image_captioning_with_keras6.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#IMPORTS\n",
    "import string\n",
    "from collections import Counter\n",
    "import datetime as dt\n",
    "from numpy import array\n",
    "import numpy as np\n",
    "import pickle\n",
    "from pickle import load, dump\n",
    "# from keras.applications.inception_v3 import InceptionV3\n",
    "# from keras.applications import ResNet152V2\n",
    "from keras.applications import InceptionResNetV2\n",
    "from keras.models import Model\n",
    "from keras.preprocessing import image\n",
    "# from keras.applications.inception_v3 import preprocess_input\n",
    "# from keras.applications.resnet_v2 import preprocess_input\n",
    "from keras.applications.inception_resnet_v2 import preprocess_input\n",
    "from keras import Input\n",
    "from keras.layers import Dropout, Dense, Embedding, LSTM, GaussianDropout, BatchNormalization\n",
    "from keras.layers.merge import add\n",
    "from keras.preprocessing.sequence import pad_sequences\n",
    "from keras.utils import to_categorical\n",
    "import keras.backend as K\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "                Logging results to <a href=\"https://wandb.com\" target=\"_blank\">Weights & Biases</a> <a href=\"https://docs.wandb.com/integrations/jupyter.html\" target=\"_blank\">(Documentation)</a>.<br/>\n",
       "                Project page: <a href=\"https://app.wandb.ai/vipul43/model7\" target=\"_blank\">https://app.wandb.ai/vipul43/model7</a><br/>\n",
       "                Run page: <a href=\"https://app.wandb.ai/vipul43/model7/runs/368ztcwf\" target=\"_blank\">https://app.wandb.ai/vipul43/model7/runs/368ztcwf</a><br/>\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[34m\u001b[1mwandb\u001b[0m: Wandb version 0.9.3 is available!  To upgrade, please run:\n",
      "\u001b[34m\u001b[1mwandb\u001b[0m:  $ pip install wandb --upgrade\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "W&B Run: https://app.wandb.ai/vipul43/model7/runs/368ztcwf"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#RUN THIS CELL ONLY ONCE\n",
    "import wandb\n",
    "wandb.init(project=\"model7\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#EXTRACTING TEXT DATA AND PREPROCESSING\n",
    "#STEP: EXTRACTING DATA\n",
    "filepath = \"../../../Downloads/Flickr8k/Flickr8k_text/Flickr8k.token.txt\"\n",
    "file = open(filepath, 'r')\n",
    "content = file.read()\n",
    "file.close()\n",
    "lines = content.split('\\n')\n",
    "\n",
    "#STEP: SAVING DATA IN A DICTIONARY FOR TIME EFFICIENCY PURPOSES\n",
    "tokens_dic = {}\n",
    "for line in lines:\n",
    "    tokens = line.split()\n",
    "    image_ = tokens[0]\n",
    "    caption = ' '.join(tokens[1:])\n",
    "    image_title = image_.split('.')[0]\n",
    "    if(image_title not in tokens_dic.keys()):\n",
    "        tokens_dic[image_title] = [caption]\n",
    "    else:\n",
    "        tokens_dic[image_title].append(caption)\n",
    "\n",
    "\n",
    "#STEP: PREPROCESSING\n",
    "table = str.maketrans('', '', string.punctuation)\n",
    "for captions in tokens_dic.values():\n",
    "    for i in range(len(captions)):\n",
    "        caption = captions[i]\n",
    "        words = caption.split()\n",
    "        words = [word.lower() for word in words]\n",
    "        words = [word.translate(table) for word in words]\n",
    "        words = [word for word in words if len(word)>1]\n",
    "        words = [word for word in words if word.isalpha()]\n",
    "        captions[i] = ' '.join(words)\n",
    "        \n",
    "\n",
    "#LOADING TRAINING SET\n",
