8V4T8pjqY2YXAlvd/e2wA+2kY/58zOy7ZrYF+DTwrRBj7Yyu+HtrcTXRX+nx1JX1SWqhrcGdjMwsFzgN+L+YS9wZ7/NtUoECok3SU4BHzWxM8IujW3VRfX4O3EH0F+sdwI+J/iPudsdaHzPLBr5O9FJH3HXR54O7fwP4hpl9DbgR+HaXBfk+dFV9gvf6BtAIPNQ10R1VDF1Wn55AyeJwKUC5u0+L3WlmEWBJsPk40S/Q2ObxcGBr8LwU+GOQHN4ws2aic8eUhRl4O465Pu6+M+a4+4Enwgy4A8dan7HAaODt4B//cOBNM5vp7jtCjr0tXfH3FushYCFxShZ0UX3M7ErgY8A58fiRFaOrP5/kFu9Ok3g/gFHEdGgB/wAuCZ4bMLWd41p3aH0k2P854Pbg+XhgC8H9LElanyExZb4EzE/mz6dVmXfoxg7ukD6fcTFlbgIWJHl95gKrgMLurEfYf2/0gA7uuAcQ18rDI8B2oIFoi+BfiP7yfAp4O/ij/VY7xxYBK4CNwN0tCQFIB34XvPYmcHaS1+e3wHJgGdFfUUOSuT6tynRrsgjp8/lDsH8Z0Xl+hiV5fTYQ/YG1NHh05+iuMOpzUfBedcBO4Onuqk9XP3QHt4iIdEijoUREpENKFiIi0iElCxER6ZCShYiIdEjJQkREOqRkIT2amVV28/n+18wmddF7NQWzya4ws790NAOrmfU1s+u74twirWnorPRoZlbp7rld+H6pfmiiu1DFxm5mDwLr3P27Ryg/CnjC3ad0R3zSu6hlIb2OmRWa2R/MbHHwOD3YP9PMFpnZW2b2DzObEOy/0sweN7Pngb+Z2Rwze9HMFgRrLzwUs37Biy3rFphZZTDJ39tm9pqZDQr2jw22l5vZ/+tk62cRhyZDzDWzv5nZm8F7XBiU+QEwNmiN/Cgoe2tQx2Vm9h9d+L9RehklC+mN/hv4qbufAvwT8L/B/jXAGe4+nejsrd+LOeZk4JPufmawPR24GZgEjAFOb+M8OcBr7j4VeBm4Nub8/+3uJ3L4bKVtCuYiOofoHfQAtcBF7n4y0fVTfhwkq9uAje4+zd1vNbPzgHHATGAaMMPMPtjR+UTaookEpTf6EDApZibRvGCG0XzgQTMbR3SW3bSYY55199i1Dt5w91IAM1tKdE6hv7c6Tz2HJl5cApwbPD+VQ+trPAz8VztxZgXvPQxYDTwb7Dfge8EXf3Pw+qA2jj8veLwVbOcSTR4vt3M+kXYpWUhvlALMdvfa2J1mdjfwgrtfFFz/fzHm5apW71EX87yJtv8tNfihTsH2yhxJjbtPC6ZWfxq4AbiT6LoVhcAMd28ws3eAzDaON+D77n7v+zyvyHvoMpT0Rs8QnaEVADNrmYI6n0NTS18Z4vlfI3r5C2BeR4XdvZrokqlfNrNUonHuChLFWcDIoGgF0Cfm0KeBq4NWE2Y2zMwGdlEdpJdRspCeLtvMSmMetxD94i0KOn1XEZ1WHuCHwPfN7C3CbXXfDNxiZsuA44H9HR3g7m8RnVn2MqLrVhSZ2XLgM0T7WnD3PcCrwVDbH7n7M0Qvcy0Kyi7g8GQi0mkaOivSzYLLSjXu7mY2D7jM3S/s6DiReFKfhUj3mwHcHYxgKidOy9SKvB9qWYiISIfUZyEiIh1SshARkQ4pWYiISIeULEREpENKFiIi0iElCxER6dD/B1W4Iryp+ijGAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.recorder.plot(skip_end=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "Total time: 08:21 <p><table style='width:375px; margin-bottom:10px'>\n",
       "  <tr>\n",
       "    <th>epoch</th>\n",
       "    <th>train_loss</th>\n",
       "    <th>valid_loss</th>\n",
       "    <th>accuracy</th>\n",
       "    <th>time</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>0</th>\n",
       "    <th>0.532072</th>\n",
       "    <th>0.501664</th>\n",
       "    <th>0.776389</th>\n",
       "    <th>00:22</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>1</th>\n",
       "    <th>0.537432</th>\n",
       "    <th>0.449404</th>\n",
       "    <th>0.785000</th>\n",
       "    <th>00:22</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>2</th>\n",
       "    <th>0.452703</th>\n",
       "    <th>0.380878</th>\n",
       "    <th>0.830000</th>\n",
       "    <th>00:23</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>3</th>\n",
       "    <th>0.451064</th>\n",
       "    <th>0.372108</th>\n",
       "    <th>0.828611</th>\n",
       "    <th>00:22</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>4</th>\n",
       "    <th>0.379762</th>\n",
       "    <th>0.378799</th>\n",
       "    <th>0.828333</th>\n",
       "    <th>00:26</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>5</th>\n",
       "    <th>0.402086</th>\n",
       "    <th>0.391328</th>\n",
       "    <th>0.828889</th>\n",
       "    <th>00:23</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>6</th>\n",
       "    <th>0.402177</th>\n",
       "    <th>0.385814</th>\n",
       "    <th>0.837500</th>\n",
       "    <th>00:22</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>7</th>\n",
       "    <th>0.364293</th>\n",
       "    <th>0.386650</th>\n",
       "    <th>0.829444</th>\n",
       "    <th>00:25</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>8</th>\n",
       "    <th>0.388546</th>\n",
       "    <th>0.426296</th>\n",
       "    <th>0.833333</th>\n",
       "    <th>00:23</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>9</th>\n",
       "    <th>0.312414</th>\n",
       "    <th>0.434039</th>\n",
       "    <th>0.830000</th>\n",
       "    <th>00:24</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>10</th>\n",
       "    <th>0.300423</th>\n",
       "    <th>0.447511</th>\n",
       "    <th>0.833611</th>\n",
       "    <th>00:24</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>11</th>\n",
       "    <th>0.263376</th>\n",
       "    <th>0.448432</th>\n",
       "    <th>0.834722</th>\n",
       "    <th>00:24</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>12</th>\n",
       "    <th>0.254895</th>\n",
       "    <th>0.450464</th>\n",
       "    <th>0.837222</th>\n",
       "    <th>00:24</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>13</th>\n",
       "    <th>0.258895</th>\n",
       "    <th>0.446566</th>\n",
       "    <th>0.828333</th>\n",
       "    <th>00:23</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>14</th>\n",
       "    <th>0.225605</th>\n",
       "    <th>0.475200</th>\n",
       "    <th>0.832222</th>\n",
       "    <th>00:25</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>15</th>\n",
       "    <th>0.203444</th>\n",
       "    <th>0.514545</th>\n",
       "    <th>0.834167</th>\n",
       "    <th>00:25</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>16</th>\n",
       "    <th>0.214901</th>\n",
       "    <th>0.502333</th>\n",
       "    <th>0.832778</th>\n",
       "    <th>00:26</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>17</th>\n",
       "    <th>0.218773</th>\n",
       "    <th>0.467870</th>\n",
       "    <th>0.828056</th>\n",
       "    <th>00:22</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>18</th>\n",
       "    <th>0.201399</th>\n",
       "    <th>0.467877</th>\n",
       "    <th>0.828889</th>\n",
       "    <th>00:24</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>19</th>\n",
       "    <th>0.181447</th>\n",
       "    <th>0.527198</th>\n",
       "    <th>0.827222</th>\n",
       "    <th>00:36</th>\n",
       "  </tr>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Better model found at epoch 0 with accuracy value: 0.7763888835906982.\n",
