DL_GESTRURE_RECOGNITION_HARIKA.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "3d941679",
   "metadata": {},
   "source": [
    "# Gesture Recognition\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9bd6cb7d",
   "metadata": {},
   "source": [
    "In this group project, you are going to build a 3D Conv model that will be able to predict the 5 gestures correctly. Please import the following libraries to get started."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0570e30d",
   "metadata": {},
   "source": [
    "Imagine you are working as a data scientist at a home electronics company which manufactures state of the art smart televisions. You want to develop a cool feature in the smart-TV that can recognise five different gestures performed by the user which will help users control the TV without using a remote.\n",
    "\n",
    "The gestures are continuously monitored by the webcam mounted on the TV. Each gesture corresponds to a specific command:\n",
    "\n",
    "Thumbs up: Increase the volume\n",
    "\n",
    "Thumbs down: Decrease the volume\n",
    "\n",
    "Left swipe: 'Jump' backwards 10 seconds\n",
    "\n",
    "Right swipe: 'Jump' forward 10 seconds\n",
    "\n",
    "Stop: Pause the movie"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "06b71d12",
   "metadata": {},
   "outputs": [],
   "source": [
    "#importing necessary libraries\n",
    "\n",
    "import numpy as np\n",
    "import os\n",
    "import scipy.misc\n",
    "import PIL\n",
    "from cv2 import imread,resize\n",
    "import datetime\n",
    "import os\n",
    "import pathlib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "2bff5cdc",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(30)\n",
    "import random as rn\n",
    "rn.seed(30)\n",
    "from keras import backend as K\n",
    "import tensorflow as tf\n",
    "tf.random.set_seed(30)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8f09f1af",
   "metadata": {},
   "source": [
    "In this block, you read the folder names for training and validation. You also set the batch_size here. Note that you set the batch size in such a way that you are able to use the GPU in full capacity. You keep increasing the batch size until the machine throws an error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "aff28ed0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home\n"
     ]
    }
   ],
   "source": [
    "path = pathlib.Path.cwd()\n",
    "print(path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "96fbf077",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_doc = np.random.permutation(open('/datasets/Project_data/train.csv').readlines())\n",
    "val_doc = np.random.permutation(open('/datasets/Project_data/val.csv').readlines())\n",
    "batch_size = 20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "2f531327",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['WIN_20180926_16_54_08_Pro_Right_Swipe_new;Right_Swipe_new;1\\n',\n",
       "       'WIN_20180925_18_02_58_Pro_Thumbs_Down_new;Thumbs_Down_new;3\\n',\n",
       "       'WIN_20180925_17_33_08_Pro_Left_Swipe_new;Left_Swipe_new;0\\n',\n",
       "       'WIN_20180925_17_51_17_Pro_Thumbs_Up_new;Thumbs_Up_new;4\\n',\n",
       "       'WIN_20180926_17_17_35_Pro_Left_Swipe_new;Left_Swipe_new;0\\n'],\n",
       "      dtype='<U88')"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_doc[:5]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "21212db3",
   "metadata": {},
   "source": [
    "Plot one image in actual dimension"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "980e689f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f90a4d2ea90>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "import matplotlib.pyplot as plt\n",
    "t = np.random.permutation(train_doc)\n",
    "train_path = '/datasets/Project_data/train'\n",
    "imgs = os.listdir(train_path+'/'+ t[0].split(';')[0])\n",
    "image = imread(train_path+'/'+ t[0].strip().split(';')[0]+'/'+imgs[1]).astype(np.float32)\n",
    "plt.imshow(image.astype(\"uint8\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "07264239",
   "metadata": {},
   "source": [
    "Compressed version of same image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "879494fa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(120, 160, 3)\n",
      "(100, 100, 3)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f909ec35700>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "h,w,_ = image.shape\n",
    "print(image.shape)\n",
    "image3 = resize(image[int(h/2)-50:int(h/2)+50,int(w/2)-50:int(w/2)+50,:],(100,100)).astype(np.float32)\n",
    "print(image3.shape)\n",
    "plt.imshow(image3.astype(\"uint8\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "11d4e321",
   "metadata": {},
   "source": [
    "# Generator"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0b05f15c",
   "metadata": {},
   "source": [
    " This is one of the most important part of the code. The overall structure of the generator has been given. In the generator, you are going to preprocess the images as you have images of 2 different dimensions as well as create a batch of video frames. You have to experiment with img_idx, y,z and normalization such that you get high accuracy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "c7839c48",
   "metadata": {},
   "outputs": [],
   "source": [
    "## Generator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "215ec289",
   "metadata": {},
   "outputs": [],
   "source": [
    "def generator(source_path, folder_list, batch_size):\n",
    "    print( 'Source path = ', source_path, '; batch size =', batch_size)\n",
    "    img_idx = [0,1,2,4,6,8,10,12,14,16,18,20,22,24,26,27,28,29]\n",
    "    while True:\n",
    "        t = np.random.permutation(folder_list)\n",
    "        num_batches = int(len(t)/batch_size)\n",
    "        for batch in range(num_batches):\n",
    "            batch_data = np.zeros((batch_size,18,100,100,3))\n",
    "            batch_labels = np.zeros((batch_size,5))\n",
    "            for folder in range(batch_size):\n",
    "                imgs = os.listdir(source_path+'/'+ t[folder + (batch*batch_size)].split(';')[0])\n",
    "                for idx,item in enumerate(img_idx):\n",
    "                    image = imread(source_path+'/'+ t[folder + (batch*batch_size)].strip().split(';')[0]+'/'+imgs[item]).astype(np.float32)\n",
    "                    if image.shape[1] == 160:\n",
    "                        image = resize(image[int(h/2)-50:int(h/2)+50,int(w/2)-50:int(w/2)+50,:],(100,100)).astype(np.float32)\n",
    "                    else:\n",
    "                        image = resize(image,(100,100)).astype(np.float32)\n",
    "                    \n",
    "                    batch_data[folder,idx,:,:,0] = (image[:,:,0] - np.min(image[:,:,0]))/(np.max(image[:,:,0])-np.min(image[:,:,0]))\n",
    "                    batch_data[folder,idx,:,:,1] = (image[:,:,1] - np.min(image[:,:,1]))/(np.max(image[:,:,1])-np.min(image[:,:,1]))\n",
    "                    batch_data[folder,idx,:,:,2] = (image[:,:,2] - np.min(image[:,:,2]))/(np.max(image[:,:,2])-np.min(image[:,:,2]))\n",
    "                    \n",
    "                batch_labels[folder, int(t[folder + (batch*batch_size)].strip().split(';')[2])] = 1\n",
    "            yield batch_data, batch_labels\n",
    "\n",
    "        if (len(t)%batch_size) != 0:\n",
    "            batch_data = np.zeros((len(t)%batch_size,18,100,100,3))\n",
    "            batch_labels = np.zeros((len(t)%batch_size,5))\n",
    "            for folder in range(len(t)%batch_size):\n",
    "                imgs = os.listdir(source_path+'/'+ t[folder + (num_batches*batch_size)].split(';')[0])\n",
    "                for idx,item in enumerate(img_idx):\n",
    "                    image = imread(source_path+'/'+ t[folder + (num_batches*batch_size)].strip().split(';')[0]+'/'+imgs[item]).astype(np.float32)\n",
    "                    if image.shape[1] == 160:\n",
    "                        image = resize(image[int(h/2)-50:int(h/2)+50,int(w/2)-50:int(w/2)+50,:],(100,100)).astype(np.float32)\n",
    "                    else:\n",
    "                        image = resize(image,(100,100)).astype(np.float32)\n",
    "\n",
    "                    batch_data[folder,idx,:,:,0] = (image[:,:,0] - np.min(image[:,:,0]))/(np.max(image[:,:,0])-np.min(image[:,:,0]))\n",
    "                    batch_data[folder,idx,:,:,1] = (image[:,:,1] - np.min(image[:,:,1]))/(np.max(image[:,:,1])-np.min(image[:,:,1]))\n",
    "                    batch_data[folder,idx,:,:,2] = (image[:,:,2] - np.min(image[:,:,2]))/(np.max(image[:,:,2])-np.min(image[:,:,2]))\n",
    "\n",
    "                batch_labels[folder, int(t[folder + (num_batches*batch_size)].strip().split(';')[2])] = 1\n",
    "\n",
    "            yield batch_data, batch_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "62de3146",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# training sequences = 663\n",
      "# validation sequences = 100\n",
      "# epochs = 5\n"
     ]
    }
   ],
   "source": [
    "curr_dt_time = datetime.datetime.now()\n",
    "train_path = '/datasets/Project_data/train'\n",
    "val_path = '/datasets/Project_data/val'\n",
    "num_train_sequences = len(train_doc)\n",
    "print('# training sequences =', num_train_sequences)\n",
    "num_val_sequences = len(val_doc)\n",
    "print('# validation sequences =', num_val_sequences)\n",
    "num_epochs = 5 # choose the number of epochs\n",
    "print ('# epochs =', num_epochs)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0cf56a4b",
   "metadata": {},
   "source": [
    "# Model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f0cbe9dc",
   "metadata": {},
   "source": [
    "Here you make the model using different functionalities that Keras provides. Remember to use `Conv3D` and `MaxPooling3D` and not `Conv2D` and `Maxpooling2D` for a 3D convolution model. You would want to use `TimeDistributed` while building a Conv2D + RNN model. Also remember that the last layer is the softmax. Design the network in such a way that the model is able to give good accuracy on the least number of parameters so that it can fit in the memory of the webcam."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6c01ab36",
