02_train_LM_es.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from decoder_head.core import *\n",
    "from fastai2.text.all import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "es.wiki.test.tokens  es.wiki.train.tokens  es.wiki.valid.tokens\r\n"
     ]
    }
   ],
   "source": [
    "ls data/es-100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "es-100/\n",
      "es-100/es.wiki.test.tokens\n",
      "es-100/es.wiki.train.tokens\n",
      "es-100/es.wiki.valid.tokens\n"
     ]
    }
   ],
   "source": [
    "# !rm -rf data/es-100\n",
    "# !tar -xvf data/es-100.tar.gz -C data\n",
    "\n",
    "# !mkdir data/es-100/train\n",
    "# !mkdir data/es-100/valid\n",
    "# !mkdir data/es-100/test\n",
    "\n",
    "# !mv data/es-100/es.wiki.train.tokens data/es-100/train/es.wiki.train.tokens\n",
    "# !mv data/es-100/es.wiki.valid.tokens data/es-100/valid/es.wiki.valid.tokens\n",
    "# !mv data/es-100/es.wiki.test.tokens data/es-100/test/es.wiki.test.tokens"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "# split_file_into_chunks('data/es-100/train/es.wiki.train.tokens')\n",
    "# split_file_into_chunks('data/es-100/valid/es.wiki.valid.tokens')\n",
    "# split_file_into_chunks('data/es-100/test/es.wiki.test.tokens')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.txt   13.txt  17.txt  20.txt  2.txt  6.txt  es.wiki.train.tokens\r\n",
      "10.txt  14.txt  18.txt  21.txt  3.txt  7.txt\r\n",
      "11.txt  15.txt  19.txt  22.txt  4.txt  8.txt\r\n",
      "12.txt  16.txt  1.txt   23.txt  5.txt  9.txt\r\n"
     ]
    }
   ],
   "source": [
    "ls data/es-100/train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# tokenize_folder('data/es-100')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "vocab = make_vocab(pd.read_pickle('data/es-100_tok/counter.pkl'), max_vocab=4000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = 'data/es-100_tok/'\n",
    "mult = 4\n",
    "bs = 80\n",
    "seq_len = 70\n",
    "\n",
    "lm = DataBlock(blocks=(TextBlock(vocab=vocab, is_lm=True),),\n",
    "                get_x=read_tokenized_file,\n",
    "                get_items=partial(get_text_files, folders=['train', 'valid']),\n",
    "                splitter=FuncSplitter(lambda itm: itm.parent.name == 'valid'))\n",
    "\n",
    "dbunch_lm = lm.databunch(path, path=path, bs=bs, seq_len=seq_len)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn = language_model_learner(\n",
    "    dbunch_lm,\n",
    "    pAWD_LSTM,\n",
    "    opt_func=opt,\n",
    "    pretrained=False,\n",
    "    config=awd_lstm_lm_config,\n",
    "    drop_mult=0.1,\n",
    "    metrics=[accuracy]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(0.1720)"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.mean((torch.cat([tup[0] for tup in dbunch_lm.valid_ds]) == 0).float())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\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",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>2.486690</td>\n",
       "      <td>2.816764</td>\n",
       "      <td>0.392779</td>\n",
       "      <td>46:39</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "IOPub message rate exceeded.\n",
      "The notebook server will temporarily stop sending output\n",
      "to the client in order to avoid crashing it.\n",
      "To change this limit, set the config variable\n",
      "`--NotebookApp.iopub_msg_rate_limit`.\n",
      "\n",
      "Current values:\n",
      "NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)\n",
      "NotebookApp.rate_limit_window=3.0 (secs)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "learn.fit_one_cycle(1, 5e-3,  moms=(0.8, 0.7, 0.8), wd=1e-7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "# learn.save('pLSTM_es')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<fastai2.text.learner.LMLearner at 0x7f55c0677c10>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "learn.load('pLSTM_es')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save(learn.model[0].encoder.weight, 'data/embeddings_es.torch')\n",
    "torch.save(learn.dbunch.vocab, 'data/vocab_es.torch')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'La Red y el Cementerio Santo Domingo , fueron los primeros en participar , que en septiembre'"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "learn.predict('La', n_words=20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn.optimize_permutation()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "preds, targs = learn.get_preds()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(2.8087, device='cuda:0', grad_fn=<AddBackward0>)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "aza_loss(learn, preds, targs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Parameter containing:\n",