    "\n",
    "#STEP: EXTRACTING DATA\n",
    "filepath = \"../../../Downloads/Flickr8k/Flickr8k_text/Flickr_8k.trainImages.txt\"\n",
    "file = open(filepath, \"r\")\n",
    "content = file.read()\n",
    "file.close()\n",
    "lines = content.split('\\n')\n",
    "\n",
    "#STEP: SAVING DATA IN A LIST\n",
    "train = []\n",
    "for line in lines:\n",
    "    [image_title, stuff] =  line.split('.')\n",
    "    train.append(image_title)\n",
    "#NOTE: SIZE OF TRAIN LIST IS 6000\n",
    "\n",
    "#STEP: SAVING IMAGE-CAPTIONS IN TRAIN DATASET\n",
    "train_dataset = {}\n",
    "max_caption_length = 0\n",
    "for image_title, captions in tokens_dic.items():\n",
    "    if(image_title in train and (image_title not in train_dataset.keys())):\n",
    "        train_dataset[image_title] = list()\n",
    "        for caption in captions:\n",
    "            refined_caption = \"startseq \" + caption + \" endseq\"\n",
    "            train_dataset[image_title].append(refined_caption)\n",
    "            max_caption_length = max(max_caption_length, len(refined_caption.split()))\n",
    "#NOTE: SIZE OF TRAIN DATASET IS 6000\n",
    "\n",
    "\n",
    "#LOADING CROSS VALIDATION(OR DEVELOPMENT) SET\n",
    "\n",
    "#STEP: EXTRACTING DATA\n",
    "filepath = \"../../../Downloads/Flickr8k/Flickr8k_text/Flickr_8k.devImages.txt\"\n",
    "file = open(filepath, \"r\")\n",
    "content = file.read()\n",
    "file.close()\n",
    "lines = content.split('\\n')\n",
    "\n",
    "#STEP: SAVING DATA IN A LIST\n",
    "dev = []\n",
    "for line in lines:\n",
    "    [image_title, stuff] =  line.split('.')\n",
    "    dev.append(image_title)\n",
    "#NOTE: SIZE OF TEST LIST IS 1000\n",
    "\n",
    "#STEP: SAVING IMAGE-CAPTIONS IN TRAIN DATASET\n",
    "dev_dataset = {}\n",
    "for image_title, captions in tokens_dic.items():\n",
    "    if(image_title in dev and (image_title not in dev_dataset.keys())):\n",
    "        dev_dataset[image_title] = list()\n",
    "        for caption in captions:\n",
    "            refined_caption = \"startseq \" + caption + \" endseq\"\n",
    "            dev_dataset[image_title].append(refined_caption)\n",
    "#NOTE: SIZE OF DEV DATASET IS 1000\n",
    "\n",
    "\n",
    "#LOADING TESTING SET\n",
    "\n",
    "#STEP: EXTRACTING DATA\n",
    "filepath = \"../../../Downloads/Flickr8k/Flickr8k_text/Flickr_8k.testImages.txt\"\n",
    "file = open(filepath, \"r\")\n",
    "content = file.read()\n",
    "file.close()\n",
    "lines = content.split('\\n')\n",
    "\n",
    "#STEP: SAVING DATA IN A LIST\n",
    "test = []\n",
    "for line in lines:\n",
    "    [image_title, stuff] =  line.split('.')\n",
    "    test.append(image_title)\n",
    "#NOTE: SIZE OF TEST LIST IS 1000\n",
    "\n",
    "#STEP: SAVING IMAGE-CAPTIONS IN TRAIN DATASET\n",
    "test_dataset = {}\n",
    "for image_title, captions in tokens_dic.items():\n",
    "    if(image_title in test and (image_title not in test_dataset.keys())):\n",
    "        test_dataset[image_title] = list()\n",
    "        for caption in captions:\n",
    "            refined_caption = \"startseq \" + caption + \" endseq\"\n",
    "            test_dataset[image_title].append(refined_caption)\n",
    "#NOTE: SIZE OF TEST DATASET IS 1000\n",
    "\n",
    "\n",
    "#CONSTRUCTING VOCABULARY FROM CAPTIONS\n",
    "vocabulary = set()\n",
    "for captions in tokens_dic.values():\n",
    "    for caption in captions:\n",
    "        for word in caption.split():\n",
    "            vocabulary.add(word)\n",
    "#NOTE: SIZE OF VOCABULARY BY NOT LETTING 1 LENGTH WORDS BE IN CAPTION IS 8763\n",
