playground.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from Decoder import Decoder\n",
    "from Encoder import Encoder\n",
    "from evaluate_captions import evaluate_captions\n",
    "from runner import Runner\n",
    "from settings import LSTM_HIDDEN_SIZE, EMBEDDED_SIZE, CAPTIONS_DIR, IMAGES_DIR, NUM_WORKERS, SHUFFLE_DATA, BATCH_SIZE, VALIDATION_SIZE, TEMPERATURE\n",
    "from utils import load_datasets, select_ann_ids, load_image_ids, get_device\n",
    "\n",
    "from data_loader import get_loader"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loading annotations into memory...\n",
      "Done (t=0.87s)\n",
      "creating index...\n",
      "index created!\n",
      "loading annotations into memory...\n",
      "Done (t=0.51s)\n",
      "creating index...\n",
      "index created!\n",
      "loading annotations into memory...\n",
      "Done (t=0.79s)\n",
      "creating index...\n",
      "index created!\n",
      "loading annotations into memory...\n",
      "Done (t=0.79s)\n",
      "creating index...\n",
      "index created!\n",
      "[1000/414113] Tokenized the captions.\n",
      "[2000/414113] Tokenized the captions.\n",
      "[3000/414113] Tokenized the captions.\n",
      "[4000/414113] Tokenized the captions.\n",
      "[5000/414113] Tokenized the captions.\n",
      "[6000/414113] Tokenized the captions.\n",
      "[7000/414113] Tokenized the captions.\n",
      "[8000/414113] Tokenized the captions.\n",
      "[9000/414113] Tokenized the captions.\n",
      "[10000/414113] Tokenized the captions.\n",
      "[11000/414113] Tokenized the captions.\n",
      "[12000/414113] Tokenized the captions.\n",
      "[13000/414113] Tokenized the captions.\n",
      "[14000/414113] Tokenized the captions.\n",
      "[15000/414113] Tokenized the captions.\n",
      "[16000/414113] Tokenized the captions.\n",
      "[17000/414113] Tokenized the captions.\n",
      "[18000/414113] Tokenized the captions.\n",
      "[19000/414113] Tokenized the captions.\n",
      "[20000/414113] Tokenized the captions.\n",
      "[21000/414113] Tokenized the captions.\n",
      "[22000/414113] Tokenized the captions.\n",
      "[23000/414113] Tokenized the captions.\n",
      "[24000/414113] Tokenized the captions.\n",
      "[25000/414113] Tokenized the captions.\n",
      "[26000/414113] Tokenized the captions.\n",
      "[27000/414113] Tokenized the captions.\n",
      "[28000/414113] Tokenized the captions.\n",
      "[29000/414113] Tokenized the captions.\n",
      "[30000/414113] Tokenized the captions.\n",
      "[31000/414113] Tokenized the captions.\n",
      "[32000/414113] Tokenized the captions.\n",
      "[33000/414113] Tokenized the captions.\n",
      "[34000/414113] Tokenized the captions.\n",
      "[35000/414113] Tokenized the captions.\n",
      "[36000/414113] Tokenized the captions.\n",
      "[37000/414113] Tokenized the captions.\n",
      "[38000/414113] Tokenized the captions.\n",
      "[39000/414113] Tokenized the captions.\n",
      "[40000/414113] Tokenized the captions.\n",
      "[41000/414113] Tokenized the captions.\n",
      "[42000/414113] Tokenized the captions.\n",
      "[43000/414113] Tokenized the captions.\n",
      "[44000/414113] Tokenized the captions.\n",
      "[45000/414113] Tokenized the captions.\n",
      "[46000/414113] Tokenized the captions.\n",
      "[47000/414113] Tokenized the captions.\n",
      "[48000/414113] Tokenized the captions.\n",
      "[49000/414113] Tokenized the captions.\n",
      "[50000/414113] Tokenized the captions.\n",
      "[51000/414113] Tokenized the captions.\n",
      "[52000/414113] Tokenized the captions.\n",
      "[53000/414113] Tokenized the captions.\n",
      "[54000/414113] Tokenized the captions.\n",
      "[55000/414113] Tokenized the captions.\n",
      "[56000/414113] Tokenized the captions.\n",
      "[57000/414113] Tokenized the captions.\n",
      "[58000/414113] Tokenized the captions.\n",
      "[59000/414113] Tokenized the captions.\n",
      "[60000/414113] Tokenized the captions.\n",
      "[61000/414113] Tokenized the captions.\n",
      "[62000/414113] Tokenized the captions.\n",
      "[63000/414113] Tokenized the captions.\n",
      "[64000/414113] Tokenized the captions.\n",
      "[65000/414113] Tokenized the captions.\n",
      "[66000/414113] Tokenized the captions.\n",
      "[67000/414113] Tokenized the captions.\n",
      "[68000/414113] Tokenized the captions.\n",
      "[69000/414113] Tokenized the captions.\n",
      "[70000/414113] Tokenized the captions.\n",