   "metadata": {},
   "source": [
    "# Experiment 1: CNN + RNN based architecture with 5 epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "00bb7120",
   "metadata": {},
   "outputs": [],
   "source": [
    "## Experiment 1 - CNN + RNN based architecture with 5 epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "f4e49b65",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-09-07 03:30:46.036796: W tensorflow/core/common_runtime/gpu/gpu_bfc_allocator.cc:39] Overriding allow_growth setting because the TF_FORCE_GPU_ALLOW_GROWTH environment variable is set. Original config value was 0.\n",
      "2023-09-07 03:30:46.036860: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1525] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 14796 MB memory:  -> device: 0, name: Quadro RTX 5000, pci bus id: 0000:1d:00.0, compute capability: 7.5\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential, Model\n",
    "from keras.layers import Dense, GRU, Dropout, Flatten, TimeDistributed\n",
    "from keras.callbacks import ModelCheckpoint, ReduceLROnPlateau\n",
    "from tensorflow.keras import optimizers\n",
    "from keras.applications.vgg16 import VGG16\n",
    "    \n",
    "base_model = VGG16(include_top=False, weights='imagenet', input_shape=(100,100,3))\n",
    "x = base_model.output\n",
    "x = Flatten()(x)\n",
    "#x.add(Dropout(0.5))\n",
    "features = Dense(64, activation='relu')(x)\n",
    "conv_model = Model(inputs=base_model.input, outputs=features)\n",
    "    \n",
    "for layer in base_model.layers:\n",
    "    layer.trainable = False\n",
    "        \n",
    "model = Sequential()\n",
    "model.add(TimeDistributed(conv_model, input_shape=(18,100,100,3)))\n",
    "model.add(GRU(32, return_sequences=True))\n",
    "model.add(GRU(16))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(8, activation='relu'))\n",
    "model.add(Dense(5, activation='softmax'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "064ff3d1",
   "metadata": {},
   "source": [
    "Now that you have written the model, the next step is to compile the model. When you print the summary of the model, you'll see the total number of parameters you have to train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "5005572e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " time_distributed (TimeDistr  (None, 18, 64)           15009664  \n",
      " ibuted)                                                         \n",
      "                                                                 \n",
      " gru (GRU)                   (None, 18, 32)            9408      \n",
      "                                                                 \n",
      " gru_1 (GRU)                 (None, 16)                2400      \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 16)                0         \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 8)                 136       \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 5)                 45        \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 15,021,653\n",
      "Trainable params: 306,965\n",
      "Non-trainable params: 14,714,688\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.8/dist-packages/keras/optimizer_v2/gradient_descent.py:102: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n",
      "  super(SGD, self).__init__(name, **kwargs)\n"
     ]
    }
   ],
   "source": [
    "\n",
    "sgd = optimizers.SGD(lr=0.001, decay=1e-6, momentum=0.7, nesterov=True)\n",
    "model.compile(optimizer=sgd, loss='categorical_crossentropy', metrics=['categorical_accuracy'])\n",
    "print (model.summary())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "15aadf2e",
   "metadata": {},
   "source": [
    "Let us create the `train_generator` and the `val_generator` which will be used in `.fit_generator`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "89a60e5a",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_generator = generator(train_path, train_doc, batch_size)\n",
    "val_generator = generator(val_path, val_doc, batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "3d022a4c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:`period` argument is deprecated. Please use `save_freq` to specify the frequency in number of batches seen.\n",
      "WARNING:tensorflow:`epsilon` argument is deprecated and will be removed, use `min_delta` instead.\n"
     ]
    }
   ],
   "source": [
    "model_name = 'model_init' + '_' + str(curr_dt_time).replace(' ','').replace(':','_') + '/'\n",
    "\n",
    "if not os.path.exists(model_name):\n",
    "    os.mkdir(model_name)\n",
    "\n",
    "filepath = model_name + 'model-{epoch:05d}-{loss:.5f}-{categorical_accuracy:.5f}-{val_loss:.5f}-{val_categorical_accuracy:.5f}.h5'\n",
    "\n",
    "checkpoint = ModelCheckpoint(filepath, monitor='val_loss', verbose=1, save_best_only=False, save_weights_only=False, mode='auto', period=1)\n",
    "\n",
    "LR = ReduceLROnPlateau(monitor='val_loss', factor=0.5, patience=2, verbose=1, mode='min', epsilon=0.0001, cooldown=0, min_lr=0.00001)# write the REducelronplateau code here\n",
    "callbacks_list = [checkpoint, LR]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3bee7ef2",
   "metadata": {},
   "source": [
    "The steps_per_epoch and validation_steps are used by fit_generator to decide the number of next() calls it need to make."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "7bbb7e0a",
   "metadata": {},
   "outputs": [],
   "source": [
    "if (num_train_sequences%batch_size) == 0:\n",
    "    steps_per_epoch = int(num_train_sequences/batch_size)\n",
    "else:\n",
    "    steps_per_epoch = (num_train_sequences//batch_size) + 1\n",
    "\n",
    "if (num_val_sequences%batch_size) == 0:\n",
    "    validation_steps = int(num_val_sequences/batch_size)\n",
    "\n",
    "else:\n",
    "    validation_steps = (num_val_sequences//batch_size) + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "f830a5f7",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_332/2741796141.py:1: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
      "  model.fit_generator(train_generator, steps_per_epoch=steps_per_epoch, epochs=num_epochs, verbose=1,\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Source path =  /datasets/Project_data/train ; batch size = 20\n",
      "Epoch 1/5\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-09-07 03:31:05.871252: I tensorflow/stream_executor/cuda/cuda_dnn.cc:377] Loaded cuDNN version 8302\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "34/34 [==============================] - ETA: 0s - loss: 1.6425 - categorical_accuracy: 0.2021Source path =  /datasets/Project_data/val ; batch size = 20\n",
      "\n",
      "Epoch 00001: saving model to model_init_2023-09-0703_30_43.005262/model-00001-1.64252-0.20211-1.61456-0.15000.h5\n",
      "34/34 [==============================] - 36s 852ms/step - loss: 1.6425 - categorical_accuracy: 0.2021 - val_loss: 1.6146 - val_categorical_accuracy: 0.1500 - lr: 0.0010\n",
      "Epoch 2/5\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.6195 - categorical_accuracy: 0.2021\n",
      "Epoch 00002: saving model to model_init_2023-09-0703_30_43.005262/model-00002-1.61954-0.20211-1.60485-0.25000.h5\n",
      "34/34 [==============================] - 28s 842ms/step - loss: 1.6195 - categorical_accuracy: 0.2021 - val_loss: 1.6048 - val_categorical_accuracy: 0.2500 - lr: 0.0010\n",
      "Epoch 3/5\n",
      "33/34 [============================>.] - ETA: 0s - loss: 1.6069 - categorical_accuracy: 0.2303\n",
      "Epoch 00003: saving model to model_init_2023-09-0703_30_43.005262/model-00003-1.60682-0.22926-1.59500-0.25000.h5\n",
      "34/34 [==============================] - 27s 812ms/step - loss: 1.6068 - categorical_accuracy: 0.2293 - val_loss: 1.5950 - val_categorical_accuracy: 0.2500 - lr: 0.0010\n",
      "Epoch 4/5\n",
      "33/34 [============================>.] - ETA: 0s - loss: 1.6025 - categorical_accuracy: 0.2273\n",
      "Epoch 00004: saving model to model_init_2023-09-0703_30_43.005262/model-00004-1.60247-0.22775-1.57097-0.20000.h5\n",
      "34/34 [==============================] - 28s 837ms/step - loss: 1.6025 - categorical_accuracy: 0.2278 - val_loss: 1.5710 - val_categorical_accuracy: 0.2000 - lr: 0.0010\n",
      "Epoch 5/5\n",
      "33/34 [============================>.] - ETA: 0s - loss: 1.5871 - categorical_accuracy: 0.2515\n",
      "Epoch 00005: saving model to model_init_2023-09-0703_30_43.005262/model-00005-1.58746-0.25038-1.57589-0.31000.h5\n",
      "34/34 [==============================] - 27s 828ms/step - loss: 1.5875 - categorical_accuracy: 0.2504 - val_loss: 1.5759 - val_categorical_accuracy: 0.3100 - lr: 0.0010\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f909cb1eb20>"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit_generator(train_generator, steps_per_epoch=steps_per_epoch, epochs=num_epochs, verbose=1,\n",
    "                    callbacks=callbacks_list, validation_data=val_generator,\n",
    "                    validation_steps=validation_steps, class_weight=None, workers=1, initial_epoch=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "82681825",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "1f5e41fc",
   "metadata": {},
   "source": [
    "Let us now fit the model. This will start training the model and with the help of the checkpoints, you'll be able to save the model at the end of each epoch."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b8d83a84",
   "metadata": {},
   "source": [
    "# Experiment 2: CNN + RNN with Trainable VGGnet layers for 10 epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "32b22416",
   "metadata": {},
   "outputs": [],
   "source": [
    "## Experiment 2 - CNN + RNN with Trainable VGGnet layers for 10 epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "689eb4d5",
   "metadata": {},
   "outputs": [],
   "source": [
    "split_at = 15\n",
    "for layer in base_model.layers[:split_at]:\n",