    "\n",
    "#CONSTRUCTING MOST PROBABLE VOCABULARY FROM ALL TRAIN WORDS\n",
    "word_count_threshold = 10\n",
    "all_train_words = []\n",
    "for captions in train_dataset.values():\n",
    "    for caption in captions:\n",
    "        words = caption.split()\n",
    "        for word in words:\n",
    "            all_train_words.append(word)\n",
    "\n",
    "counter = Counter(all_train_words)\n",
    "commons = counter.most_common()\n",
    "most_probable_vocabulary = set()\n",
    "for ele in commons:\n",
    "    if(ele[1]>=word_count_threshold):\n",
    "        most_probable_vocabulary.add(ele[0])\n",
    "vocabulary_size = len(most_probable_vocabulary)\n",
    "#NOTE: SIZE OF MOST PROBABLE VOCABULARY BY NOT LETTING 1 LENGTH WORDS BE IN CAPTION IS 1651\n",
    "\n",
    "\n",
    "#READING ENCODED TRAIN IMAGES, ENCODED DEV IMAGES AND ENCODED TEST IMAGES FROM PKL FILES\n",
    "encoded_train_images = {}\n",
    "encoded_dev_images = {}\n",
    "encoded_test_images = {}\n",
    "with open(\"encoded_train_images3.pkl\", \"rb\") as encoded_pickle:\n",
    "    encoded_train_images = load(encoded_pickle)\n",
    "    \n",
    "with open(\"encoded_test_images3.pkl\", \"rb\") as encoded_pickle:\n",
    "    encoded_test_images = load(encoded_pickle)\n",
    "    \n",
    "with open(\"encoded_dev_images3.pkl\", \"rb\") as encoded_pickle:\n",
    "    encoded_dev_images = load(encoded_pickle)\n",
    "\n",
    "\n",
    "def data_generator(descriptions, photos, wordtoix, max_length, num_photos_per_batch):\n",
    "    X1, X2, y = list(), list(), list()\n",
    "    n=0\n",
    "    while 1:\n",
    "        for key, desc_list in descriptions.items():\n",
    "            n+=1\n",
    "            photo = photos[key]\n",
    "            for desc in desc_list:\n",
    "                seq = [wordtoix[word] for word in desc.split(' ') if word in wordtoix]\n",
    "                for i in range(1, len(seq)):\n",
    "                    in_seq, out_seq = seq[:i], seq[i]\n",
    "                    in_seq = pad_sequences([in_seq], maxlen=max_length)[0]\n",
    "                    out_seq = to_categorical([out_seq], num_classes=vocabulary_size+1)[0]\n",
    "                    X1.append(photo)\n",
    "                    X2.append(in_seq)\n",
    "                    y.append(out_seq)\n",
    "            if n==num_photos_per_batch:\n",
    "                yield [[array(X1), array(X2)], array(y)]\n",
    "                X1, X2, y = list(), list(), list()\n",
    "                n=0\n",
    "                \n",
    "                \n",
    "#USEFUL DICTIONARIES\n",
    "index_to_word = {}\n",
    "word_to_index = {}\n",
    "index = 1\n",
    "for word in most_probable_vocabulary:\n",
    "    index_to_word[index] = word\n",
    "    word_to_index[word] = index\n",
    "    index+=1\n",
    "\n",
    "\n",
    "#WORD EMBEDDINGS, MAPPING EVERY WORD OF OUR MOST PROBABALE VOCABULARY TO 200 DIMENSION VECTOR, FOR THAT WE WILL BE USING GLOVE\n",
    "embeddings = {}\n",
    "file = open(\"../../../Downloads/glove/glove.6B.200d.txt\", \"r\")\n",
    "content = file.read()\n",
    "file.close()\n",
    "lines = content.split('\\n')\n",
    "for line in lines:\n",
    "    word_and_vector = line.split()\n",
    "    word = word_and_vector[0]\n",
    "    vector = word_and_vector[1:]\n",
    "    vector = np.asarray(vector, dtype='float32')\n",
    "    embeddings[word] = vector\n",
    "    \n",
    "    \n",
    "#MAPPING MOST PROBABLE VOCABULARY TO VECTOR BY CREATING A DENSE MATRIX WITH ROW AS WORD AND COLUMNS ARE 200 DIMENSIONS OF VECTOR\n",