      "Better model found at epoch 1 with accuracy value: 0.7850000262260437.\n",
      "Better model found at epoch 2 with accuracy value: 0.8299999833106995.\n",
      "Better model found at epoch 6 with accuracy value: 0.8374999761581421.\n"
     ]
    }
   ],
   "source": [
    "## replace with attention head for classifier?\n",
    "learn.freeze()\n",
    "learn.fit_one_cycle(20, 1e-3, moms=(0.8,0.7),\n",
    "                   callbacks=[SaveModelCallback(learn,every='improvement',mode='max',monitor='accuracy',name='best_acc_model')])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.recorder.plot_losses()\n",
    "learn.recorder.plot_metrics()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn.save('Tfm_trained_head')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#learn = learn.load('rnn_trained_head')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "Total time: 15:02 <p><table style='width:375px; margin-bottom:10px'>\n",
       "  <tr>\n",
       "    <th>epoch</th>\n",
       "    <th>train_loss</th>\n",
       "    <th>valid_loss</th>\n",
       "    <th>accuracy</th>\n",
       "    <th>time</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>0</th>\n",
       "    <th>0.403401</th>\n",
       "    <th>0.405786</th>\n",
       "    <th>0.823611</th>\n",
       "    <th>01:00</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>1</th>\n",
       "    <th>0.485345</th>\n",
       "    <th>0.416100</th>\n",
       "    <th>0.820556</th>\n",
       "    <th>01:00</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>2</th>\n",
       "    <th>0.438312</th>\n",
       "    <th>0.415085</th>\n",
       "    <th>0.828333</th>\n",
       "    <th>00:29</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>3</th>\n",
       "    <th>0.440661</th>\n",
       "    <th>1.419594</th>\n",
       "    <th>0.657222</th>\n",
       "    <th>00:30</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>4</th>\n",
       "    <th>0.364594</th>\n",
       "    <th>0.345021</th>\n",
       "    <th>0.855000</th>\n",
       "    <th>00:30</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>5</th>\n",
       "    <th>0.339439</th>\n",
       "    <th>0.466150</th>\n",
       "    <th>0.830833</th>\n",
       "    <th>00:27</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>6</th>\n",
       "    <th>0.284175</th>\n",
       "    <th>0.431246</th>\n",
       "    <th>0.848333</th>\n",
       "    <th>00:26</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>7</th>\n",
       "    <th>0.201899</th>\n",
       "    <th>0.480610</th>\n",
       "    <th>0.849722</th>\n",
       "    <th>00:45</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>8</th>\n",
       "    <th>0.157626</th>\n",
       "    <th>0.944617</th>\n",
       "    <th>0.830833</th>\n",
       "    <th>00:58</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>9</th>\n",
       "    <th>0.135186</th>\n",
       "    <th>1.519732</th>\n",
       "    <th>0.828611</th>\n",
       "    <th>00:58</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>10</th>\n",
       "    <th>0.085516</th>\n",
       "    <th>1.062499</th>\n",
       "    <th>0.821667</th>\n",
       "    <th>01:01</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>11</th>\n",
       "    <th>0.065590</th>\n",
       "    <th>0.852063</th>\n",
       "    <th>0.850833</th>\n",
       "    <th>01:01</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>12</th>\n",
       "    <th>0.045483</th>\n",
       "    <th>0.847496</th>\n",
       "    <th>0.843056</th>\n",
       "    <th>01:04</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>13</th>\n",
       "    <th>0.049319</th>\n",
       "    <th>0.704194</th>\n",
       "    <th>0.853333</th>\n",
       "    <th>00:26</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>14</th>\n",
       "    <th>0.034208</th>\n",
       "    <th>0.969585</th>\n",
       "    <th>0.843333</th>\n",
       "    <th>00:28</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>15</th>\n",
       "    <th>0.028909</th>\n",
       "    <th>0.857532</th>\n",
       "    <th>0.855000</th>\n",
       "    <th>00:32</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>16</th>\n",
       "    <th>0.023507</th>\n",
       "    <th>0.784188</th>\n",
       "    <th>0.853056</th>\n",
       "    <th>01:00</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>17</th>\n",
       "    <th>0.022352</th>\n",
       "    <th>0.771314</th>\n",
       "    <th>0.854167</th>\n",
       "    <th>01:00</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>18</th>\n",
       "    <th>0.018394</th>\n",
       "    <th>0.822610</th>\n",
       "    <th>0.851944</th>\n",
       "    <th>00:41</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>19</th>\n",
       "    <th>0.026852</th>\n",
       "    <th>0.912473</th>\n",
       "    <th>0.848056</th>\n",
       "    <th>00:29</th>\n",
       "  </tr>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Better model found at epoch 0 with accuracy value: 0.8236111402511597.\n",
      "Better model found at epoch 2 with accuracy value: 0.82833331823349.\n",
      "Better model found at epoch 4 with accuracy value: 0.8550000190734863.\n"
     ]
    },
    {
     "ename": "RuntimeError",
