AWD_LSTM_train.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "from ipyexperiments import *\n",
    "from fastai import *\n",
    "from fastai.text import *\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from path import Path"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "MODELS = Path('./models/AWD_LSTM')\n",
    "DATA = Path('./data')\n",
    "\n",
    "df = pd.read_parquet(DATA / 'mbti_1.pqt', 'fastparquet')\n",
    "train = df[df.is_valid == False]\n",
    "val = df[df.is_valid == True]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "*** Experiment started with the Pytorch backend\n",
      "Device: ID 0, GeForce GTX 1070 (8117 RAM)\n",
      "\n",
      "\n",
      "*** Current state:\n",
      "RAM:    Used    Free   Total       Util\n",
      "CPU:   2,778  48,954  64,352 MB   4.32% \n",
      "GPU:   1,600   6,517   8,117 MB  19.71% \n",
      "\n",
      "\n",
      "･ RAM:  △Consumed    △Peaked    Used Total | Exec time 0:00:00.180\n",
      "･ CPU:          0          0      2,778 MB |\n",
      "･ GPU:          4          2      1,600 MB |\n",
      "\n",
      "IPyExperimentsPytorch: Finishing\n",
      "\n",
      "*** Experiment finished in 02:08:45 (elapsed wallclock time)\n",
      "\n",
      "*** Newly defined local variables:\n",
      "Deleted: data_lm\n",
      "\n",
      "*** Experiment memory:\n",
      "RAM: Consumed       Reclaimed\n",
      "CPU:      329        4 MB (  1.35%)\n",
      "GPU:        4        0 MB (  0.00%)\n",
      "\n",
      "*** Current state:\n",
      "RAM:    Used    Free   Total       Util\n",
      "CPU:   2,774  48,958  64,352 MB   4.31% \n",
      "GPU:   1,600   6,517   8,117 MB  19.71% \n",
      "\n",
      "\n",
      "･ RAM:  △Consumed    △Peaked    Used Total | Exec time 0:00:03.794\n",
      "･ CPU:          0          0      2,859 MB |\n",
      "･ GPU:        234          8      1,834 MB |\n"
     ]
    }
   ],
   "source": [
    "exp1 = IPyExperimentsPytorch()\n",
    "\n",
    "data_lm = load_data(MODELS,'data_lm_export.pkl')\n",
    "learn = language_model_learner(data_lm, arch=AWD_LSTM )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "･ RAM:  △Consumed    △Peaked    Used Total | Exec time 0:00:33.478\n",
      "･ CPU:          0          2      2,888 MB |\n",
      "･ GPU:        406      5,249      2,240 MB |\n"
     ]
    }
   ],
   "source": [
    "learn.lr_find()\n",
    "learn.recorder.plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "Total time: 15:23 <p><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>4.700811</td>\n",
       "      <td>4.366804</td>\n",
       "      <td>0.242623</td>\n",
       "      <td>15:23</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Total time: 2:55:42 <p><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>4.469680</td>\n",
       "      <td>4.251969</td>\n",
       "      <td>0.253254</td>\n",
       "      <td>17:53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>4.392018</td>\n",
       "      <td>4.133133</td>\n",
       "      <td>0.265819</td>\n",
       "      <td>17:40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>4.264731</td>\n",
       "      <td>4.047349</td>\n",
       "      <td>0.274892</td>\n",
       "      <td>17:39</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>4.171680</td>\n",
       "      <td>3.991757</td>\n",
       "      <td>0.281432</td>\n",
       "      <td>17:28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>4.139320</td>\n",
       "      <td>3.955577</td>\n",
       "      <td>0.285707</td>\n",
       "      <td>17:31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>4.091193</td>\n",
       "      <td>3.927975</td>\n",
       "      <td>0.289342</td>\n",
       "      <td>17:30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>4.067306</td>\n",
       "      <td>3.910633</td>\n",
       "      <td>0.291621</td>\n",
       "      <td>17:25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>4.012954</td>\n",
       "      <td>3.898835</td>\n",
       "      <td>0.292884</td>\n",
       "      <td>17:27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>3.989483</td>\n",
       "      <td>3.893325</td>\n",
       "      <td>0.293600</td>\n",
       "      <td>17:28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>3.979686</td>\n",
       "      <td>3.892325</td>\n",
       "      <td>0.293740</td>\n",
       "      <td>17:37</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "･ RAM:  △Consumed    △Peaked    Used Total | Exec time 3:11:05.809\n",
      "･ CPU:          3          0      2,892 MB |\n",
      "･ GPU:      1,425      4,229      3,666 MB |\n"
     ]
    }
   ],
   "source": [
    "learn.fit_one_cycle(1, 2e-02)\n",
    "learn.unfreeze()\n",
    "learn.fit_one_cycle(10)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "･ RAM:  △Consumed    △Peaked    Used Total | Exec time 0:00:12.600\n",
      "･ CPU:          0          0      2,904 MB |\n",
      "･ GPU:          0          0      4,111 MB |\n"
     ]
    }
   ],
   "source": [
    "learn.save('awd1')\n",
    "learn.save_encoder('awd1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}