    "embedding_dim = 200\n",
    "embeddings_matrix = np.zeros((vocabulary_size+1, embedding_dim))\n",
    "for word, index in word_to_index.items():\n",
    "    vector = embeddings.get(word)\n",
    "    if(vector is not None):\n",
    "        embeddings_matrix[index]=vector\n",
    "#NOTE: DIMENSION OF EMBEDDINGS MATRIX IS (MOST PROBABLE VOCABULARY X 200), HERE (1951 X 200)\n",
    "#NOTE: 1ST ROW OF EMBEDDINGS MATRIX IS ALL ZEROS"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# model\n",
    "<img src=\"files/images/model1_arch.png\">\n",
    "<img src=\"files/images/model_summary.png\">\n",
    "<img src=\"files/images/model_performance.png\">\n",
    "\n",
    "# model\n",
    "<img src=\"files/images/model2_arch.png\">\n",
    "<img src=\"files/images/model2_summary.png\">\n",
    "<img src=\"files/images/model2_performance.png\">\n",
    "\n",
    "# model\n",
    "<img src=\"files/images/model3_arch.png\">\n",
    "<img src=\"files/images/model3_summary.png\">\n",
    "<img src=\"files/images/model3_performance.png\">\n",
    "\n",
    "# model\n",
    "<img src=\"files/images/model4_arch.png\">\n",
    "<img src=\"files/images/model4_summary.png\">\n",
    "<img src=\"files/images/model4_performance.png\">\n",
    "\n",
    "# model\n",
    "<img src=\"files/images/model5_arch.png\">\n",
    "<img src=\"files/images/model5_summary.png\">\n",
    "<img src=\"files/images/model5_performance.png\">\n",
    "\n",
    "# model\n",
    "<img src=\"files/images/model6_arch.png\">\n",
    "<img src=\"files/images/model6_summary.png\">\n",
    "<img src=\"files/images/model6_performance.png\">"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model7\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_15 (InputLayer)           (None, 34)           0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_14 (InputLayer)           (None, 1536)         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "embedding_7 (Embedding)         (None, 34, 200)      330400      input_15[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "gaussian_dropout_7 (GaussianDro (None, 1536)         0           input_14[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "gaussian_dropout_8 (GaussianDro (None, 34, 200)      0           embedding_7[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_13 (Dense)                (None, 256)          393472      gaussian_dropout_7[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "lstm_9 (LSTM)                   (None, 256)          467968      gaussian_dropout_8[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "add_5 (Add)                     (None, 256)          0           dense_13[0][0]                   \n",
      "                                                                 lstm_9[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_2 (BatchNor (None, 256)          1024        add_5[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "dense_14 (Dense)                (None, 256)          65792       batch_normalization_2[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "dense_15 (Dense)                (None, 1652)         424564      dense_14[0][0]                   \n",
      "==================================================================================================\n",
      "Total params: 1,683,220\n",
      "Trainable params: 1,682,708\n",
      "Non-trainable params: 512\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#SAME ARCHITECTURE AS MODEL\n",
    "input1 = Input(shape=(1536,))\n",
    "fe1 = GaussianDropout(0.5)(input1)\n",
    "fe2 = Dense(256, activation='relu')(fe1)\n",