    "    layer.trainable= False\n",
    "for layer in base_model.layers[split_at:]:\n",
    "    layer.trainable=True\n",
    "\n",
    "model = Sequential()\n",
    "model.add(TimeDistributed(conv_model, input_shape=(18,100,100,3)))\n",
    "model.add(GRU(32, return_sequences=True))\n",
    "model.add(GRU(16))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(8, activation='relu'))\n",
    "model.add(Dense(5, activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "26f6ffec",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " time_distributed_1 (TimeDis  (None, 18, 64)           15009664  \n",
      " tributed)                                                       \n",
      "                                                                 \n",
      " gru_2 (GRU)                 (None, 18, 32)            9408      \n",
      "                                                                 \n",
      " gru_3 (GRU)                 (None, 16)                2400      \n",
      "                                                                 \n",
      " dropout_1 (Dropout)         (None, 16)                0         \n",
      "                                                                 \n",
      " dense_3 (Dense)             (None, 8)                 136       \n",
      "                                                                 \n",
      " dense_4 (Dense)             (None, 5)                 45        \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 15,021,653\n",
      "Trainable params: 7,386,389\n",
      "Non-trainable params: 7,635,264\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "sgd = optimizers.SGD(lr=0.001, decay=1e-6, momentum=0.7, nesterov=True)\n",
    "model.compile(optimizer=sgd, loss='categorical_crossentropy', metrics=['categorical_accuracy'])\n",
    "print (model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "a37ae23c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# epochs = 10\n"
     ]
    }
   ],
   "source": [
    "num_epochs = 10 # choose the number of epochs\n",
    "print ('# epochs =', num_epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "0c34e095",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:`period` argument is deprecated. Please use `save_freq` to specify the frequency in number of batches seen.\n",
      "WARNING:tensorflow:`epsilon` argument is deprecated and will be removed, use `min_delta` instead.\n"
     ]
    }
   ],
   "source": [
    "model_name = 'model_init' + '_' + str(curr_dt_time).replace(' ','').replace(':','_') + '/'\n",
    "\n",
    "if not os.path.exists(model_name):\n",
    "    os.mkdir(model_name)\n",
    "\n",
    "filepath = model_name + 'model-{epoch:05d}-{loss:.5f}-{categorical_accuracy:.5f}-{val_loss:.5f}-{val_categorical_accuracy:.5f}.h5'\n",
    "\n",
    "checkpoint = ModelCheckpoint(filepath, monitor='val_loss', verbose=1, save_best_only=False, save_weights_only=False, mode='auto', period=1)\n",
    "\n",
    "LR = ReduceLROnPlateau(monitor='val_loss', factor=0.5, patience=2, verbose=1, mode='min', epsilon=0.0001, cooldown=0, min_lr=0.00001)# write the REducelronplateau code here\n",
    "callbacks_list = [checkpoint, LR]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "b190c591",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_332/2741796141.py:1: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
      "  model.fit_generator(train_generator, steps_per_epoch=steps_per_epoch, epochs=num_epochs, verbose=1,\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.6490 - categorical_accuracy: 0.2081\n",
      "Epoch 00001: saving model to model_init_2023-09-0703_30_43.005262/model-00001-1.64898-0.20814-1.58070-0.26000.h5\n",
      "34/34 [==============================] - 32s 840ms/step - loss: 1.6490 - categorical_accuracy: 0.2081 - val_loss: 1.5807 - val_categorical_accuracy: 0.2600 - lr: 0.0010\n",
      "Epoch 2/10\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.5583 - categorical_accuracy: 0.2700\n",
      "Epoch 00002: saving model to model_init_2023-09-0703_30_43.005262/model-00002-1.55829-0.26998-1.53626-0.39000.h5\n",
      "34/34 [==============================] - 27s 822ms/step - loss: 1.5583 - categorical_accuracy: 0.2700 - val_loss: 1.5363 - val_categorical_accuracy: 0.3900 - lr: 0.0010\n",
      "Epoch 3/10\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.4781 - categorical_accuracy: 0.3409\n",
      "Epoch 00003: saving model to model_init_2023-09-0703_30_43.005262/model-00003-1.47813-0.34087-1.41489-0.48000.h5\n",
      "34/34 [==============================] - 28s 836ms/step - loss: 1.4781 - categorical_accuracy: 0.3409 - val_loss: 1.4149 - val_categorical_accuracy: 0.4800 - lr: 0.0010\n",
      "Epoch 4/10\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.3726 - categorical_accuracy: 0.3831\n",
      "Epoch 00004: saving model to model_init_2023-09-0703_30_43.005262/model-00004-1.37263-0.38311-1.39104-0.44000.h5\n",
      "34/34 [==============================] - 27s 810ms/step - loss: 1.3726 - categorical_accuracy: 0.3831 - val_loss: 1.3910 - val_categorical_accuracy: 0.4400 - lr: 0.0010\n",
      "Epoch 5/10\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.2769 - categorical_accuracy: 0.4465\n",
      "Epoch 00005: saving model to model_init_2023-09-0703_30_43.005262/model-00005-1.27693-0.44646-1.26283-0.45000.h5\n",
      "34/34 [==============================] - 28s 838ms/step - loss: 1.2769 - categorical_accuracy: 0.4465 - val_loss: 1.2628 - val_categorical_accuracy: 0.4500 - lr: 0.0010\n",
      "Epoch 6/10\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.1990 - categorical_accuracy: 0.4796\n",
      "Epoch 00006: saving model to model_init_2023-09-0703_30_43.005262/model-00006-1.19904-0.47964-1.18219-0.56000.h5\n",
      "34/34 [==============================] - 27s 826ms/step - loss: 1.1990 - categorical_accuracy: 0.4796 - val_loss: 1.1822 - val_categorical_accuracy: 0.5600 - lr: 0.0010\n",
      "Epoch 7/10\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.1324 - categorical_accuracy: 0.5098\n",
      "Epoch 00007: saving model to model_init_2023-09-0703_30_43.005262/model-00007-1.13244-0.50980-1.18358-0.50000.h5\n",
      "34/34 [==============================] - 28s 836ms/step - loss: 1.1324 - categorical_accuracy: 0.5098 - val_loss: 1.1836 - val_categorical_accuracy: 0.5000 - lr: 0.0010\n",
      "Epoch 8/10\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.0984 - categorical_accuracy: 0.5701\n",
      "Epoch 00008: saving model to model_init_2023-09-0703_30_43.005262/model-00008-1.09843-0.57014-1.12706-0.48000.h5\n",
      "34/34 [==============================] - 27s 824ms/step - loss: 1.0984 - categorical_accuracy: 0.5701 - val_loss: 1.1271 - val_categorical_accuracy: 0.4800 - lr: 0.0010\n",
      "Epoch 9/10\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.0692 - categorical_accuracy: 0.5535\n",
      "Epoch 00009: saving model to model_init_2023-09-0703_30_43.005262/model-00009-1.06921-0.55354-1.00232-0.64000.h5\n",
      "34/34 [==============================] - 27s 824ms/step - loss: 1.0692 - categorical_accuracy: 0.5535 - val_loss: 1.0023 - val_categorical_accuracy: 0.6400 - lr: 0.0010\n",
      "Epoch 10/10\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.9921 - categorical_accuracy: 0.5732\n",
      "Epoch 00010: saving model to model_init_2023-09-0703_30_43.005262/model-00010-0.99209-0.57315-1.01207-0.61000.h5\n",
      "34/34 [==============================] - 27s 819ms/step - loss: 0.9921 - categorical_accuracy: 0.5732 - val_loss: 1.0121 - val_categorical_accuracy: 0.6100 - lr: 0.0010\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f909c53bdc0>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit_generator(train_generator, steps_per_epoch=steps_per_epoch, epochs=num_epochs, verbose=1,\n",
    "                    callbacks=callbacks_list, validation_data=val_generator,\n",
    "                    validation_steps=validation_steps, class_weight=None, workers=1, initial_epoch=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "99046d5e",
   "metadata": {},
   "source": [
    "# Experiment 3 - CNN + RNN with Trainable VGGnet layers for 20 epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "24e1952b",
   "metadata": {},
   "outputs": [],
   "source": [
    "## Experiment 3 - CNN + RNN with Trainable VGGnet layers for 20 epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "17a1d00e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# epochs = 20\n"
     ]
    }
   ],
   "source": [
    "num_epochs = 20 # choose the number of epochs\n",
    "print ('# epochs =', num_epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "ee002c6f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:`period` argument is deprecated. Please use `save_freq` to specify the frequency in number of batches seen.\n",
      "WARNING:tensorflow:`epsilon` argument is deprecated and will be removed, use `min_delta` instead.\n"
     ]
    }
   ],
   "source": [
    "model_name = 'model_init' + '_' + str(curr_dt_time).replace(' ','').replace(':','_') + '/'\n",
    "    \n",
    "if not os.path.exists(model_name):\n",
    "    os.mkdir(model_name)\n",
    "        \n",
    "filepath = model_name + 'model-{epoch:05d}-{loss:.5f}-{categorical_accuracy:.5f}-{val_loss:.5f}-{val_categorical_accuracy:.5f}.h5'\n",
    "\n",
    "checkpoint = ModelCheckpoint(filepath, monitor='val_loss', verbose=1, save_best_only=False, save_weights_only=False, mode='auto', period=1)\n",
    "\n",
    "LR = ReduceLROnPlateau(monitor='val_loss', factor=0.5, patience=2, verbose=1, mode='min', epsilon=0.0001, cooldown=0, min_lr=0.00001)# write the REducelronplateau code here\n",
    "callbacks_list = [checkpoint, LR]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "93e77b52",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_332/385748685.py:1: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
      "  model.fit_generator(train_generator, steps_per_epoch=steps_per_epoch, epochs=num_epochs, verbose=1,\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.9582 - categorical_accuracy: 0.6169\n",