      "[71000/414113] Tokenized the captions.\n",
      "[72000/414113] Tokenized the captions.\n",
      "[73000/414113] Tokenized the captions.\n",
      "[74000/414113] Tokenized the captions.\n",
      "[75000/414113] Tokenized the captions.\n",
      "[76000/414113] Tokenized the captions.\n",
      "[77000/414113] Tokenized the captions.\n",
      "[78000/414113] Tokenized the captions.\n",
      "[79000/414113] Tokenized the captions.\n",
      "[80000/414113] Tokenized the captions.\n",
      "[81000/414113] Tokenized the captions.\n",
      "[82000/414113] Tokenized the captions.\n",
      "[83000/414113] Tokenized the captions.\n",
      "[84000/414113] Tokenized the captions.\n",
      "[85000/414113] Tokenized the captions.\n",
      "[86000/414113] Tokenized the captions.\n",
      "[87000/414113] Tokenized the captions.\n",
      "[88000/414113] Tokenized the captions.\n",
      "[89000/414113] Tokenized the captions.\n",
      "[90000/414113] Tokenized the captions.\n",
      "[91000/414113] Tokenized the captions.\n",
      "[92000/414113] Tokenized the captions.\n",
      "[93000/414113] Tokenized the captions.\n",
      "[94000/414113] Tokenized the captions.\n",
      "[95000/414113] Tokenized the captions.\n",
      "[96000/414113] Tokenized the captions.\n",
      "[97000/414113] Tokenized the captions.\n",
      "[98000/414113] Tokenized the captions.\n",
      "[99000/414113] Tokenized the captions.\n",
      "[100000/414113] Tokenized the captions.\n",
      "[101000/414113] Tokenized the captions.\n",
      "[102000/414113] Tokenized the captions.\n",
      "[103000/414113] Tokenized the captions.\n",
      "[104000/414113] Tokenized the captions.\n",
      "[105000/414113] Tokenized the captions.\n",
      "[106000/414113] Tokenized the captions.\n",
      "[107000/414113] Tokenized the captions.\n",
      "[108000/414113] Tokenized the captions.\n",
      "[109000/414113] Tokenized the captions.\n",
      "[110000/414113] Tokenized the captions.\n",
      "[111000/414113] Tokenized the captions.\n",
      "[112000/414113] Tokenized the captions.\n",
      "[113000/414113] Tokenized the captions.\n",
      "[114000/414113] Tokenized the captions.\n",
      "[115000/414113] Tokenized the captions.\n",
      "[116000/414113] Tokenized the captions.\n",
      "[117000/414113] Tokenized the captions.\n",
      "[118000/414113] Tokenized the captions.\n",
      "[119000/414113] Tokenized the captions.\n",
      "[120000/414113] Tokenized the captions.\n",
      "[121000/414113] Tokenized the captions.\n",
      "[122000/414113] Tokenized the captions.\n",
      "[123000/414113] Tokenized the captions.\n",
      "[124000/414113] Tokenized the captions.\n",
      "[125000/414113] Tokenized the captions.\n",
      "[126000/414113] Tokenized the captions.\n",
      "[127000/414113] Tokenized the captions.\n",
      "[128000/414113] Tokenized the captions.\n",
      "[129000/414113] Tokenized the captions.\n",
      "[130000/414113] Tokenized the captions.\n",
      "[131000/414113] Tokenized the captions.\n",
      "[132000/414113] Tokenized the captions.\n",
      "[133000/414113] Tokenized the captions.\n",
      "[134000/414113] Tokenized the captions.\n",
      "[135000/414113] Tokenized the captions.\n",
      "[136000/414113] Tokenized the captions.\n",
      "[137000/414113] Tokenized the captions.\n",
      "[138000/414113] Tokenized the captions.\n",
      "[139000/414113] Tokenized the captions.\n",
      "[140000/414113] Tokenized the captions.\n",
      "[141000/414113] Tokenized the captions.\n",
      "[142000/414113] Tokenized the captions.\n",
      "[143000/414113] Tokenized the captions.\n",
      "[144000/414113] Tokenized the captions.\n",
      "[145000/414113] Tokenized the captions.\n",
      "[146000/414113] Tokenized the captions.\n",
      "[147000/414113] Tokenized the captions.\n",
      "[148000/414113] Tokenized the captions.\n",
      "[149000/414113] Tokenized the captions.\n",
      "[150000/414113] Tokenized the captions.\n",
      "[151000/414113] Tokenized the captions.\n",
      "[152000/414113] Tokenized the captions.\n",
      "[153000/414113] Tokenized the captions.\n",
      "[154000/414113] Tokenized the captions.\n",
      "[155000/414113] Tokenized the captions.\n",
      "[156000/414113] Tokenized the captions.\n",
      "[157000/414113] Tokenized the captions.\n",
      "[158000/414113] Tokenized the captions.\n",
      "[159000/414113] Tokenized the captions.\n",
      "[160000/414113] Tokenized the captions.\n",
      "[161000/414113] Tokenized the captions.\n",