    "\n",
    "input2 = Input(shape=(max_caption_length,))\n",
    "se1 = Embedding(vocabulary_size+1, embedding_dim, mask_zero=True)(input2)\n",
    "se2 = GaussianDropout(0.5)(se1)\n",
    "se3 = LSTM(256)(se2)\n",
    "\n",
    "decoder1 = add([fe2, se3])\n",
    "decoder2 = BatchNormalization()(decoder1)\n",
    "decoder3 = Dense(256, activation='relu')(decoder2)\n",
    "\n",
    "output = Dense(vocabulary_size+1, activation='softmax')(decoder3)\n",
    "model7 = Model(inputs=[input1, input2], outputs=output, name=\"model7\")\n",
    "\n",
    "model7.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SETTING ADDITIONAL PROPERTIES FOR EMBEDDING LAYER\n",
    "model7.layers[2].set_weights([embeddings_matrix])\n",
    "model7.layers[2].trainable = False\n",
    "\n",
    "model7.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[34m\u001b[1mwandb\u001b[0m: Wandb version 0.9.3 is available!  To upgrade, please run:\n",
      "\u001b[34m\u001b[1mwandb\u001b[0m:  $ pip install wandb --upgrade\n"
     ]
    }
   ],
   "source": [
    "wandb.config.dropout = 0.5\n",
    "wandb.config.hidden_layer_size = 3\n",
    "wandb.config.epochs=30"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "#HYPERPARAMETERS OF TRAINING SESSION 1\n",
    "#NOTE: DEFAULT LEARNING RATE OF KERAS MODEL IS 0.001\n",
    "epochs = 20\n",
    "batch_size = 3\n",
    "steps = len(train_dataset)//batch_size\n",
    "#NOTE: SIZE OF TRAIN DATASET IS 6000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 721s 360ms/step - loss: 4.0434 - accuracy: 0.2551\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[34m\u001b[1mwandb\u001b[0m: Wandb version 0.9.3 is available!  To upgrade, please run:\n",
      "\u001b[34m\u001b[1mwandb\u001b[0m:  $ pip install wandb --upgrade\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 679s 340ms/step - loss: 3.4167 - accuracy: 0.3048\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 676s 338ms/step - loss: 3.2032 - accuracy: 0.3259\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 679s 339ms/step - loss: 3.0660 - accuracy: 0.3396\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 679s 340ms/step - loss: 2.9692 - accuracy: 0.3489\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 681s 340ms/step - loss: 2.8915 - accuracy: 0.3569\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 681s 340ms/step - loss: 2.8263 - accuracy: 0.3640\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 683s 342ms/step - loss: 2.7696 - accuracy: 0.3694\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 689s 345ms/step - loss: 2.7231 - accuracy: 0.3744\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 684s 342ms/step - loss: 2.6815 - accuracy: 0.3802\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 682s 341ms/step - loss: 2.6461 - accuracy: 0.3842\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 680s 340ms/step - loss: 2.6128 - accuracy: 0.3881\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 679s 340ms/step - loss: 2.5827 - accuracy: 0.3910\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 681s 340ms/step - loss: 2.5532 - accuracy: 0.3954\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 682s 341ms/step - loss: 2.5303 - accuracy: 0.3980\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 683s 341ms/step - loss: 2.5058 - accuracy: 0.4023\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 691s 345ms/step - loss: 2.4860 - accuracy: 0.4037\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 691s 346ms/step - loss: 2.4663 - accuracy: 0.4066\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 688s 344ms/step - loss: 2.4487 - accuracy: 0.4093\n",