      "Epoch 00001: saving model to model_init_2023-09-0703_30_43.005262/model-00001-0.95823-0.61689-0.92433-0.67000.h5\n",
      "34/34 [==============================] - 29s 862ms/step - loss: 0.9582 - categorical_accuracy: 0.6169 - val_loss: 0.9243 - val_categorical_accuracy: 0.6700 - lr: 0.0010\n",
      "Epoch 2/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.9139 - categorical_accuracy: 0.6425\n",
      "Epoch 00002: saving model to model_init_2023-09-0703_30_43.005262/model-00002-0.91390-0.64253-0.86018-0.67000.h5\n",
      "34/34 [==============================] - 28s 848ms/step - loss: 0.9139 - categorical_accuracy: 0.6425 - val_loss: 0.8602 - val_categorical_accuracy: 0.6700 - lr: 0.0010\n",
      "Epoch 3/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.8410 - categorical_accuracy: 0.6757\n",
      "Epoch 00003: saving model to model_init_2023-09-0703_30_43.005262/model-00003-0.84097-0.67572-0.89325-0.61000.h5\n",
      "34/34 [==============================] - 28s 848ms/step - loss: 0.8410 - categorical_accuracy: 0.6757 - val_loss: 0.8932 - val_categorical_accuracy: 0.6100 - lr: 0.0010\n",
      "Epoch 4/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.8191 - categorical_accuracy: 0.6968\n",
      "Epoch 00004: saving model to model_init_2023-09-0703_30_43.005262/model-00004-0.81914-0.69683-0.80687-0.66000.h5\n",
      "34/34 [==============================] - 27s 810ms/step - loss: 0.8191 - categorical_accuracy: 0.6968 - val_loss: 0.8069 - val_categorical_accuracy: 0.6600 - lr: 0.0010\n",
      "Epoch 5/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.7791 - categorical_accuracy: 0.7179\n",
      "Epoch 00005: saving model to model_init_2023-09-0703_30_43.005262/model-00005-0.77915-0.71795-0.84490-0.80000.h5\n",
      "34/34 [==============================] - 28s 834ms/step - loss: 0.7791 - categorical_accuracy: 0.7179 - val_loss: 0.8449 - val_categorical_accuracy: 0.8000 - lr: 0.0010\n",
      "Epoch 6/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.7351 - categorical_accuracy: 0.7677\n",
      "Epoch 00006: saving model to model_init_2023-09-0703_30_43.005262/model-00006-0.73512-0.76772-0.77544-0.82000.h5\n",
      "34/34 [==============================] - 27s 824ms/step - loss: 0.7351 - categorical_accuracy: 0.7677 - val_loss: 0.7754 - val_categorical_accuracy: 0.8200 - lr: 0.0010\n",
      "Epoch 7/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.7173 - categorical_accuracy: 0.8039\n",
      "Epoch 00007: saving model to model_init_2023-09-0703_30_43.005262/model-00007-0.71727-0.80392-0.72051-0.82000.h5\n",
      "34/34 [==============================] - 27s 810ms/step - loss: 0.7173 - categorical_accuracy: 0.8039 - val_loss: 0.7205 - val_categorical_accuracy: 0.8200 - lr: 0.0010\n",
      "Epoch 8/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.6791 - categorical_accuracy: 0.8341\n",
      "Epoch 00008: saving model to model_init_2023-09-0703_30_43.005262/model-00008-0.67915-0.83409-0.79618-0.84000.h5\n",
      "34/34 [==============================] - 28s 836ms/step - loss: 0.6791 - categorical_accuracy: 0.8341 - val_loss: 0.7962 - val_categorical_accuracy: 0.8400 - lr: 0.0010\n",
      "Epoch 9/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.6610 - categorical_accuracy: 0.8567\n",
      "Epoch 00009: saving model to model_init_2023-09-0703_30_43.005262/model-00009-0.66099-0.85671-0.67676-0.88000.h5\n",
      "34/34 [==============================] - 28s 829ms/step - loss: 0.6610 - categorical_accuracy: 0.8567 - val_loss: 0.6768 - val_categorical_accuracy: 0.8800 - lr: 0.0010\n",
      "Epoch 10/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.6148 - categorical_accuracy: 0.8627\n",
      "Epoch 00010: saving model to model_init_2023-09-0703_30_43.005262/model-00010-0.61478-0.86275-0.69187-0.87000.h5\n",
      "34/34 [==============================] - 27s 814ms/step - loss: 0.6148 - categorical_accuracy: 0.8627 - val_loss: 0.6919 - val_categorical_accuracy: 0.8700 - lr: 0.0010\n",
      "Epoch 11/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5787 - categorical_accuracy: 0.9065\n",
      "Epoch 00011: saving model to model_init_2023-09-0703_30_43.005262/model-00011-0.57869-0.90649-0.65601-0.84000.h5\n",
      "34/34 [==============================] - 27s 825ms/step - loss: 0.5787 - categorical_accuracy: 0.9065 - val_loss: 0.6560 - val_categorical_accuracy: 0.8400 - lr: 0.0010\n",
      "Epoch 12/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5614 - categorical_accuracy: 0.8959\n",
      "Epoch 00012: saving model to model_init_2023-09-0703_30_43.005262/model-00012-0.56137-0.89593-0.58255-0.89000.h5\n",
      "34/34 [==============================] - 28s 835ms/step - loss: 0.5614 - categorical_accuracy: 0.8959 - val_loss: 0.5826 - val_categorical_accuracy: 0.8900 - lr: 0.0010\n",
      "Epoch 13/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5408 - categorical_accuracy: 0.9065\n",
      "Epoch 00013: saving model to model_init_2023-09-0703_30_43.005262/model-00013-0.54078-0.90649-0.65718-0.80000.h5\n",
      "34/34 [==============================] - 27s 814ms/step - loss: 0.5408 - categorical_accuracy: 0.9065 - val_loss: 0.6572 - val_categorical_accuracy: 0.8000 - lr: 0.0010\n",
      "Epoch 14/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5159 - categorical_accuracy: 0.9095\n",
      "Epoch 00014: saving model to model_init_2023-09-0703_30_43.005262/model-00014-0.51586-0.90950-0.58961-0.87000.h5\n",
      "\n",
      "Epoch 00014: ReduceLROnPlateau reducing learning rate to 0.0005000000237487257.\n",
      "34/34 [==============================] - 27s 809ms/step - loss: 0.5159 - categorical_accuracy: 0.9095 - val_loss: 0.5896 - val_categorical_accuracy: 0.8700 - lr: 0.0010\n",
      "Epoch 15/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.4672 - categorical_accuracy: 0.9397\n",
      "Epoch 00015: saving model to model_init_2023-09-0703_30_43.005262/model-00015-0.46716-0.93967-0.57246-0.86000.h5\n",
      "34/34 [==============================] - 27s 824ms/step - loss: 0.4672 - categorical_accuracy: 0.9397 - val_loss: 0.5725 - val_categorical_accuracy: 0.8600 - lr: 5.0000e-04\n",
      "Epoch 16/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.4714 - categorical_accuracy: 0.9276\n",
      "Epoch 00016: saving model to model_init_2023-09-0703_30_43.005262/model-00016-0.47141-0.92760-0.57210-0.86000.h5\n",
      "34/34 [==============================] - 27s 806ms/step - loss: 0.4714 - categorical_accuracy: 0.9276 - val_loss: 0.5721 - val_categorical_accuracy: 0.8600 - lr: 5.0000e-04\n",
      "Epoch 17/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.4676 - categorical_accuracy: 0.9140\n",
      "Epoch 00017: saving model to model_init_2023-09-0703_30_43.005262/model-00017-0.46758-0.91403-0.59083-0.84000.h5\n",
      "34/34 [==============================] - 28s 857ms/step - loss: 0.4676 - categorical_accuracy: 0.9140 - val_loss: 0.5908 - val_categorical_accuracy: 0.8400 - lr: 5.0000e-04\n",
      "Epoch 18/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.4447 - categorical_accuracy: 0.9382\n",
      "Epoch 00018: saving model to model_init_2023-09-0703_30_43.005262/model-00018-0.44472-0.93816-0.56862-0.84000.h5\n",
      "34/34 [==============================] - 28s 836ms/step - loss: 0.4447 - categorical_accuracy: 0.9382 - val_loss: 0.5686 - val_categorical_accuracy: 0.8400 - lr: 5.0000e-04\n",
      "Epoch 19/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.4340 - categorical_accuracy: 0.9427\n",
      "Epoch 00019: saving model to model_init_2023-09-0703_30_43.005262/model-00019-0.43403-0.94268-0.53922-0.86000.h5\n",
      "34/34 [==============================] - 28s 845ms/step - loss: 0.4340 - categorical_accuracy: 0.9427 - val_loss: 0.5392 - val_categorical_accuracy: 0.8600 - lr: 5.0000e-04\n",
      "Epoch 20/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.4239 - categorical_accuracy: 0.9502\n",
      "Epoch 00020: saving model to model_init_2023-09-0703_30_43.005262/model-00020-0.42394-0.95023-0.54457-0.85000.h5\n",
      "34/34 [==============================] - 27s 823ms/step - loss: 0.4239 - categorical_accuracy: 0.9502 - val_loss: 0.5446 - val_categorical_accuracy: 0.8500 - lr: 5.0000e-04\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f9089fc2e80>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit_generator(train_generator, steps_per_epoch=steps_per_epoch, epochs=num_epochs, verbose=1, \n",
    "                    callbacks=callbacks_list, validation_data=val_generator, \n",
    "                    validation_steps=validation_steps, class_weight=None, workers=1, initial_epoch=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eea1ff02",
   "metadata": {},
   "source": [
    "# Experiment 4 -CNN + RNN without trainable VGGnet layers with additional dropout for 20 epoch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "69f12852",
   "metadata": {},
   "outputs": [],
   "source": [
    "## Experiment 4 -CNN + RNN without trainable VGGnet layers with additional dropout for 20 epoch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "1b67ecba",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential, Model\n",
    "from keras.layers import Dense, GRU, Dropout, Flatten, TimeDistributed\n",
    "from keras.callbacks import ModelCheckpoint, ReduceLROnPlateau\n",
    "from tensorflow.keras import optimizers\n",
    "from keras.applications.vgg16 import VGG16\n",
    "    \n",
    "base_model = VGG16(include_top=False, weights='imagenet', input_shape=(100,100,3))\n",
    "x = base_model.output\n",
    "x = Flatten()(x)\n",
    "x = Dropout(0.5)(x)\n",
    "features = Dense(64, activation='relu')(x)\n",
    "conv_model = Model(inputs=base_model.input, outputs=features)\n",
    "    \n",
    "for layer in base_model.layers:\n",
    "    layer.trainable = False\n",
    "        \n",
    "model = Sequential()\n",
    "model.add(TimeDistributed(conv_model, input_shape=(18,100,100,3)))\n",
    "model.add(GRU(32, return_sequences=True))\n",
    "model.add(GRU(16))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(8, activation='relu'))\n",