     "evalue": "Error(s) in loading state_dict for SequentialRNN:\n\tMissing key(s) in state_dict: \"0.module.pos_enc.weight\", \"0.module.layers.0.mhra.attention.weight\", \"0.module.layers.0.mhra.attention.bias\", \"0.module.layers.0.mhra.out.weight\", \"0.module.layers.0.mhra.out.bias\", \"0.module.layers.0.mhra.ln.weight\", \"0.module.layers.0.mhra.ln.bias\", \"0.module.layers.0.ff.layers.0.weight\", \"0.module.layers.0.ff.layers.0.bias\", \"0.module.layers.0.ff.layers.2.weight\", \"0.module.layers.0.ff.layers.2.bias\", \"0.module.layers.0.ff.layers.5.weight\", \"0.module.layers.0.ff.layers.5.bias\", \"0.module.layers.1.mhra.attention.weight\", \"0.module.layers.1.mhra.attention.bias\", \"0.module.layers.1.mhra.out.weight\", \"0.module.layers.1.mhra.out.bias\", \"0.module.layers.1.mhra.ln.weight\", \"0.module.layers.1.mhra.ln.bias\", \"0.module.layers.1.ff.layers.0.weight\", \"0.module.layers.1.ff.layers.0.bias\", \"0.module.layers.1.ff.layers.2.weight\", \"0.module.layers.1.ff.layers.2.bias\", \"0.module.layers.1.ff.layers.5.weight\", \"0.module.layers.1.ff.layers.5.bias\", \"0.module.layers.2.mhra.attention.weight\", \"0.module.layers.2.mhra.attention.bias\", \"0.module.layers.2.mhra.out.weight\", \"0.module.layers.2.mhra.out.bias\", \"0.module.layers.2.mhra.ln.weight\", \"0.module.layers.2.mhra.ln.bias\", \"0.module.layers.2.ff.layers.0.weight\", \"0.module.layers.2.ff.layers.0.bias\", \"0.module.layers.2.ff.layers.2.weight\", \"0.module.layers.2.ff.layers.2.bias\", \"0.module.layers.2.ff.layers.5.weight\", \"0.module.layers.2.ff.layers.5.bias\", \"0.module.layers.3.mhra.attention.weight\", \"0.module.layers.3.mhra.attention.bias\", \"0.module.layers.3.mhra.out.weight\", \"0.module.layers.3.mhra.out.bias\", \"0.module.layers.3.mhra.ln.weight\", \"0.module.layers.3.mhra.ln.bias\", \"0.module.layers.3.ff.layers.0.weight\", \"0.module.layers.3.ff.layers.0.bias\", \"0.module.layers.3.ff.layers.2.weight\", \"0.module.layers.3.ff.layers.2.bias\", \"0.module.layers.3.ff.layers.5.weight\", \"0.module.layers.3.ff.layers.5.bias\", \"0.module.layers.4.mhra.attention.weight\", \"0.module.layers.4.mhra.attention.bias\", \"0.module.layers.4.mhra.out.weight\", \"0.module.layers.4.mhra.out.bias\", \"0.module.layers.4.mhra.ln.weight\", \"0.module.layers.4.mhra.ln.bias\", \"0.module.layers.4.ff.layers.0.weight\", \"0.module.layers.4.ff.layers.0.bias\", \"0.module.layers.4.ff.layers.2.weight\", \"0.module.layers.4.ff.layers.2.bias\", \"0.module.layers.4.ff.layers.5.weight\", \"0.module.layers.4.ff.layers.5.bias\", \"0.module.layers.5.mhra.attention.weight\", \"0.module.layers.5.mhra.attention.bias\", \"0.module.layers.5.mhra.out.weight\", \"0.module.layers.5.mhra.out.bias\", \"0.module.layers.5.mhra.ln.weight\", \"0.module.layers.5.mhra.ln.bias\", \"0.module.layers.5.ff.layers.0.weight\", \"0.module.layers.5.ff.layers.0.bias\", \"0.module.layers.5.ff.layers.2.weight\", \"0.module.layers.5.ff.layers.2.bias\", \"0.module.layers.5.ff.layers.5.weight\", \"0.module.layers.5.ff.layers.5.bias\", \"0.module.layers.6.mhra.attention.weight\", \"0.module.layers.6.mhra.attention.bias\", \"0.module.layers.6.mhra.out.weight\", \"0.module.layers.6.mhra.out.bias\", \"0.module.layers.6.mhra.ln.weight\", \"0.module.layers.6.mhra.ln.bias\", \"0.module.layers.6.ff.layers.0.weight\", \"0.module.layers.6.ff.layers.0.bias\", \"0.module.layers.6.ff.layers.2.weight\", \"0.module.layers.6.ff.layers.2.bias\", \"0.module.layers.6.ff.layers.5.weight\", \"0.module.layers.6.ff.layers.5.bias\", \"0.module.layers.7.mhra.attention.weight\", \"0.module.layers.7.mhra.attention.bias\", \"0.module.layers.7.mhra.out.weight\", \"0.module.layers.7.mhra.out.bias\", \"0.module.layers.7.mhra.ln.weight\", \"0.module.layers.7.mhra.ln.bias\", \"0.module.layers.7.ff.layers.0.weight\", \"0.module.layers.7.ff.layers.0.bias\", \"0.module.layers.7.ff.layers.2.weight\", \"0.module.layers.7.ff.layers.2.bias\", \"0.module.layers.7.ff.layers.5.weight\", \"0.module.layers.7.ff.layers.5.bias\", \"1.layers.1.weight\", \"1.layers.1.bias\", \"1.layers.3.weight\", \"1.layers.3.bias\", \"1.layers.3.running_mean\", \"1.layers.3.running_var\", \"1.layers.5.weight\", \"1.layers.5.bias\". \n\tUnexpected key(s) in state_dict: \"0.module.encoder_dp.emb.weight\", \"0.module.rnns.0.weight_hh_l0_raw\", \"0.module.rnns.0.module.weight_ih_l0\", \"0.module.rnns.0.module.weight_hh_l0\", \"0.module.rnns.0.module.bias_ih_l0\", \"0.module.rnns.0.module.bias_hh_l0\", \"0.module.rnns.1.weight_hh_l0_raw\", \"0.module.rnns.1.module.weight_ih_l0\", \"0.module.rnns.1.module.weight_hh_l0\", \"0.module.rnns.1.module.bias_ih_l0\", \"0.module.rnns.1.module.bias_hh_l0\", \"0.module.rnns.2.weight_hh_l0_raw\", \"0.module.rnns.2.module.weight_ih_l0\", \"0.module.rnns.2.module.weight_hh_l0\", \"0.module.rnns.2.module.bias_ih_l0\", \"0.module.rnns.2.module.bias_hh_l0\", \"1.mhra.attention.weight\", \"1.mhra.attention.bias\", \"1.layers.2.weight\", \"1.layers.2.bias\". \n\tsize mismatch for 0.module.encoder.weight: copying a param with shape torch.Size([60000, 400]) from checkpoint, the shape in current model is torch.Size([60000, 512]).\n\tsize mismatch for 1.layers.0.weight: copying a param with shape torch.Size([100]) from checkpoint, the shape in current model is torch.Size([1536]).\n\tsize mismatch for 1.layers.0.bias: copying a param with shape torch.Size([100]) from checkpoint, the shape in current model is torch.Size([1536]).\n\tsize mismatch for 1.layers.0.running_mean: copying a param with shape torch.Size([100]) from checkpoint, the shape in current model is torch.Size([1536]).\n\tsize mismatch for 1.layers.0.running_var: copying a param with shape torch.Size([100]) from checkpoint, the shape in current model is torch.Size([1536]).",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mRuntimeError\u001b[0m                              Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-32-7c957a8c3dca>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mlearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfreeze_to\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m learn.fit_one_cycle(20,slice(1e-8,1e-3),moms=(0.7,0.4), wd=0.10,\n\u001b[0;32m----> 3\u001b[0;31m                    callbacks=[SaveModelCallback(learn,every='improvement',mode='max',monitor='accuracy',name='best_acc_model')])\n\u001b[0m",