      "Epoch 1/1\n",
      "2000/2000 [==============================] - 692s 346ms/step - loss: 2.4343 - accuracy: 0.4108\n"
     ]
    }
   ],
   "source": [
    "#TRAINING SESSION 1\n",
    "for i in range(epochs):\n",
    "    generator = data_generator(train_dataset, encoded_train_images, word_to_index, max_caption_length, batch_size)\n",
    "    hist = model7.fit_generator(generator, epochs=1, steps_per_epoch=steps, verbose=1)\n",
    "    loss = hist.history['loss']\n",
    "    wandb.log({'epoch': i+1, 'loss': loss})\n",
    "model7.save('./model7_weights/model_' + str(0) + '.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "#HYPERPARAMETERS OF TRAINING SESSION 2\n",
    "#NOTE: LEARNING RATE IS CHANGED TO 0.0003\n",
    "K.set_value(model7.optimizer.lr, 0.0003)\n",
    "epochs = 5\n",
    "batch_size = 4\n",
    "steps = len(train_dataset)//batch_size\n",
    "#NOTE: SIZE OF TRAIN DATASET IS 6000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1\n",
      "1500/1500 [==============================] - 640s 427ms/step - loss: 2.4496 - accuracy: 0.4076\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[34m\u001b[1mwandb\u001b[0m: Wandb version 0.9.3 is available!  To upgrade, please run:\n",
      "\u001b[34m\u001b[1mwandb\u001b[0m:  $ pip install wandb --upgrade\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1\n",
      "1500/1500 [==============================] - 633s 422ms/step - loss: 2.4070 - accuracy: 0.4144\n",
      "Epoch 1/1\n",
      "1500/1500 [==============================] - 627s 418ms/step - loss: 2.3842 - accuracy: 0.4177\n",
      "Epoch 1/1\n",
      "1500/1500 [==============================] - 629s 419ms/step - loss: 2.3612 - accuracy: 0.4214\n",
      "Epoch 1/1\n",
      "1500/1500 [==============================] - 628s 419ms/step - loss: 2.3443 - accuracy: 0.4228\n"
     ]
    }
   ],
   "source": [
    "#TRAINING SESSION 2\n",
    "for i in range(epochs):\n",
    "    generator = data_generator(train_dataset, encoded_train_images, word_to_index, max_caption_length, batch_size)\n",
    "    hist = model7.fit_generator(generator, epochs=1, steps_per_epoch=steps, verbose=1)\n",
    "    loss = hist.history['loss']\n",
    "    wandb.log({'epoch': i+21, 'loss': loss})\n",
    "model7.save('./model7_weights/model_' + str(1) + '.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "#HYPERPARAMETERS OF TRAINING SESSION 3\n",
    "#NOTE: LEARNING RATE IS CHANGED TO 0.0001\n",
    "K.set_value(model7.optimizer.lr, 0.0001)\n",
    "epochs = 5\n",
    "batch_size = 6\n",
    "steps = len(train_dataset)//batch_size\n",
    "#NOTE: SIZE OF TRAIN DATASET IS 6000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1\n",
      "1000/1000 [==============================] - 579s 579ms/step - loss: 2.3212 - accuracy: 0.4232\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[34m\u001b[1mwandb\u001b[0m: Wandb version 0.9.3 is available!  To upgrade, please run:\n",
      "\u001b[34m\u001b[1mwandb\u001b[0m:  $ pip install wandb --upgrade\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1\n",
      "1000/1000 [==============================] - 582s 582ms/step - loss: 2.3101 - accuracy: 0.4263\n",
      "Epoch 1/1\n",
      "1000/1000 [==============================] - 577s 577ms/step - loss: 2.3044 - accuracy: 0.4262\n",
      "Epoch 1/1\n",
      "1000/1000 [==============================] - 576s 576ms/step - loss: 2.2969 - accuracy: 0.4277\n",
      "Epoch 1/1\n",
      "1000/1000 [==============================] - 576s 576ms/step - loss: 2.2877 - accuracy: 0.4288\n"
     ]
    }
   ],
   "source": [