    "model.add(Dense(5, activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "9cbc7832",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " time_distributed_2 (TimeDis  (None, 18, 64)           15009664  \n",
      " tributed)                                                       \n",
      "                                                                 \n",
      " gru_4 (GRU)                 (None, 18, 32)            9408      \n",
      "                                                                 \n",
      " gru_5 (GRU)                 (None, 16)                2400      \n",
      "                                                                 \n",
      " dropout_3 (Dropout)         (None, 16)                0         \n",
      "                                                                 \n",
      " dense_6 (Dense)             (None, 8)                 136       \n",
      "                                                                 \n",
      " dense_7 (Dense)             (None, 5)                 45        \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 15,021,653\n",
      "Trainable params: 306,965\n",
      "Non-trainable params: 14,714,688\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "sgd = optimizers.SGD(lr=0.001, decay=1e-6, momentum=0.7, nesterov=True)\n",
    "model.compile(optimizer=sgd, loss='categorical_crossentropy', metrics=['categorical_accuracy'])\n",
    "print (model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "1ddff1f9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# epochs = 20\n"
     ]
    }
   ],
   "source": [
    "num_epochs = 20 # choose the number of epochs\n",
    "print ('# epochs =', num_epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "d7065f36",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:`period` argument is deprecated. Please use `save_freq` to specify the frequency in number of batches seen.\n",
      "WARNING:tensorflow:`epsilon` argument is deprecated and will be removed, use `min_delta` instead.\n"
     ]
    }
   ],
   "source": [
    "model_name = 'model_init' + '_' + str(curr_dt_time).replace(' ','').replace(':','_') + '/'\n",
    "    \n",
    "if not os.path.exists(model_name):\n",
    "    os.mkdir(model_name)\n",
    "        \n",
    "filepath = model_name + 'model-{epoch:05d}-{loss:.5f}-{categorical_accuracy:.5f}-{val_loss:.5f}-{val_categorical_accuracy:.5f}.h5'\n",
    "\n",
    "checkpoint = ModelCheckpoint(filepath, monitor='val_loss', verbose=1, save_best_only=False, save_weights_only=False, mode='auto', period=1)\n",
    "\n",
    "LR = ReduceLROnPlateau(monitor='val_loss', factor=0.5, patience=2, verbose=1, mode='min', epsilon=0.0001, cooldown=0, min_lr=0.00001)# write the REducelronplateau code here\n",
    "callbacks_list = [checkpoint, LR]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "ab539072",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_332/385748685.py:1: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
      "  model.fit_generator(train_generator, steps_per_epoch=steps_per_epoch, epochs=num_epochs, verbose=1,\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.6326 - categorical_accuracy: 0.2036\n",
      "Epoch 00001: saving model to model_init_2023-09-0703_30_43.005262/model-00001-1.63262-0.20362-1.58231-0.31000.h5\n",
      "34/34 [==============================] - 31s 850ms/step - loss: 1.6326 - categorical_accuracy: 0.2036 - val_loss: 1.5823 - val_categorical_accuracy: 0.3100 - lr: 0.0010\n",
      "Epoch 2/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.6213 - categorical_accuracy: 0.2232\n",
      "Epoch 00002: saving model to model_init_2023-09-0703_30_43.005262/model-00002-1.62131-0.22323-1.56488-0.32000.h5\n",
      "34/34 [==============================] - 27s 808ms/step - loss: 1.6213 - categorical_accuracy: 0.2232 - val_loss: 1.5649 - val_categorical_accuracy: 0.3200 - lr: 0.0010\n",
      "Epoch 3/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.5902 - categorical_accuracy: 0.2308\n",
      "Epoch 00003: saving model to model_init_2023-09-0703_30_43.005262/model-00003-1.59018-0.23077-1.56824-0.27000.h5\n",
      "34/34 [==============================] - 28s 839ms/step - loss: 1.5902 - categorical_accuracy: 0.2308 - val_loss: 1.5682 - val_categorical_accuracy: 0.2700 - lr: 0.0010\n",
      "Epoch 4/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.5648 - categorical_accuracy: 0.2836\n",
      "Epoch 00004: saving model to model_init_2023-09-0703_30_43.005262/model-00004-1.56485-0.28356-1.52009-0.38000.h5\n",
      "34/34 [==============================] - 27s 810ms/step - loss: 1.5648 - categorical_accuracy: 0.2836 - val_loss: 1.5201 - val_categorical_accuracy: 0.3800 - lr: 0.0010\n",
      "Epoch 5/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.5521 - categorical_accuracy: 0.3002\n",
      "Epoch 00005: saving model to model_init_2023-09-0703_30_43.005262/model-00005-1.55212-0.30015-1.54089-0.32000.h5\n",
      "34/34 [==============================] - 28s 843ms/step - loss: 1.5521 - categorical_accuracy: 0.3002 - val_loss: 1.5409 - val_categorical_accuracy: 0.3200 - lr: 0.0010\n",
      "Epoch 6/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.5392 - categorical_accuracy: 0.3243\n",
      "Epoch 00006: saving model to model_init_2023-09-0703_30_43.005262/model-00006-1.53917-0.32428-1.48763-0.38000.h5\n",
      "34/34 [==============================] - 27s 819ms/step - loss: 1.5392 - categorical_accuracy: 0.3243 - val_loss: 1.4876 - val_categorical_accuracy: 0.3800 - lr: 0.0010\n",
      "Epoch 7/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.5145 - categorical_accuracy: 0.3529\n",
      "Epoch 00007: saving model to model_init_2023-09-0703_30_43.005262/model-00007-1.51453-0.35294-1.48063-0.38000.h5\n",
      "34/34 [==============================] - 28s 835ms/step - loss: 1.5145 - categorical_accuracy: 0.3529 - val_loss: 1.4806 - val_categorical_accuracy: 0.3800 - lr: 0.0010\n",
      "Epoch 8/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.5177 - categorical_accuracy: 0.3258\n",
      "Epoch 00008: saving model to model_init_2023-09-0703_30_43.005262/model-00008-1.51769-0.32579-1.42669-0.47000.h5\n",
      "34/34 [==============================] - 27s 825ms/step - loss: 1.5177 - categorical_accuracy: 0.3258 - val_loss: 1.4267 - val_categorical_accuracy: 0.4700 - lr: 0.0010\n",
      "Epoch 9/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.4973 - categorical_accuracy: 0.3560\n",
      "Epoch 00009: saving model to model_init_2023-09-0703_30_43.005262/model-00009-1.49733-0.35596-1.47581-0.38000.h5\n",
      "34/34 [==============================] - 27s 824ms/step - loss: 1.4973 - categorical_accuracy: 0.3560 - val_loss: 1.4758 - val_categorical_accuracy: 0.3800 - lr: 0.0010\n",
      "Epoch 10/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.4688 - categorical_accuracy: 0.3876\n",
      "Epoch 00010: saving model to model_init_2023-09-0703_30_43.005262/model-00010-1.46883-0.38763-1.40261-0.44000.h5\n",
      "34/34 [==============================] - 28s 839ms/step - loss: 1.4688 - categorical_accuracy: 0.3876 - val_loss: 1.4026 - val_categorical_accuracy: 0.4400 - lr: 0.0010\n",
      "Epoch 11/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.4681 - categorical_accuracy: 0.3967\n",
      "Epoch 00011: saving model to model_init_2023-09-0703_30_43.005262/model-00011-1.46814-0.39668-1.40225-0.44000.h5\n",
      "34/34 [==============================] - 27s 813ms/step - loss: 1.4681 - categorical_accuracy: 0.3967 - val_loss: 1.4022 - val_categorical_accuracy: 0.4400 - lr: 0.0010\n",
      "Epoch 12/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.4520 - categorical_accuracy: 0.3605\n",
      "Epoch 00012: saving model to model_init_2023-09-0703_30_43.005262/model-00012-1.45203-0.36048-1.42772-0.37000.h5\n",
      "34/34 [==============================] - 27s 817ms/step - loss: 1.4520 - categorical_accuracy: 0.3605 - val_loss: 1.4277 - val_categorical_accuracy: 0.3700 - lr: 0.0010\n",
      "Epoch 13/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.4260 - categorical_accuracy: 0.3952\n",
      "Epoch 00013: saving model to model_init_2023-09-0703_30_43.005262/model-00013-1.42597-0.39517-1.35662-0.48000.h5\n",
      "34/34 [==============================] - 27s 825ms/step - loss: 1.4260 - categorical_accuracy: 0.3952 - val_loss: 1.3566 - val_categorical_accuracy: 0.4800 - lr: 0.0010\n",
      "Epoch 14/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.4221 - categorical_accuracy: 0.3997\n",
      "Epoch 00014: saving model to model_init_2023-09-0703_30_43.005262/model-00014-1.42212-0.39970-1.36531-0.44000.h5\n",
      "34/34 [==============================] - 28s 830ms/step - loss: 1.4221 - categorical_accuracy: 0.3997 - val_loss: 1.3653 - val_categorical_accuracy: 0.4400 - lr: 0.0010\n",
      "Epoch 15/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.3992 - categorical_accuracy: 0.4118\n",
      "Epoch 00015: saving model to model_init_2023-09-0703_30_43.005262/model-00015-1.39921-0.41176-1.40619-0.38000.h5\n",
      "\n",
      "Epoch 00015: ReduceLROnPlateau reducing learning rate to 0.0005000000237487257.\n",
      "34/34 [==============================] - 27s 809ms/step - loss: 1.3992 - categorical_accuracy: 0.4118 - val_loss: 1.4062 - val_categorical_accuracy: 0.3800 - lr: 0.0010\n",
      "Epoch 16/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.3862 - categorical_accuracy: 0.4178\n",
      "Epoch 00016: saving model to model_init_2023-09-0703_30_43.005262/model-00016-1.38622-0.41780-1.33325-0.46000.h5\n",
      "34/34 [==============================] - 28s 840ms/step - loss: 1.3862 - categorical_accuracy: 0.4178 - val_loss: 1.3332 - val_categorical_accuracy: 0.4600 - lr: 5.0000e-04\n",
      "Epoch 17/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.3827 - categorical_accuracy: 0.4329\n",
      "Epoch 00017: saving model to model_init_2023-09-0703_30_43.005262/model-00017-1.38272-0.43288-1.35179-0.45000.h5\n",
      "34/34 [==============================] - 27s 826ms/step - loss: 1.3827 - categorical_accuracy: 0.4329 - val_loss: 1.3518 - val_categorical_accuracy: 0.4500 - lr: 5.0000e-04\n",