      "[162000/414113] Tokenized the captions.\n",
      "[163000/414113] Tokenized the captions.\n",
      "[164000/414113] Tokenized the captions.\n",
      "[165000/414113] Tokenized the captions.\n",
      "[166000/414113] Tokenized the captions.\n",
      "[167000/414113] Tokenized the captions.\n",
      "[168000/414113] Tokenized the captions.\n",
      "[169000/414113] Tokenized the captions.\n",
      "[170000/414113] Tokenized the captions.\n",
      "[171000/414113] Tokenized the captions.\n",
      "[172000/414113] Tokenized the captions.\n",
      "[173000/414113] Tokenized the captions.\n",
      "[174000/414113] Tokenized the captions.\n",
      "[175000/414113] Tokenized the captions.\n",
      "[176000/414113] Tokenized the captions.\n",
      "[177000/414113] Tokenized the captions.\n",
      "[178000/414113] Tokenized the captions.\n",
      "[179000/414113] Tokenized the captions.\n",
      "[180000/414113] Tokenized the captions.\n",
      "[181000/414113] Tokenized the captions.\n",
      "[182000/414113] Tokenized the captions.\n",
      "[183000/414113] Tokenized the captions.\n",
      "[184000/414113] Tokenized the captions.\n",
      "[185000/414113] Tokenized the captions.\n",
      "[186000/414113] Tokenized the captions.\n",
      "[187000/414113] Tokenized the captions.\n",
      "[188000/414113] Tokenized the captions.\n",
      "[189000/414113] Tokenized the captions.\n",
      "[190000/414113] Tokenized the captions.\n",
      "[191000/414113] Tokenized the captions.\n",
      "[192000/414113] Tokenized the captions.\n",
      "[193000/414113] Tokenized the captions.\n",
      "[194000/414113] Tokenized the captions.\n",
      "[195000/414113] Tokenized the captions.\n",
      "[196000/414113] Tokenized the captions.\n",
      "[197000/414113] Tokenized the captions.\n",
      "[198000/414113] Tokenized the captions.\n",
      "[199000/414113] Tokenized the captions.\n",
      "[200000/414113] Tokenized the captions.\n",
      "[201000/414113] Tokenized the captions.\n",
      "[202000/414113] Tokenized the captions.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[203000/414113] Tokenized the captions.\n",
      "[204000/414113] Tokenized the captions.\n",
      "[205000/414113] Tokenized the captions.\n",
      "[206000/414113] Tokenized the captions.\n",
      "[207000/414113] Tokenized the captions.\n",
      "[208000/414113] Tokenized the captions.\n",
      "[209000/414113] Tokenized the captions.\n",
      "[210000/414113] Tokenized the captions.\n",
      "[211000/414113] Tokenized the captions.\n",
      "[212000/414113] Tokenized the captions.\n",
      "[213000/414113] Tokenized the captions.\n",
      "[214000/414113] Tokenized the captions.\n",
      "[215000/414113] Tokenized the captions.\n",
      "[216000/414113] Tokenized the captions.\n",
      "[217000/414113] Tokenized the captions.\n",
      "[218000/414113] Tokenized the captions.\n",
      "[219000/414113] Tokenized the captions.\n",
      "[220000/414113] Tokenized the captions.\n",
      "[221000/414113] Tokenized the captions.\n",
      "[222000/414113] Tokenized the captions.\n",
      "[223000/414113] Tokenized the captions.\n",
      "[224000/414113] Tokenized the captions.\n",
      "[225000/414113] Tokenized the captions.\n",
      "[226000/414113] Tokenized the captions.\n",
      "[227000/414113] Tokenized the captions.\n",
      "[228000/414113] Tokenized the captions.\n",
      "[229000/414113] Tokenized the captions.\n",
      "[230000/414113] Tokenized the captions.\n",
      "[231000/414113] Tokenized the captions.\n",
      "[232000/414113] Tokenized the captions.\n",
      "[233000/414113] Tokenized the captions.\n",
      "[234000/414113] Tokenized the captions.\n",
      "[235000/414113] Tokenized the captions.\n",
      "[236000/414113] Tokenized the captions.\n",
      "[237000/414113] Tokenized the captions.\n",
      "[238000/414113] Tokenized the captions.\n",
      "[239000/414113] Tokenized the captions.\n",
      "[240000/414113] Tokenized the captions.\n",
      "[241000/414113] Tokenized the captions.\n",
      "[242000/414113] Tokenized the captions.\n",
      "[243000/414113] Tokenized the captions.\n",
      "[244000/414113] Tokenized the captions.\n",
      "[245000/414113] Tokenized the captions.\n",
      "[246000/414113] Tokenized the captions.\n",
      "[247000/414113] Tokenized the captions.\n",
      "[248000/414113] Tokenized the captions.\n",
      "[249000/414113] Tokenized the captions.\n",
      "[250000/414113] Tokenized the captions.\n",