      "\u001b[0;32m~/fast_ai/fastai-fork/fastai/train.py\u001b[0m in \u001b[0;36mfit_one_cycle\u001b[0;34m(learn, cyc_len, max_lr, moms, div_factor, pct_start, final_div, wd, callbacks, tot_epochs, start_epoch)\u001b[0m\n\u001b[1;32m     20\u001b[0m     callbacks.append(OneCycleScheduler(learn, max_lr, moms=moms, div_factor=div_factor, pct_start=pct_start,\n\u001b[1;32m     21\u001b[0m                                        final_div=final_div, tot_epochs=tot_epochs, start_epoch=start_epoch))\n\u001b[0;32m---> 22\u001b[0;31m     \u001b[0mlearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcyc_len\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmax_lr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwd\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mwd\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcallbacks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mcallbacks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     23\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     24\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mlr_find\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlearn\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mLearner\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstart_lr\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mFloats\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1e-7\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mend_lr\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mFloats\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m10\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum_it\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mint\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstop_div\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mbool\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwd\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mfloat\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/fast_ai/fastai-fork/fastai/basic_train.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, epochs, lr, wd, callbacks)\u001b[0m\n\u001b[1;32m    198\u001b[0m         \u001b[0mcallbacks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mcb\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mcb\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcallback_fns\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mlistify\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdefaults\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mextra_callback_fns\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mlistify\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcallbacks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    199\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mdefaults\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mextra_callbacks\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mcallbacks\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0mdefaults\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mextra_callbacks\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 200\u001b[0;31m         \u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mepochs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmetrics\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmetrics\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcallbacks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcallbacks\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mcallbacks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    201\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    202\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mcreate_opt\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlr\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mFloats\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwd\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mFloats\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0.\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m->\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/fast_ai/fastai-fork/fastai/basic_train.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(epochs, learn, callbacks, metrics)\u001b[0m\n\u001b[1;32m    110\u001b[0m         \u001b[0mexception\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    111\u001b[0m         \u001b[0;32mraise\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 112\u001b[0;31m     \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mcb_handler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mon_train_end\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mexception\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    113\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    114\u001b[0m loss_func_name2activ = {'cross_entropy_loss': F.softmax, 'nll_loss': torch.exp, 'poisson_nll_loss': torch.exp,\n",
      "\u001b[0;32m~/fast_ai/fastai-fork/fastai/callback.py\u001b[0m in \u001b[0;36mon_train_end\u001b[0;34m(self, exception)\u001b[0m\n\u001b[1;32m    321\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mon_train_end\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mexception\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mUnion\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mbool\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mException\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m->\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    322\u001b[0m         \u001b[0;34m\"Handle end of training, `exception` is an `Exception` or False if no exceptions during training.\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 323\u001b[0;31m         \u001b[0mself\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'train_end'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mexception\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mexception\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    324\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    325\u001b[0m     \u001b[0;34m@\u001b[0m\u001b[0mproperty\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/fast_ai/fastai-fork/fastai/callback.