      "Epoch 18/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.3754 - categorical_accuracy: 0.4208\n",
      "Epoch 00018: saving model to model_init_2023-09-0703_30_43.005262/model-00018-1.37541-0.42081-1.33610-0.46000.h5\n",
      "\n",
      "Epoch 00018: ReduceLROnPlateau reducing learning rate to 0.0002500000118743628.\n",
      "34/34 [==============================] - 27s 820ms/step - loss: 1.3754 - categorical_accuracy: 0.4208 - val_loss: 1.3361 - val_categorical_accuracy: 0.4600 - lr: 5.0000e-04\n",
      "Epoch 19/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.3773 - categorical_accuracy: 0.4284\n",
      "Epoch 00019: saving model to model_init_2023-09-0703_30_43.005262/model-00019-1.37732-0.42836-1.27643-0.52000.h5\n",
      "34/34 [==============================] - 28s 829ms/step - loss: 1.3773 - categorical_accuracy: 0.4284 - val_loss: 1.2764 - val_categorical_accuracy: 0.5200 - lr: 2.5000e-04\n",
      "Epoch 20/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.3448 - categorical_accuracy: 0.4525\n",
      "Epoch 00020: saving model to model_init_2023-09-0703_30_43.005262/model-00020-1.34480-0.45249-1.35038-0.44000.h5\n",
      "34/34 [==============================] - 27s 822ms/step - loss: 1.3448 - categorical_accuracy: 0.4525 - val_loss: 1.3504 - val_categorical_accuracy: 0.4400 - lr: 2.5000e-04\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f909c068910>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit_generator(train_generator, steps_per_epoch=steps_per_epoch, epochs=num_epochs, verbose=1, \n",
    "                    callbacks=callbacks_list, validation_data=val_generator, \n",
    "                    validation_steps=validation_steps, class_weight=None, workers=1, initial_epoch=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1d8d1c0d",
   "metadata": {},
   "source": [
    "# Experiment 5: CNN + RNN without trainable VGGnet layers for 20 epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "3bbe7749",
   "metadata": {},
   "outputs": [],
   "source": [
    "## Experiment 5: CNN + RNN without trainable VGGnet layers for 20 epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "40115a75",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential, Model\n",
    "from keras.layers import Dense, GRU, Dropout, Flatten, TimeDistributed\n",
    "from keras.callbacks import ModelCheckpoint, ReduceLROnPlateau\n",
    "from tensorflow.keras import optimizers\n",
    "from keras.applications.vgg16 import VGG16\n",
    "    \n",
    "base_model = VGG16(include_top=False, weights='imagenet', input_shape=(100,100,3))\n",
    "x = base_model.output\n",
    "x = Flatten()(x)\n",
    "#x.add(Dropout(0.5))\n",
    "features = Dense(64, activation='relu')(x)\n",
    "conv_model = Model(inputs=base_model.input, outputs=features)\n",
    "    \n",
    "for layer in base_model.layers:\n",
    "    layer.trainable = False\n",
    "        \n",
    "model = Sequential()\n",
    "model.add(TimeDistributed(conv_model, input_shape=(18,100,100,3)))\n",
    "model.add(GRU(32, return_sequences=True))\n",
    "model.add(GRU(16))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(8, activation='relu'))\n",
    "model.add(Dense(5, activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "e2e9b35f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_3\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " time_distributed_3 (TimeDis  (None, 18, 64)           15009664  \n",
      " tributed)                                                       \n",
      "                                                                 \n",
      " gru_6 (GRU)                 (None, 18, 32)            9408      \n",
      "                                                                 \n",
      " gru_7 (GRU)                 (None, 16)                2400      \n",
      "                                                                 \n",
      " dropout_4 (Dropout)         (None, 16)                0         \n",
      "                                                                 \n",
      " dense_9 (Dense)             (None, 8)                 136       \n",
      "                                                                 \n",
      " dense_10 (Dense)            (None, 5)                 45        \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 15,021,653\n",
      "Trainable params: 306,965\n",
      "Non-trainable params: 14,714,688\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "sgd = optimizers.SGD(lr=0.001, decay=1e-6, momentum=0.7, nesterov=True)\n",
    "model.compile(optimizer=sgd, loss='categorical_crossentropy', metrics=['categorical_accuracy'])\n",
    "print (model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "2a2cb643",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# epochs = 20\n"
     ]
    }
   ],
   "source": [
    "num_epochs = 20 # choose the number of epochs\n",
    "print ('# epochs =', num_epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "9a98d8cb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:`period` argument is deprecated. Please use `save_freq` to specify the frequency in number of batches seen.\n",
      "WARNING:tensorflow:`epsilon` argument is deprecated and will be removed, use `min_delta` instead.\n"
     ]
    }
   ],
   "source": [
    "model_name = 'model_init' + '_' + str(curr_dt_time).replace(' ','').replace(':','_') + '/'\n",
    "    \n",
    "if not os.path.exists(model_name):\n",
    "    os.mkdir(model_name)\n",
    "        \n",
    "filepath = model_name + 'model-{epoch:05d}-{loss:.5f}-{categorical_accuracy:.5f}-{val_loss:.5f}-{val_categorical_accuracy:.5f}.h5'\n",
    "\n",
    "checkpoint = ModelCheckpoint(filepath, monitor='val_loss', verbose=1, save_best_only=False, save_weights_only=False, mode='auto', period=1)\n",
    "\n",
    "LR = ReduceLROnPlateau(monitor='val_loss', factor=0.5, patience=2, verbose=1, mode='min', epsilon=0.0001, cooldown=0, min_lr=0.00001)# write the REducelronplateau code here\n",
    "callbacks_list = [checkpoint, LR]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "a28c6db6",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_332/385748685.py:1: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
      "  model.fit_generator(train_generator, steps_per_epoch=steps_per_epoch, epochs=num_epochs, verbose=1,\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.6410 - categorical_accuracy: 0.1931\n",
      "Epoch 00001: saving model to model_init_2023-09-0703_30_43.005262/model-00001-1.64103-0.19306-1.59567-0.24000.h5\n",
      "34/34 [==============================] - 30s 830ms/step - loss: 1.6410 - categorical_accuracy: 0.1931 - val_loss: 1.5957 - val_categorical_accuracy: 0.2400 - lr: 0.0010\n",
      "Epoch 2/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.6062 - categorical_accuracy: 0.2338\n",
      "Epoch 00002: saving model to model_init_2023-09-0703_30_43.005262/model-00002-1.60621-0.23379-1.60026-0.21000.h5\n",
      "34/34 [==============================] - 28s 831ms/step - loss: 1.6062 - categorical_accuracy: 0.2338 - val_loss: 1.6003 - val_categorical_accuracy: 0.2100 - lr: 0.0010\n",
      "Epoch 3/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.5889 - categorical_accuracy: 0.2700\n",
      "Epoch 00003: saving model to model_init_2023-09-0703_30_43.005262/model-00003-1.58893-0.26998-1.58612-0.29000.h5\n",
      "34/34 [==============================] - 27s 828ms/step - loss: 1.5889 - categorical_accuracy: 0.2700 - val_loss: 1.5861 - val_categorical_accuracy: 0.2900 - lr: 0.0010\n",
      "Epoch 4/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.5764 - categorical_accuracy: 0.2760\n",
      "Epoch 00004: saving model to model_init_2023-09-0703_30_43.005262/model-00004-1.57643-0.27602-1.54189-0.34000.h5\n",
      "34/34 [==============================] - 27s 817ms/step - loss: 1.5764 - categorical_accuracy: 0.2760 - val_loss: 1.5419 - val_categorical_accuracy: 0.3400 - lr: 0.0010\n",
      "Epoch 5/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.5571 - categorical_accuracy: 0.2896\n",
      "Epoch 00005: saving model to model_init_2023-09-0703_30_43.005262/model-00005-1.55706-0.28959-1.52018-0.36000.h5\n",
      "34/34 [==============================] - 28s 833ms/step - loss: 1.5571 - categorical_accuracy: 0.2896 - val_loss: 1.5202 - val_categorical_accuracy: 0.3600 - lr: 0.0010\n",
      "Epoch 6/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.5448 - categorical_accuracy: 0.3077\n",
      "Epoch 00006: saving model to model_init_2023-09-0703_30_43.005262/model-00006-1.54476-0.30769-1.51338-0.41000.h5\n",
      "34/34 [==============================] - 27s 812ms/step - loss: 1.5448 - categorical_accuracy: 0.3077 - val_loss: 1.5134 - val_categorical_accuracy: 0.4100 - lr: 0.0010\n",
      "Epoch 7/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.5264 - categorical_accuracy: 0.3394\n",
      "Epoch 00007: saving model to model_init_2023-09-0703_30_43.005262/model-00007-1.52637-0.33937-1.48182-0.46000.h5\n",
      "34/34 [==============================] - 27s 827ms/step - loss: 1.5264 - categorical_accuracy: 0.3394 - val_loss: 1.4818 - val_categorical_accuracy: 0.4600 - lr: 0.0010\n",
      "Epoch 8/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.5180 - categorical_accuracy: 0.3409\n",
      "Epoch 00008: saving model to model_init_2023-09-0703_30_43.005262/model-00008-1.51801-0.34087-1.49127-0.41000.h5\n",
      "34/34 [==============================] - 28s 851ms/step - loss: 1.5180 - categorical_accuracy: 0.3409 - val_loss: 1.4913 - val_categorical_accuracy: 0.4100 - lr: 0.0010\n",
      "Epoch 9/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.4967 - categorical_accuracy: 0.3801\n",
      "Epoch 00009: saving model to model_init_2023-09-0703_30_43.005262/model-00009-1.49675-0.38009-1.44471-0.45000.h5\n",
      "34/34 [==============================] - 28s 837ms/step - loss: 1.4967 - categorical_accuracy: 0.3801 - val_loss: 1.4447 - val_categorical_accuracy: 0.4500 - lr: 0.0010\n",