      "[251000/414113] Tokenized the captions.\n",
      "[252000/414113] Tokenized the captions.\n",
      "[253000/414113] Tokenized the captions.\n",
      "[254000/414113] Tokenized the captions.\n",
      "[255000/414113] Tokenized the captions.\n",
      "[256000/414113] Tokenized the captions.\n",
      "[257000/414113] Tokenized the captions.\n",
      "[258000/414113] Tokenized the captions.\n",
      "[259000/414113] Tokenized the captions.\n",
      "[260000/414113] Tokenized the captions.\n",
      "[261000/414113] Tokenized the captions.\n",
      "[262000/414113] Tokenized the captions.\n",
      "[263000/414113] Tokenized the captions.\n",
      "[264000/414113] Tokenized the captions.\n",
      "[265000/414113] Tokenized the captions.\n",
      "[266000/414113] Tokenized the captions.\n",
      "[267000/414113] Tokenized the captions.\n",
      "[268000/414113] Tokenized the captions.\n",
      "[269000/414113] Tokenized the captions.\n",
      "[270000/414113] Tokenized the captions.\n",
      "[271000/414113] Tokenized the captions.\n",
      "[272000/414113] Tokenized the captions.\n",
      "[273000/414113] Tokenized the captions.\n",
      "[274000/414113] Tokenized the captions.\n",
      "[275000/414113] Tokenized the captions.\n",
      "[276000/414113] Tokenized the captions.\n",
      "[277000/414113] Tokenized the captions.\n",
      "[278000/414113] Tokenized the captions.\n",
      "[279000/414113] Tokenized the captions.\n",
      "[280000/414113] Tokenized the captions.\n",
      "[281000/414113] Tokenized the captions.\n",
      "[282000/414113] Tokenized the captions.\n",
      "[283000/414113] Tokenized the captions.\n",
      "[284000/414113] Tokenized the captions.\n",
      "[285000/414113] Tokenized the captions.\n",
      "[286000/414113] Tokenized the captions.\n",
      "[287000/414113] Tokenized the captions.\n",
      "[288000/414113] Tokenized the captions.\n",
      "[289000/414113] Tokenized the captions.\n",
      "[290000/414113] Tokenized the captions.\n",
      "[291000/414113] Tokenized the captions.\n",
      "[292000/414113] Tokenized the captions.\n",
      "[293000/414113] Tokenized the captions.\n",
      "[294000/414113] Tokenized the captions.\n",
      "[295000/414113] Tokenized the captions.\n",
      "[296000/414113] Tokenized the captions.\n",
      "[297000/414113] Tokenized the captions.\n",
      "[298000/414113] Tokenized the captions.\n",
      "[299000/414113] Tokenized the captions.\n",
      "[300000/414113] Tokenized the captions.\n",
      "[301000/414113] Tokenized the captions.\n",
      "[302000/414113] Tokenized the captions.\n",
      "[303000/414113] Tokenized the captions.\n",
      "[304000/414113] Tokenized the captions.\n",
      "[305000/414113] Tokenized the captions.\n",
      "[306000/414113] Tokenized the captions.\n",
      "[307000/414113] Tokenized the captions.\n",
      "[308000/414113] Tokenized the captions.\n",
      "[309000/414113] Tokenized the captions.\n",
      "[310000/414113] Tokenized the captions.\n",
      "[311000/414113] Tokenized the captions.\n",
      "[312000/414113] Tokenized the captions.\n",
      "[313000/414113] Tokenized the captions.\n",
      "[314000/414113] Tokenized the captions.\n",
      "[315000/414113] Tokenized the captions.\n",
      "[316000/414113] Tokenized the captions.\n",
      "[317000/414113] Tokenized the captions.\n",
      "[318000/414113] Tokenized the captions.\n",
      "[319000/414113] Tokenized the captions.\n",
      "[320000/414113] Tokenized the captions.\n",
      "[321000/414113] Tokenized the captions.\n",
      "[322000/414113] Tokenized the captions.\n",
      "[323000/414113] Tokenized the captions.\n",
      "[324000/414113] Tokenized the captions.\n",
      "[325000/414113] Tokenized the captions.\n",
      "[326000/414113] Tokenized the captions.\n",
      "[327000/414113] Tokenized the captions.\n",
      "[328000/414113] Tokenized the captions.\n",
      "[329000/414113] Tokenized the captions.\n",
      "[330000/414113] Tokenized the captions.\n",
      "[331000/414113] Tokenized the captions.\n",
      "[332000/414113] Tokenized the captions.\n",
      "[333000/414113] Tokenized the captions.\n",
      "[334000/414113] Tokenized the captions.\n",
      "[335000/414113] Tokenized the captions.\n",
      "[336000/414113] Tokenized the captions.\n",
      "[337000/414113] Tokenized the captions.\n",
      "[338000/414113] Tokenized the captions.\n",
      "[339000/414113] Tokenized the captions.\n",
      "[340000/414113] Tokenized the captions.\n",