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, cb_name, call_mets, **kwargs)\u001b[0m\n\u001b[1;32m    249\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mcall_mets\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    250\u001b[0m             \u001b[0;32mfor\u001b[0m \u001b[0mmet\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmetrics\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_call_and_update\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmet\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcb_name\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 251\u001b[0;31m         \u001b[0;32mfor\u001b[0m \u001b[0mcb\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcallbacks\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_call_and_update\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcb_name\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    252\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    253\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mset_dl\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdl\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mDataLoader\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/fast_ai/fastai-fork/fastai/callback.py\u001b[0m in \u001b[0;36m_call_and_update\u001b[0;34m(self, cb, cb_name, **kwargs)\u001b[0m\n\u001b[1;32m    239\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_call_and_update\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcb_name\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m->\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    240\u001b[0m         \u001b[0;34m\"Call `cb_name` on `cb` and update the inner state.\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 241\u001b[0;31m         \u001b[0mnew\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mifnone\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgetattr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34mf'on_{cb_name}'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m**\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstate_dict\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    242\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mk\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mv\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mnew\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mitems\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    243\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mk\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstate_dict\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/fast_ai/fastai-fork/fastai/callbacks/tracker.py\u001b[0m in \u001b[0;36mon_train_end\u001b[0;34m(self, **kwargs)\u001b[0m\n\u001b[1;32m    104\u001b[0m         \u001b[0;34m\"Load the best model.\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    105\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mevery\u001b[0m\u001b[0;34m==\u001b[0m\u001b[0;34m\"improvement\"\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34mf'{self.learn.model_dir}/{self.name}.pth'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mis_file\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 106\u001b[0;31m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mload\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf'{self.name}'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpurge\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    107\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    108\u001b[0m \u001b[0;32mclass\u001b[0m \u001b[0mReduceLROnPlateauCallback\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mTrackerCallback\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/fast_ai/fastai-fork/fastai/basic_train.py\u001b[0m in \u001b[0;36mload\u001b[0;34m(self, file, device, strict, with_opt, purge, remove_module)\u001b[0m\n\u001b[1;32m    271\u001b[0m             \u001b[0mmodel_state\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mstate\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'model'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    272\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mremove_module\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mmodel_state\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mremove_module_load\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel_state\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 273\u001b[0;31m             \u001b[0mget_model\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mload_state_dict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel_state\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstrict\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mstrict\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    274\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mifnone\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mwith_opt\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    275\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mhasattr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'opt'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcreate_opt\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdefaults\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwd\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/fastaiv1_dev/lib/python3.7/site-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36mload_state_dict\u001b[0;34m(self, state_dict, strict)\u001b[0m\n\u001b[1;32m    767\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0merror_msgs\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    768\u001b[0m             raise RuntimeError('Error(s) in loading state_dict for {}:\\n\\t{}'.format(\n\u001b[0;32m--> 769\u001b[0;31m                                self.__class__.