      "Epoch 10/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.4805 - categorical_accuracy: 0.3590\n",
      "Epoch 00010: saving model to model_init_2023-09-0703_30_43.005262/model-00010-1.48051-0.35897-1.47805-0.39000.h5\n",
      "34/34 [==============================] - 27s 804ms/step - loss: 1.4805 - categorical_accuracy: 0.3590 - val_loss: 1.4781 - val_categorical_accuracy: 0.3900 - lr: 0.0010\n",
      "Epoch 11/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.4675 - categorical_accuracy: 0.4027\n",
      "Epoch 00011: saving model to model_init_2023-09-0703_30_43.005262/model-00011-1.46752-0.40271-1.41703-0.49000.h5\n",
      "34/34 [==============================] - 28s 831ms/step - loss: 1.4675 - categorical_accuracy: 0.4027 - val_loss: 1.4170 - val_categorical_accuracy: 0.4900 - lr: 0.0010\n",
      "Epoch 12/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.4396 - categorical_accuracy: 0.4193\n",
      "Epoch 00012: saving model to model_init_2023-09-0703_30_43.005262/model-00012-1.43958-0.41931-1.43760-0.48000.h5\n",
      "34/34 [==============================] - 27s 827ms/step - loss: 1.4396 - categorical_accuracy: 0.4193 - val_loss: 1.4376 - val_categorical_accuracy: 0.4800 - lr: 0.0010\n",
      "Epoch 13/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.3918 - categorical_accuracy: 0.4691\n",
      "Epoch 00013: saving model to model_init_2023-09-0703_30_43.005262/model-00013-1.39181-0.46908-1.37893-0.51000.h5\n",
      "34/34 [==============================] - 28s 829ms/step - loss: 1.3918 - categorical_accuracy: 0.4691 - val_loss: 1.3789 - val_categorical_accuracy: 0.5100 - lr: 0.0010\n",
      "Epoch 14/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.3922 - categorical_accuracy: 0.4495\n",
      "Epoch 00014: saving model to model_init_2023-09-0703_30_43.005262/model-00014-1.39221-0.44947-1.32692-0.62000.h5\n",
      "34/34 [==============================] - 27s 822ms/step - loss: 1.3922 - categorical_accuracy: 0.4495 - val_loss: 1.3269 - val_categorical_accuracy: 0.6200 - lr: 0.0010\n",
      "Epoch 15/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.3797 - categorical_accuracy: 0.4510\n",
      "Epoch 00015: saving model to model_init_2023-09-0703_30_43.005262/model-00015-1.37972-0.45098-1.37000-0.46000.h5\n",
      "34/34 [==============================] - 29s 874ms/step - loss: 1.3797 - categorical_accuracy: 0.4510 - val_loss: 1.3700 - val_categorical_accuracy: 0.4600 - lr: 0.0010\n",
      "Epoch 16/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.3609 - categorical_accuracy: 0.4736\n",
      "Epoch 00016: saving model to model_init_2023-09-0703_30_43.005262/model-00016-1.36090-0.47360-1.31038-0.56000.h5\n",
      "34/34 [==============================] - 28s 843ms/step - loss: 1.3609 - categorical_accuracy: 0.4736 - val_loss: 1.3104 - val_categorical_accuracy: 0.5600 - lr: 0.0010\n",
      "Epoch 17/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.3279 - categorical_accuracy: 0.5083\n",
      "Epoch 00017: saving model to model_init_2023-09-0703_30_43.005262/model-00017-1.32792-0.50830-1.31158-0.56000.h5\n",
      "34/34 [==============================] - 27s 819ms/step - loss: 1.3279 - categorical_accuracy: 0.5083 - val_loss: 1.3116 - val_categorical_accuracy: 0.5600 - lr: 0.0010\n",
      "Epoch 18/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.3057 - categorical_accuracy: 0.5023\n",
      "Epoch 00018: saving model to model_init_2023-09-0703_30_43.005262/model-00018-1.30575-0.50226-1.25557-0.60000.h5\n",
      "34/34 [==============================] - 28s 830ms/step - loss: 1.3057 - categorical_accuracy: 0.5023 - val_loss: 1.2556 - val_categorical_accuracy: 0.6000 - lr: 0.0010\n",
      "Epoch 19/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.2916 - categorical_accuracy: 0.5249\n",
      "Epoch 00019: saving model to model_init_2023-09-0703_30_43.005262/model-00019-1.29157-0.52489-1.25309-0.53000.h5\n",
      "34/34 [==============================] - 27s 822ms/step - loss: 1.2916 - categorical_accuracy: 0.5249 - val_loss: 1.2531 - val_categorical_accuracy: 0.5300 - lr: 0.0010\n",
      "Epoch 20/20\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.2455 - categorical_accuracy: 0.5385\n",
      "Epoch 00020: saving model to model_init_2023-09-0703_30_43.005262/model-00020-1.24550-0.53846-1.26903-0.59000.h5\n",
      "34/34 [==============================] - 27s 822ms/step - loss: 1.2455 - categorical_accuracy: 0.5385 - val_loss: 1.2690 - val_categorical_accuracy: 0.5900 - lr: 0.0010\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f9088bc43a0>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit_generator(train_generator, steps_per_epoch=steps_per_epoch, epochs=num_epochs, verbose=1, \n",
    "                    callbacks=callbacks_list, validation_data=val_generator, \n",
    "                    validation_steps=validation_steps, class_weight=None, workers=1, initial_epoch=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "82bf1803",
   "metadata": {},
   "source": [
    "# Experiment 6: Conv3D architecture with 25 epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "63ed2aa1",
   "metadata": {},
   "outputs": [],
   "source": [
    "## Experiment 6: Conv3D architecture with 25 epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "12b9bc01",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, GRU, Dropout, Flatten, BatchNormalization, Activation\n",
    "from keras.layers.convolutional import Conv3D, MaxPooling3D\n",
    "\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Conv3D(64, (3,3,3), strides=(1,1,1), padding='same', input_shape=(18,100,100,3)))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('elu'))\n",
    "model.add(MaxPooling3D(pool_size=(2,2,1), strides=(2,2,1)))\n",
    "\n",
    "model.add(Conv3D(128, (3,3,3), strides=(1,1,1), padding='same'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('elu'))\n",
    "model.add(MaxPooling3D(pool_size=(2,2,2), strides=(2,2,2)))\n",
    "\n",
    "# model.add(Dropout(0.25))\n",
    "\n",
    "model.add(Conv3D(256, (3,3,3), strides=(1,1,1), padding='same'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('elu'))\n",
    "model.add(MaxPooling3D(pool_size=(2,2,2), strides=(2,2,2)))\n",
    "\n",
    "# model.add(Dropout(0.25))\n",
    "\n",
    "model.add(Conv3D(256, (3,3,3), strides=(1,1,1), padding='same'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Activation('elu'))\n",
    "model.add(MaxPooling3D(pool_size=(2,2,2), strides=(2,2,2)))\n",
    "\n",
    "model.add(Flatten())\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(512, activation='elu'))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(5, activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "09750cf6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_4\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv3d (Conv3D)             (None, 18, 100, 100, 64)  5248      \n",
      "                                                                 \n",
      " batch_normalization (BatchN  (None, 18, 100, 100, 64)  256      \n",
      " ormalization)                                                   \n",
      "                                                                 \n",
      " activation (Activation)     (None, 18, 100, 100, 64)  0         \n",
      "                                                                 \n",
      " max_pooling3d (MaxPooling3D  (None, 9, 50, 100, 64)   0         \n",
      " )                                                               \n",
      "                                                                 \n",
      " conv3d_1 (Conv3D)           (None, 9, 50, 100, 128)   221312    \n",
      "                                                                 \n",
      " batch_normalization_1 (Batc  (None, 9, 50, 100, 128)  512       \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " activation_1 (Activation)   (None, 9, 50, 100, 128)   0         \n",
      "                                                                 \n",
      " max_pooling3d_1 (MaxPooling  (None, 4, 25, 50, 128)   0         \n",
      " 3D)                                                             \n",
      "                                                                 \n",
      " conv3d_2 (Conv3D)           (None, 4, 25, 50, 256)    884992    \n",
      "                                                                 \n",
      " batch_normalization_2 (Batc  (None, 4, 25, 50, 256)   1024      \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " activation_2 (Activation)   (None, 4, 25, 50, 256)    0         \n",
      "                                                                 \n",
      " max_pooling3d_2 (MaxPooling  (None, 2, 12, 25, 256)   0         \n",
      " 3D)                                                             \n",
      "                                                                 \n",
      " conv3d_3 (Conv3D)           (None, 2, 12, 25, 256)    1769728   \n",
      "                                                                 \n",
      " batch_normalization_3 (Batc  (None, 2, 12, 25, 256)   1024      \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " activation_3 (Activation)   (None, 2, 12, 25, 256)    0         \n",
      "                                                                 \n",
      " max_pooling3d_3 (MaxPooling  (None, 1, 6, 12, 256)    0         \n",
      " 3D)                                                             \n",
      "                                                                 \n",
      " flatten_3 (Flatten)         (None, 18432)             0         \n",
      "                                                                 \n",
      " dropout_5 (Dropout)         (None, 18432)             0         \n",
      "                                                                 \n",
      " dense_11 (Dense)            (None, 512)               9437696   \n",
      "                                                                 \n",
      " dropout_6 (Dropout)         (None, 512)               0         \n",