    "#TRAINING SESSION 3\n",
    "for i in range(epochs):\n",
    "    generator = data_generator(train_dataset, encoded_train_images, word_to_index, max_caption_length, batch_size)\n",
    "    hist = model7.fit_generator(generator, epochs=1, steps_per_epoch=steps, verbose=1)\n",
    "    loss = hist.history['loss']\n",
    "    wandb.log({'epoch': i+26, 'loss': loss})\n",
    "model7.save('./model7_weights/model_' + str(2) + '.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "#LOADING TRAINED WEIGHTS TO MODEL\n",
    "model7.load_weights('./model7_weights/model_2.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict7(feature_vec):\n",
    "    partial_caption = \"startseq\"\n",
    "    for i in range(max_caption_length):\n",
    "        seq = [word_to_index[word] for word in partial_caption.split() if word in word_to_index]\n",
    "        seq = pad_sequences([seq], maxlen=max_caption_length)\n",
    "        yhat = model7.predict([feature_vec,seq], verbose=0)\n",
    "        yhat = np.argmax(yhat)\n",
    "        word = index_to_word[yhat]\n",
    "        partial_caption += ' ' + word\n",
    "        if word == 'endseq':\n",
    "            break\n",
    "    final_caption = partial_caption.split()[1:-1]\n",
    "    final_caption = ' '.join(final_caption)\n",
    "    return final_caption"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "model7\n",
      "3115901702_f07aa0ef74\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": [
      "prediction: two girls in skirts walking on the sidewalk\n",
      "Ground Truth: woman walks as she looks at her cellphone and others look on\n",
      "Ground Truth: several ladies talking together\n",
      "Ground Truth: three girls walking one is checking her cellphone\n",
      "Ground Truth: three girls walk together and talk\n",
      "Ground Truth: three high school girls walk and talk on astroturf\n"
     ]
    }
   ],
   "source": [
    "#OBSERVATION OF PERFORMANCE ON RANDOM IMAGES FROM DEV DATASET\n",
    "key = list(dev_dataset.keys())[np.random.randint(0, 1000)]\n",
    "print(\"model7\")\n",
    "print(key)\n",
    "feature_vec = encoded_dev_images[key].reshape((1, 1536))\n",
    "image_path = \"../../../Downloads/Flickr8k/Flicker8k_dataset/\"+key+\".jpg\"\n",
    "image_ = plt.imread(image_path)\n",
    "plt.imshow(image_)\n",
    "plt.show()\n",
    "print(\"prediction: {}\".format(predict7(feature_vec)))\n",
    "for caption in dev_dataset[key]:\n",
    "    caption = ' '.join(caption.split()[1:-1])\n",
    "    print(\"Ground Truth: {}\".format(caption))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LOSS ON TRAIN DATASET: [2.672255039215088, 0.4269621670246124]\n",
      "LOSS ON DEV DATASET: [4.131709098815918, 0.3226974308490753]\n"
     ]
    }
   ],
   "source": [
    "#PERFORMANCE ON TRAIN DATASET, DEV DATASET\n",
    "batch_size = 3\n",
    "steps = len(train_dataset)//batch_size\n",
    "train_generator = data_generator(train_dataset, encoded_train_images, word_to_index, max_caption_length, batch_size)\n",
    "train_performance = model7.evaluate(train_generator, steps=steps, verbose=0)\n",
    "\n",
    "batch_size = 4\n",
    "steps = len(dev_dataset)//batch_size\n",
    "dev_generator = data_generator(dev_dataset, encoded_dev_images, word_to_index, max_caption_length, batch_size)\n",
    "dev_performance = model7.evaluate(dev_generator, steps=steps, verbose=0)\n",
    "\n",
    "print(\"LOSS ON TRAIN DATASET: {}\".format(train_performance))\n",
    "print(\"LOSS ON DEV DATASET: {}\".format(dev_performance))"
   ]
  },
  {
   "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}