__name__, \"\\n\\t\".join(error_msgs)))\n\u001b[0m\u001b[1;32m    770\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    771\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_named_members\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mget_members_fn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mprefix\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m''\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrecurse\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mRuntimeError\u001b[0m: Error(s) in loading state_dict for SequentialRNN:\n\tMissing key(s) in state_dict: \"0.module.pos_enc.weight\", \"0.module.layers.0.mhra.attention.weight\", \"0.module.layers.0.mhra.attention.bias\", \"0.module.layers.0.mhra.out.weight\", \"0.module.layers.0.mhra.out.bias\", \"0.module.layers.0.mhra.ln.weight\", \"0.module.layers.0.mhra.ln.bias\", \"0.module.layers.0.ff.layers.0.weight\", \"0.module.layers.0.ff.layers.0.bias\", \"0.module.layers.0.ff.layers.2.weight\", \"0.module.layers.0.ff.layers.2.bias\", \"0.module.layers.0.ff.layers.5.weight\", \"0.module.layers.0.ff.layers.5.bias\", \"0.module.layers.1.mhra.attention.weight\", \"0.module.layers.1.mhra.attention.bias\", \"0.module.layers.1.mhra.out.weight\", \"0.module.layers.1.mhra.out.bias\", \"0.module.layers.1.mhra.ln.weight\", \"0.module.layers.1.mhra.ln.bias\", \"0.module.layers.1.ff.layers.0.weight\", \"0.module.layers.1.ff.layers.0.bias\", \"0.module.layers.1.ff.layers.2.weight\", \"0.module.layers.1.ff.layers.2.bias\", \"0.module.layers.1.ff.layers.5.weight\", \"0.module.layers.1.ff.layers.5.bias\", \"0.module.layers.2.mhra.attention.weight\", \"0.module.layers.2.mhra.attention.bias\", \"0.module.layers.2.mhra.out.weight\", \"0.module.layers.2.mhra.out.bias\", \"0.module.layers.2.mhra.ln.weight\", \"0.module.layers.2.mhra.ln.bias\", \"0.module.layers.2.ff.layers.0.weight\", \"0.module.layers.2.ff.layers.0.bias\", \"0.module.layers.2.ff.layers.2.weight\", \"0.module.layers.2.ff.layers.2.bias\", \"0.module.layers.2.ff.layers.5.weight\", \"0.module.layers.2.ff.layers.5.bias\", \"0.module.layers.3.mhra.attention.weight\", \"0.module.layers.3.mhra.attention.bias\", \"0.module.layers.3.mhra.out.weight\", \"0.module.layers.3.mhra.out.bias\", \"0.module.layers.3.mhra.ln.weight\", \"0.module.layers.3.mhra.ln.bias\", \"0.module.layers.3.ff.layers.0.weight\", \"0.module.layers.3.ff.layers.0.bias\", \"0.module.layers.3.ff.layers.2.weight\", \"0.module.layers.3.ff.layers.2.bias\", \"0.module.layers.3.ff.layers.5.weight\", \"0.module.layers.3.ff.layers.5.bias\", \"0.module.layers.4.mhra.attention.weight\", \"0.module.layers.4.mhra.attention.bias\", \"0.module.layers.4.mhra.out.weight\", \"0.module.layers.4.mhra.out.bias\", \"0.module.layers.4.mhra.ln.weight\", \"0.module.layers.4.mhra.ln.bias\", \"0.module.layers.4.ff.layers.0.weight\", \"0.module.layers.4.ff.layers.0.bias\", \"0.module.layers.4.ff.layers.2.weight\", \"0.module.layers.4.ff.layers.2.bias\", \"0.module.layers.4.ff.layers.5.weight\", \"0.module.layers.4.ff.layers.5.bias\", \"0.module.layers.5.mhra.attention.weight\", \"0.module.layers.5.mhra.attention.bias\", \"0.module.layers.5.mhra.out.weight\", \"0.module.layers.5.mhra.out.bias\", \"0.module.layers.5.mhra.ln.weight\", \"0.module.layers.5.mhra.ln.bias\", \"0.module.layers.5.ff.layers.0.weight\", \"0.module.layers.5.ff.layers.0.bias\", \"0.module.layers.5.ff.layers.2.weight\", \"0.module.layers.5.ff.layers.2.bias\", \"0.module.layers.5.ff.layers.5.weight\", \"0.module.layers.5.ff.layers.5.bias\", \"0.module.layers.6.mhra.attention.weight\", \"0.module.layers.6.mhra.attention.bias\", \"0.module.layers.6.mhra.out.weight\", \"0.module.layers.6.mhra.out.bias\", \"0.module.layers.6.mhra.ln.weight\", \"0.module.layers.6.mhra.ln.bias\", \"0.module.layers.6.ff.layers.0.weight\", \"0.module.layers.6.ff.layers.0.bias\", \"0.module.layers.6.ff.layers.2.weight\", \"0.module.layers.6.ff.layers.2.bias\", \"0.module.layers.6.ff.layers.5.weight\", \"0.module.layers.6.ff.layers.5.bias\", \"0.module.layers.7.mhra.attention.weight\", \"0.module.layers.7.mhra.attention.bias\", \"0.module.layers.7.mhra.out.weight\", \"0.module.layers.7.mhra.out.bias\", \"0.module.layers.7.mhra.ln.weight\", \"0.module.layers.7.mhra.ln.bias\", \"0.module.layers.7.ff.layers.0.weight\", \"0.module.layers.7.ff.layers.0.bias\", \"0.module.layers.7.ff.layers.2.weight\", \"0.module.layers.7.ff.layers.2.bias\", \"0.module.layers.7.ff.layers.5.weight\", \"0.module.layers.7.ff.layers.5.bias\", \"1.layers.1.weight\", \"1.layers.1.bias\", \"1.layers.3.weight\", \"1.layers.3.bias\", \"1.layers.3.running_mean\", \"1.layers.3.running_var\", \"1.layers.5.weight\", \"1.layers.5.bias\". \n\tUnexpected key(s) in state_dict: \"0.module.encoder_dp.emb.weight\", \"0.module.rnns.0.weight_hh_l0_raw\", \"0.module.rnns.0.module.weight_ih_l0\", \"0.module.rnns.0.module.weight_hh_l0\", \"0.module.rnns.0.module.bias_ih_l0\", \"0.module.rnns.0.module.bias_hh_l0\", \"0.module.rnns.1.weight_hh_l0_raw\", \"0.module.rnns.1.module.weight_ih_l0\", \"0.module.rnns.1.module.weight_hh_l0\", \"0.module.rnns.1.module.bias_ih_l0\", \"0.module.rnns.1.module.bias_hh_l0\", \"0.module.rnns.2.weight_hh_l0_raw\", \"0.module.rnns.2.module.weight_ih_l0\", \"0.module.rnns.2.module.weight_hh_l0\", \"0.module.rnns.2.module.bias_ih_l0\", \"0.module.rnns.2.module.bias_hh_l0\", \"1.mhra.attention.weight\", \"1.mhra.attention.bias\", \"1.layers.2.weight\", \"1.layers.2.bias\". \n\tsize mismatch for 0.module.encoder.weight: copying a param with shape torch.Size([60000, 400]) from checkpoint, the shape in current model is torch.Size([60000, 512]).\n\tsize mismatch for 1.layers.0.weight: copying a param with shape torch.Size([100]) from checkpoint, the shape in current model is torch.Size([1536]).\n\tsize mismatch for 1.layers.0.bias: copying a param with shape torch.Size([100]) from checkpoint, the shape in current model is torch.Size([1536]).\n\tsize mismatch for 1.layers.0.running_mean: copying a param with shape torch.Size([100]) from checkpoint, the shape in current model is torch.Size([1536]).\n\tsize mismatch for 1.layers.0.running_var: copying a param with shape torch.Size([100]) from checkpoint, the shape in current model is torch.Size([1536])."
     ]
    }
   ],
   "source": [
    "learn.freeze_to(-2)\n",
    "learn.fit_one_cycle(20,slice(1e-8,1e-3),moms=(0.7,0.4), wd=0.10,\n",
    "                   callbacks=[SaveModelCallback(learn,every='improvement',mode='max',monitor='accuracy',name='best_acc_model')])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.recorder.plot_losses()\n",
    "learn.recorder.plot_metrics()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "Total time: 05:00 <p><table style='width:375px; margin-bottom:10px'>\n",