      "[341000/414113] Tokenized the captions.\n",
      "[342000/414113] Tokenized the captions.\n",
      "[343000/414113] Tokenized the captions.\n",
      "[344000/414113] Tokenized the captions.\n",
      "[345000/414113] Tokenized the captions.\n",
      "[346000/414113] Tokenized the captions.\n",
      "[347000/414113] Tokenized the captions.\n",
      "[348000/414113] Tokenized the captions.\n",
      "[349000/414113] Tokenized the captions.\n",
      "[350000/414113] Tokenized the captions.\n",
      "[351000/414113] Tokenized the captions.\n",
      "[352000/414113] Tokenized the captions.\n",
      "[353000/414113] Tokenized the captions.\n",
      "[354000/414113] Tokenized the captions.\n",
      "[355000/414113] Tokenized the captions.\n",
      "[356000/414113] Tokenized the captions.\n",
      "[357000/414113] Tokenized the captions.\n",
      "[358000/414113] Tokenized the captions.\n",
      "[359000/414113] Tokenized the captions.\n",
      "[360000/414113] Tokenized the captions.\n",
      "[361000/414113] Tokenized the captions.\n",
      "[362000/414113] Tokenized the captions.\n",
      "[363000/414113] Tokenized the captions.\n",
      "[364000/414113] Tokenized the captions.\n",
      "[365000/414113] Tokenized the captions.\n",
      "[366000/414113] Tokenized the captions.\n",
      "[367000/414113] Tokenized the captions.\n",
      "[368000/414113] Tokenized the captions.\n",
      "[369000/414113] Tokenized the captions.\n",
      "[370000/414113] Tokenized the captions.\n",
      "[371000/414113] Tokenized the captions.\n",
      "[372000/414113] Tokenized the captions.\n",
      "[373000/414113] Tokenized the captions.\n",
      "[374000/414113] Tokenized the captions.\n",
      "[375000/414113] Tokenized the captions.\n",
      "[376000/414113] Tokenized the captions.\n",
      "[377000/414113] Tokenized the captions.\n",
      "[378000/414113] Tokenized the captions.\n",
      "[379000/414113] Tokenized the captions.\n",
      "[380000/414113] Tokenized the captions.\n",
      "[381000/414113] Tokenized the captions.\n",
      "[382000/414113] Tokenized the captions.\n",
      "[383000/414113] Tokenized the captions.\n",
      "[384000/414113] Tokenized the captions.\n",
      "[385000/414113] Tokenized the captions.\n",
      "[386000/414113] Tokenized the captions.\n",
      "[387000/414113] Tokenized the captions.\n",
      "[388000/414113] Tokenized the captions.\n",
      "[389000/414113] Tokenized the captions.\n",
      "[390000/414113] Tokenized the captions.\n",
      "[391000/414113] Tokenized the captions.\n",
      "[392000/414113] Tokenized the captions.\n",
      "[393000/414113] Tokenized the captions.\n",
      "[394000/414113] Tokenized the captions.\n",
      "[395000/414113] Tokenized the captions.\n",
      "[396000/414113] Tokenized the captions.\n",
      "[397000/414113] Tokenized the captions.\n",
      "[398000/414113] Tokenized the captions.\n",
      "[399000/414113] Tokenized the captions.\n",
      "[400000/414113] Tokenized the captions.\n",
      "[401000/414113] Tokenized the captions.\n",
      "[402000/414113] Tokenized the captions.\n",
      "[403000/414113] Tokenized the captions.\n",
      "[404000/414113] Tokenized the captions.\n",
      "[405000/414113] Tokenized the captions.\n",
      "[406000/414113] Tokenized the captions.\n",
      "[407000/414113] Tokenized the captions.\n",
      "[408000/414113] Tokenized the captions.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[409000/414113] Tokenized the captions.\n",
      "[410000/414113] Tokenized the captions.\n",
      "[411000/414113] Tokenized the captions.\n",
      "[412000/414113] Tokenized the captions.\n",
      "[413000/414113] Tokenized the captions.\n",
      "[414000/414113] Tokenized the captions.\n"
     ]
    }
   ],
   "source": [
    "training_ids = select_ann_ids(load_image_ids(True), \"captions_train2014.json\")\n",
    "\n",
    "validation_ids = training_ids[:int(len(training_ids) * (VALIDATION_SIZE))]\n",
    "training_ids = training_ids[int(len(training_ids) * (VALIDATION_SIZE)):]\n",
    "testing_ids = select_ann_ids(load_image_ids(False), \"captions_val2014.json\")\n",
    "\n",
    "# load data and transform images\n",
    "train_dataset = get_loader(IMAGES_DIR + \"/train/\", CAPTIONS_DIR + \"/captions_train2014.json\", training_ids,\n",
    "                           BATCH_SIZE, SHUFFLE_DATA,\n",
    "                           NUM_WORKERS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|          | 0/5 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Minibatch 0\n",