       "tensor([[ 0.0202,  0.0092,  0.0202,  ...,  0.0065, -0.0105, -0.0176],\n",
       "        [ 0.0157,  0.0335,  0.0423,  ...,  0.0039,  0.0298, -0.0334],\n",
       "        [ 0.0148,  0.0290, -0.0304,  ...,  0.0317, -0.0044,  0.0185],\n",
       "        ...,\n",
       "        [-0.0219, -0.0451, -0.0346,  ..., -0.0046, -0.0389,  0.0021],\n",
       "        [-0.0031,  0.0020,  0.0253,  ..., -0.0223,  0.0424, -0.0107],\n",
       "        [-0.0175, -0.0138, -0.0025,  ..., -0.0412,  0.0085, -0.0363]],\n",
       "       device='cuda:0', requires_grad=True)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nn.init.kaiming_normal_(learn.model[0].encoder.p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "(tensor(564654.6875, device='cuda:0', grad_fn=<AddBackward0>), tensor(6.0058))"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "preds, targs = learn.get_preds()\n",
    "aza_loss(learn, preds, targs, debug=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn.loss_func = lambda preds, targs: aza_loss(learn, preds, targs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\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",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>2.613976</td>\n",
       "      <td>2.948374</td>\n",
       "      <td>0.377296</td>\n",
       "      <td>39:42</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.fit_one_cycle(1, 1e-3,  moms=(0.8, 0.7, 0.8), wd=1e-7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# learn.save('pLSTM_es_permute_embeddings')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<fastai2.text.learner.LMLearner at 0x7f55c0677c10>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "learn.load('pLSTM_es_permute_embeddings')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.9899999499320984, 1.0)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "learn.hitInTopN(1, 100), learn.hitInTopN(5, 100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "acc1 = [(x, learn.hitInTopN(1, x)) for x in range(50,1500,200)]\n",
    "acc2 = [(x, learn.hitInTopN(2, x)) for x in range(50,1500,200)]\n",
    "acc3 = [(x, learn.hitInTopN(3, x)) for x in range(50,1500,200)]\n",
    "acc5 = [(x, learn.hitInTopN(5, x)) for x in range(50,1500,200)]\n",
    "acc10 = [(x, learn.hitInTopN(10, x)) for x in range(50,1500,200)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fb8fc4fad50>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,10))\n",
    "plt.plot(*zip(*acc1))\n",
    "plt.plot(*zip(*acc2))\n",
    "plt.plot(*zip(*acc3))\n",
    "plt.plot(*zip(*acc5))\n",
    "plt.plot(*zip(*acc10))\n",
    "plt.legend([f'acc @ {x}' for x in [1,2,3,5,10]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "actual word\t\tpredictions\n",
      "\n",
      "años\t\t['años', 'siglos', 'meses', 'décadas', 'temporadas']\n",
      "son\t\t['son', 'están', 'incluyen', 'serán', 'eran']\n",
      "año\t\t['año', 'minuto', 'temporada', 'mes', 'siglos']\n",
      "desde\t\t['desde', 'véase', 'hacia', 'entre', 'durante']\n",
      "pero\t\t['pero', 'aunque', 'pues', 'etc', 'mientras']\n",
      "según\t\t['según', 'siguiendo', 'pese', 'véase', 'dependiendo']\n",
      "durante\t\t['durante', 'tras', 'siguiendo', 'falleció', 'desde']\n",
      "esta\t\t['esta', 'dicha', 'estas', 'esa', 'este']\n",
      "hasta\t\t['hasta', 'hacia', 'desde', 'ua', 'llegando']\n",
      "parte\t\t['parte', 'depende', 'lado', 'mitad', 'consecuencia']\n",
      "ha\t\t['ha', 'han', 'he', 'ee.uu', 'nos']\n",
      "sobre\t\t['sobre', 'ante', 'hacia', 'siguiendo', 'acerca']\n",
      "censo\t\t['censo', 'contaba', 'township', 'plazo', 'ayuntamiento']\n",
      "había\t\t['había', 'habían', 'hubiera', 'habría', 'puso']\n",
      "personas\t\t['personas', 'pacientes', 'individuos', 'familias', 'mujeres']\n",
      "«\t\t['«', '‘', '¿', '“', 'frase']\n",
      "»\t\t['»', '”', '’', '¡', '!']\n",
      "donde\t\t['donde', 'quienes', 'cuyo', 'cuya', 'cuyos']\n",
      "le\t\t['le', 'les', 'nos', 'murió', 'decía']\n",
      "tiene\t\t['tiene', 'tienen', 'tendrá', 'lleva', 'tenían']\n"
     ]
    }
   ],
   "source": [
    "learn.print_synonyms(60, 20)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}