       "  <tr>\n",
       "    <th>epoch</th>\n",
       "    <th>train_loss</th>\n",
       "    <th>valid_loss</th>\n",
       "    <th>accuracy</th>\n",
       "    <th>time</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>0</th>\n",
       "    <th>0.031394</th>\n",
       "    <th>0.833714</th>\n",
       "    <th>0.851389</th>\n",
       "    <th>00:26</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>1</th>\n",
       "    <th>0.032294</th>\n",
       "    <th>0.775586</th>\n",
       "    <th>0.849722</th>\n",
       "    <th>00:28</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>2</th>\n",
       "    <th>0.034081</th>\n",
       "    <th>1.447592</th>\n",
       "    <th>0.830000</th>\n",
       "    <th>00:29</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>3</th>\n",
       "    <th>0.039692</th>\n",
       "    <th>0.953342</th>\n",
       "    <th>0.835278</th>\n",
       "    <th>00:30</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>4</th>\n",
       "    <th>0.027997</th>\n",
       "    <th>1.081962</th>\n",
       "    <th>0.839167</th>\n",
       "    <th>00:28</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>5</th>\n",
       "    <th>0.020747</th>\n",
       "    <th>0.815173</th>\n",
       "    <th>0.842778</th>\n",
       "    <th>00:30</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>6</th>\n",
       "    <th>0.012641</th>\n",
       "    <th>0.806318</th>\n",
       "    <th>0.849444</th>\n",
       "    <th>00:33</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>7</th>\n",
       "    <th>0.010216</th>\n",
       "    <th>0.968975</th>\n",
       "    <th>0.851944</th>\n",
       "    <th>00:32</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>8</th>\n",
       "    <th>0.008404</th>\n",
       "    <th>0.991545</th>\n",
       "    <th>0.847222</th>\n",
       "    <th>00:31</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>9</th>\n",
       "    <th>0.007599</th>\n",
       "    <th>0.965352</th>\n",
       "    <th>0.850833</th>\n",
       "    <th>00:30</th>\n",
       "  </tr>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.freeze_to(-3)\n",
    "learn.fit_one_cycle(10,slice(1e-8,1e-3),moms=(0.7,0.4), wd=0.10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.recorder.plot_losses()\n",
    "learn.recorder.plot_metrics()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "Total time: 05:48 <p><table style='width:375px; margin-bottom:10px'>\n",
       "  <tr>\n",
       "    <th>epoch</th>\n",
       "    <th>train_loss</th>\n",
       "    <th>valid_loss</th>\n",
       "    <th>accuracy</th>\n",
       "    <th>time</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>0</th>\n",
       "    <th>0.005337</th>\n",
       "    <th>0.933734</th>\n",
       "    <th>0.849722</th>\n",
       "    <th>00:34</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>1</th>\n",
       "    <th>0.004274</th>\n",
       "    <th>1.104189</th>\n",
       "    <th>0.851111</th>\n",
       "    <th>00:35</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>2</th>\n",
       "    <th>0.009081</th>\n",
       "    <th>0.898218</th>\n",
       "    <th>0.835556</th>\n",
       "    <th>00:35</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>3</th>\n",
       "    <th>0.007863</th>\n",
       "    <th>1.469672</th>\n",
       "    <th>0.852500</th>\n",
       "    <th>00:34</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>4</th>\n",
       "    <th>0.021667</th>\n",
       "    <th>1.500991</th>\n",
       "    <th>0.842222</th>\n",
       "    <th>00:34</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>5</th>\n",
       "    <th>0.019359</th>\n",
       "    <th>1.146566</th>\n",
       "    <th>0.847222</th>\n",
       "    <th>00:34</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>6</th>\n",
       "    <th>0.030465</th>\n",
       "    <th>0.876468</th>\n",
       "    <th>0.841111</th>\n",
       "    <th>00:33</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>7</th>\n",
       "    <th>0.013433</th>\n",
       "    <th>1.276104</th>\n",
       "    <th>0.838056</th>\n",
       "    <th>00:35</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>8</th>\n",
       "    <th>0.009762</th>\n",
       "    <th>1.014557</th>\n",
       "    <th>0.843611</th>\n",
       "    <th>00:33</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>9</th>\n",
       "    <th>0.007876</th>\n",
       "    <th>1.097269</th>\n",
       "    <th>0.844722</th>\n",
       "    <th>00:36</th>\n",
       "  </tr>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.unfreeze()\n",
    "learn.fit_one_cycle(10,slice(1e-8,1e-3),moms=(0.7,0.4), pct_start=0.75)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.recorder.plot_losses()\n",
    "learn.recorder.plot_metrics()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "preds_t,y_t,losses_t = learn.get_preds(DatasetType.Train,with_loss=True)\n",
    "interp_t = ClassificationInterpretation(learn, preds_t, y_t, losses_t)\n",
    "interp_t.plot_confusion_matrix()\n",
    "\n",
    "preds,y,losses = learn.get_preds(with_loss=True)\n",
    "interp = ClassificationInterpretation(learn, preds, y, losses)\n",
    "interp.plot_confusion_matrix()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "##next try label smoothing?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------\n"
     ]
    },
    {
     "ename": "TypeError",