      "torch.Size([2728])\n",
      "Running encoder\n",
      "Running decoder\n",
      "Running backward\n",
      "Minibatch 1\n",
      "torch.Size([2668])\n",
      "Running encoder\n",
      "Running decoder\n",
      "Running backward\n",
      "Minibatch 2\n",
      "torch.Size([2630])\n",
      "Running encoder\n",
      "Running decoder\n",
      "Running backward\n",
      "Minibatch 3\n",
      "torch.Size([2688])\n",
      "Running encoder\n",
      "Running decoder\n",
      "Running backward\n",
      "Minibatch 4\n",
      "torch.Size([383])\n",
      "Running encoder\n",
      "Running decoder\n",
      "Running backward\n",
      "Minibatch 0\n",
      "torch.Size([2687])\n",
      "Running encoder\n",
      "Running decoder\n",
      "torch.Size([2687])\n",
      "vs\n",
      "torch.Size([200, 25])\n"
     ]
    },
    {
     "ename": "RuntimeError",
     "evalue": "\"host_softmax\" not implemented for 'Long'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mRuntimeError\u001b[0m                              Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-3-7b8a29de8ab6>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      7\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      8\u001b[0m \u001b[0mrunner\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mRunner\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mencoder\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdecoder\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtrain_dataset\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtrain_dataset\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtrain_dataset\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 9\u001b[0;31m \u001b[0mrunner\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain\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[0;32m~/PA4/runner.py\u001b[0m in \u001b[0;36mtrain\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m     41\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mepoch\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtqdm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mEPOCHS\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     42\u001b[0m             \u001b[0mtrain_loss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpass_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain_dataset\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---> 43\u001b[0;31m             \u001b[0mval_loss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpass_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mval_dataset\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     44\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     45\u001b[0m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtraining_data\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mtrain_loss\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mval_loss\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~/PA4/runner.py\u001b[0m in \u001b[0;36mpass_data\u001b[0;34m(self, dataset, backward)\u001b[0m\n\u001b[1;32m    101\u001b[0m                 \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"vs\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    102\u001b[0m                 \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpredicted\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msize\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--> 103\u001b[0;31m             \u001b[0mbatch_loss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcriterion\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpredicted\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtargets\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    104\u001b[0m             \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Running backward\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    105\u001b[0m             \u001b[0;31m# backward\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/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    539\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    540\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--> 541\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    542\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    543\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;32m/opt/conda/lib/python3.7/site-packages/torch/nn/modules/loss.