     "evalue": "forward() got an unexpected keyword argument 'from_embeddings'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-40-fe7bce1d6eff>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0midx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mto_np\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0midxs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"-\"\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0;36m10\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 7\u001b[0;31m     \u001b[0mtxt_ci\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshow_intrinsic_attention\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdf_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0miloc\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0midx\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mcmap\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mcm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mBlues\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m#textify breaks?\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m~/fast_ai/fastai-fork/fastai/text/models/awd_lstm.py\u001b[0m in \u001b[0;36mshow_intrinsic_attention\u001b[0;34m(self, text, class_id, **kwargs)\u001b[0m\n\u001b[1;32m    241\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    242\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mshow_intrinsic_attention\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtext\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclass_id\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mint\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m->\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 243\u001b[0;31m         \u001b[0mtext\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mattn\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mintrinsic_attention\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtext\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclass_id\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    244\u001b[0m         \u001b[0mshow_piece_attn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtext\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtext\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msplit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mto_np\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mattn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    245\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/fast_ai/fastai-fork/fastai/text/models/awd_lstm.py\u001b[0m in \u001b[0;36mintrinsic_attention\u001b[0;34m(self, text, class_id)\u001b[0m\n\u001b[1;32m    226\u001b[0m         \u001b[0mids\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mone_item\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtext\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    227\u001b[0m         \u001b[0memb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmodule\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mencoder\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mids\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdetach\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrequires_grad_\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 228\u001b[0;31m         \u001b[0mlstm_output\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmodule\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0memb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfrom_embeddings\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    229\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0meval\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    230\u001b[0m         \u001b[0mcl\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlstm_output\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mzeros_like\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mids\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbyte\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msoftmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdim\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/fastaiv1_dev/lib/python3.7/site-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    487\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    488\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 489\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    490\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mhook\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    491\u001b[0m             \u001b[0mhook_result\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mhook\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mresult\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mTypeError\u001b[0m: forward() got an unexpected keyword argument 'from_embeddings'"
     ]
    }
   ],
   "source": [
    "from fastai.text.models.awd_lstm import TextClassificationInterpretation\n",
    "from matplotlib.pyplot import cm\n",
    "txt_ci = TextClassificationInterpretation.from_learner(learn)\n",
    "(_,idxs) = interp.top_losses(5)\n",
    "for idx in to_np(idxs):\n",
    "    print(\"-\"*10)\n",
    "    txt_ci.show_intrinsic_attention(df_train.iloc[idx,1],cmap=cm.Blues) #textify breaks?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn.save(\"haha_20190516_tweet_vocab_qrnn\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "### Save predictions for use in Ensemble method (not used)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "hidden": true
   },
   "outputs": [],
   "source": [
    "pd.DataFrame(to_np(preds)).to_csv('rnn_sent_preds.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "hidden": true
   },
   "outputs": [],
   "source": [
    "## fastai \"test\" has no labels.  It is a true test, not a whitholding set!!  We need a second valid set?\n",
    "## move the df_test into the \"valid\" position so we can run inference and get preds on randomized vals\n",
    "data_t = TextClasDataBunch.from_df(PATH,df_train,df_test,\n",
    "                               tokenizer=mycust_tok, vocab=sp_vocab,\n",
    "                               text_cols='new_text', label_cols='is_humor')\n",
    "learn.data = data_t\n",
    "t_preds,t_y,t_losses = learn.get_preds(with_loss=True,ds_type=DatasetType.Valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "hidden": true
   },
   "outputs": [],
   "source": [
    "pd.DataFrame(to_np(t_preds)).to_csv('rnn_sent_t_preds.csv')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.7 fasta.ai1 DEV",
   "language": "python",
   "name": "fastai1_dev"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}