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, input, target)\u001b[0m\n\u001b[1;32m    914\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mforward\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[0mtarget\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    915\u001b[0m         return F.cross_entropy(input, target, weight=self.weight,\n\u001b[0;32m--> 916\u001b[0;31m                                ignore_index=self.ignore_index, reduction=self.reduction)\n\u001b[0m\u001b[1;32m    917\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    918\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/torch/nn/functional.py\u001b[0m in \u001b[0;36mcross_entropy\u001b[0;34m(input, target, weight, size_average, ignore_index, reduce, reduction)\u001b[0m\n\u001b[1;32m   2007\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0msize_average\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mreduce\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\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   2008\u001b[0m         \u001b[0mreduction\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_Reduction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlegacy_get_string\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msize_average\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mreduce\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2009\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mnll_loss\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlog_softmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtarget\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mweight\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mignore_index\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mreduction\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   2010\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2011\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/torch/nn/functional.py\u001b[0m in \u001b[0;36mlog_softmax\u001b[0;34m(input, dim, _stacklevel, dtype)\u001b[0m\n\u001b[1;32m   1315\u001b[0m         \u001b[0mdim\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_get_softmax_dim\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'log_softmax'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdim\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0m_stacklevel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1316\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mdtype\u001b[0m \u001b[0;32mis\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-> 1317\u001b[0;31m         \u001b[0mret\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlog_softmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdim\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   1318\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[1;32m   1319\u001b[0m         \u001b[0mret\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlog_softmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdim\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdtype\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: \"host_softmax\" not implemented for 'Long'"
     ]
    }
   ],
   "source": [
    "vocabulary_size = len(train_dataset.dataset.vocab.word2idx)\n",
    "\n",
    "computing_device = get_device()\n",
    "\n",
    "encoder = Encoder(EMBEDDED_SIZE).to(computing_device)\n",
    "decoder = Decoder(EMBEDDED_SIZE, LSTM_HIDDEN_SIZE, vocabulary_size).to(computing_device)\n",
    "\n",
    "runner = Runner(encoder, decoder, train_dataset, train_dataset, train_dataset)\n",
    "runner.train()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "vocabulary_size = len(train_dataset.dataset.vocab.word2idx)\n",
    "\n",
    "encoder = Encoder(EMBEDDED_SIZE)\n",
    "decoder = Decoder(EMBEDDED_SIZE, LSTM_HIDDEN_SIZE, vocabulary_size)\n",
    "\n",
    "runner = Runner(encoder, decoder, train_dataset, train_dataset, train_dataset)\n",
    "runner.train()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "val_dataset = get_loader(IMAGES_DIR + \"/train/\", CAPTIONS_DIR + \"/captions_train2014.json\", validation_ids,\n",
    "                             BATCH_SIZE, SHUFFLE_DATA,\n",
    "                             NUM_WORKERS)\n",
    "test_dataset = get_loader(IMAGES_DIR + \"/test/\", CAPTIONS_DIR + \"/captions_val2014.json/\", testing_ids, BATCH_SIZE,\n",
    "                              SHUFFLE_DATA,\n",
    "                              NUM_WORKERS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for param in encoder.mod.param():\n",
    "    print(param.requires_grad)\n",
    "    param.requires_grad = False\n",
    "    print(type(param))"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}