      "                                                                 \n",
      " dense_12 (Dense)            (None, 5)                 2565      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 12,324,357\n",
      "Trainable params: 12,322,949\n",
      "Non-trainable params: 1,408\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "sgd = optimizers.SGD(lr=0.001, decay=1e-6, momentum=0.7, nesterov=True)\n",
    "model.compile(optimizer=sgd, loss='categorical_crossentropy', metrics=['categorical_accuracy'])\n",
    "print (model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "0b7f9a05",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# epochs = 25\n"
     ]
    }
   ],
   "source": [
    "num_epochs = 25 # choose the number of epochs\n",
    "print ('# epochs =', num_epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "2b710a59",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:`period` argument is deprecated. Please use `save_freq` to specify the frequency in number of batches seen.\n",
      "WARNING:tensorflow:`epsilon` argument is deprecated and will be removed, use `min_delta` instead.\n"
     ]
    }
   ],
   "source": [
    "model_name = 'model_init' + '_' + str(curr_dt_time).replace(' ','').replace(':','_') + '/'\n",
    "    \n",
    "if not os.path.exists(model_name):\n",
    "    os.mkdir(model_name)\n",
    "        \n",
    "filepath = model_name + 'model-{epoch:05d}-{loss:.5f}-{categorical_accuracy:.5f}-{val_loss:.5f}-{val_categorical_accuracy:.5f}.h5'\n",
    "\n",
    "checkpoint = ModelCheckpoint(filepath, monitor='val_loss', verbose=1, save_best_only=False, save_weights_only=False, mode='auto', period=1)\n",
    "\n",
    "LR = ReduceLROnPlateau(monitor='val_loss', factor=0.5, patience=2, verbose=1, mode='min', epsilon=0.0001, cooldown=0, min_lr=0.00001)# write the REducelronplateau code here\n",
    "callbacks_list = [checkpoint, LR]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "92b264b4",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_332/385748685.py:1: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
      "  model.fit_generator(train_generator, steps_per_epoch=steps_per_epoch, epochs=num_epochs, verbose=1,\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 3.2418 - categorical_accuracy: 0.3529\n",
      "Epoch 00001: saving model to model_init_2023-09-0703_30_43.005262/model-00001-3.24182-0.35294-2.41320-0.28000.h5\n",
      "34/34 [==============================] - 34s 919ms/step - loss: 3.2418 - categorical_accuracy: 0.3529 - val_loss: 2.4132 - val_categorical_accuracy: 0.2800 - lr: 0.0010\n",
      "Epoch 2/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.6019 - categorical_accuracy: 0.5339\n",
      "Epoch 00002: saving model to model_init_2023-09-0703_30_43.005262/model-00002-1.60187-0.53394-5.36249-0.26000.h5\n",
      "34/34 [==============================] - 28s 836ms/step - loss: 1.6019 - categorical_accuracy: 0.5339 - val_loss: 5.3625 - val_categorical_accuracy: 0.2600 - lr: 0.0010\n",
      "Epoch 3/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.2860 - categorical_accuracy: 0.5928\n",
      "Epoch 00003: saving model to model_init_2023-09-0703_30_43.005262/model-00003-1.28600-0.59276-7.02718-0.13000.h5\n",
      "\n",
      "Epoch 00003: ReduceLROnPlateau reducing learning rate to 0.0005000000237487257.\n",
      "34/34 [==============================] - 28s 834ms/step - loss: 1.2860 - categorical_accuracy: 0.5928 - val_loss: 7.0272 - val_categorical_accuracy: 0.1300 - lr: 0.0010\n",
      "Epoch 4/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 1.0489 - categorical_accuracy: 0.6576\n",
      "Epoch 00004: saving model to model_init_2023-09-0703_30_43.005262/model-00004-1.04889-0.65762-7.79261-0.14000.h5\n",
      "34/34 [==============================] - 28s 822ms/step - loss: 1.0489 - categorical_accuracy: 0.6576 - val_loss: 7.7926 - val_categorical_accuracy: 0.1400 - lr: 5.0000e-04\n",
      "Epoch 5/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.9722 - categorical_accuracy: 0.6682\n",
      "Epoch 00005: saving model to model_init_2023-09-0703_30_43.005262/model-00005-0.97221-0.66817-7.36874-0.20000.h5\n",
      "\n",
      "Epoch 00005: ReduceLROnPlateau reducing learning rate to 0.0002500000118743628.\n",
      "34/34 [==============================] - 28s 823ms/step - loss: 0.9722 - categorical_accuracy: 0.6682 - val_loss: 7.3687 - val_categorical_accuracy: 0.2000 - lr: 5.0000e-04\n",
      "Epoch 6/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.7807 - categorical_accuracy: 0.7285\n",
      "Epoch 00006: saving model to model_init_2023-09-0703_30_43.005262/model-00006-0.78067-0.72851-6.87998-0.31000.h5\n",
      "34/34 [==============================] - 28s 832ms/step - loss: 0.7807 - categorical_accuracy: 0.7285 - val_loss: 6.8800 - val_categorical_accuracy: 0.3100 - lr: 2.5000e-04\n",
      "Epoch 7/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.6709 - categorical_accuracy: 0.7632\n",
      "Epoch 00007: saving model to model_init_2023-09-0703_30_43.005262/model-00007-0.67088-0.76320-7.15384-0.23000.h5\n",
      "\n",
      "Epoch 00007: ReduceLROnPlateau reducing learning rate to 0.0001250000059371814.\n",
      "34/34 [==============================] - 28s 829ms/step - loss: 0.6709 - categorical_accuracy: 0.7632 - val_loss: 7.1538 - val_categorical_accuracy: 0.2300 - lr: 2.5000e-04\n",
      "Epoch 8/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5951 - categorical_accuracy: 0.7617\n",
      "Epoch 00008: saving model to model_init_2023-09-0703_30_43.005262/model-00008-0.59508-0.76169-7.26835-0.23000.h5\n",
      "34/34 [==============================] - 27s 809ms/step - loss: 0.5951 - categorical_accuracy: 0.7617 - val_loss: 7.2683 - val_categorical_accuracy: 0.2300 - lr: 1.2500e-04\n",
      "Epoch 9/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5535 - categorical_accuracy: 0.7964\n",
      "Epoch 00009: saving model to model_init_2023-09-0703_30_43.005262/model-00009-0.55345-0.79638-6.64700-0.27000.h5\n",
      "\n",
      "Epoch 00009: ReduceLROnPlateau reducing learning rate to 6.25000029685907e-05.\n",
      "34/34 [==============================] - 27s 819ms/step - loss: 0.5535 - categorical_accuracy: 0.7964 - val_loss: 6.6470 - val_categorical_accuracy: 0.2700 - lr: 1.2500e-04\n",
      "Epoch 10/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5718 - categorical_accuracy: 0.7949\n",
      "Epoch 00010: saving model to model_init_2023-09-0703_30_43.005262/model-00010-0.57184-0.79487-5.69763-0.31000.h5\n",
      "34/34 [==============================] - 28s 831ms/step - loss: 0.5718 - categorical_accuracy: 0.7949 - val_loss: 5.6976 - val_categorical_accuracy: 0.3100 - lr: 6.2500e-05\n",
      "Epoch 11/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.6038 - categorical_accuracy: 0.7828\n",
      "Epoch 00011: saving model to model_init_2023-09-0703_30_43.005262/model-00011-0.60381-0.78281-5.11321-0.29000.h5\n",
      "\n",
      "Epoch 00011: ReduceLROnPlateau reducing learning rate to 3.125000148429535e-05.\n",
      "34/34 [==============================] - 28s 839ms/step - loss: 0.6038 - categorical_accuracy: 0.7828 - val_loss: 5.1132 - val_categorical_accuracy: 0.2900 - lr: 6.2500e-05\n",
      "Epoch 12/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5247 - categorical_accuracy: 0.8024\n",
      "Epoch 00012: saving model to model_init_2023-09-0703_30_43.005262/model-00012-0.52465-0.80241-4.21977-0.31000.h5\n",
      "34/34 [==============================] - 28s 836ms/step - loss: 0.5247 - categorical_accuracy: 0.8024 - val_loss: 4.2198 - val_categorical_accuracy: 0.3100 - lr: 3.1250e-05\n",
      "Epoch 13/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5832 - categorical_accuracy: 0.7828\n",
      "Epoch 00013: saving model to model_init_2023-09-0703_30_43.005262/model-00013-0.58317-0.78281-3.07333-0.30000.h5\n",
      "\n",
      "Epoch 00013: ReduceLROnPlateau reducing learning rate to 1.5625000742147677e-05.\n",
      "34/34 [==============================] - 28s 827ms/step - loss: 0.5832 - categorical_accuracy: 0.7828 - val_loss: 3.0733 - val_categorical_accuracy: 0.3000 - lr: 3.1250e-05\n",
      "Epoch 14/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5627 - categorical_accuracy: 0.7843\n",
      "Epoch 00014: saving model to model_init_2023-09-0703_30_43.005262/model-00014-0.56269-0.78431-2.20658-0.44000.h5\n",
      "34/34 [==============================] - 27s 811ms/step - loss: 0.5627 - categorical_accuracy: 0.7843 - val_loss: 2.2066 - val_categorical_accuracy: 0.4400 - lr: 1.5625e-05\n",
      "Epoch 15/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5706 - categorical_accuracy: 0.8054\n",
      "Epoch 00015: saving model to model_init_2023-09-0703_30_43.005262/model-00015-0.57059-0.80543-1.77355-0.51000.h5\n",
      "34/34 [==============================] - 28s 822ms/step - loss: 0.5706 - categorical_accuracy: 0.8054 - val_loss: 1.7735 - val_categorical_accuracy: 0.5100 - lr: 1.5625e-05\n",
      "Epoch 16/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5368 - categorical_accuracy: 0.7964\n",
      "Epoch 00016: saving model to model_init_2023-09-0703_30_43.005262/model-00016-0.53683-0.79638-1.28365-0.60000.h5\n",
      "34/34 [==============================] - 28s 825ms/step - loss: 0.5368 - categorical_accuracy: 0.7964 - val_loss: 1.2837 - val_categorical_accuracy: 0.6000 - lr: 1.5625e-05\n",
      "Epoch 17/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5177 - categorical_accuracy: 0.8084\n",
      "Epoch 00017: saving model to model_init_2023-09-0703_30_43.005262/model-00017-0.51771-0.80845-1.04876-0.72000.h5\n",
      "34/34 [==============================] - 27s 820ms/step - loss: 0.5177 - categorical_accuracy: 0.8084 - val_loss: 1.0488 - val_categorical_accuracy: 0.7200 - lr: 1.5625e-05\n",
      "Epoch 18/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5313 - categorical_accuracy: 0.8100\n",
      "Epoch 00018: saving model to model_init_2023-09-0703_30_43.005262/model-00018-0.53134-0.80995-0.82139-0.76000.h5\n",
      "34/34 [==============================] - 27s 813ms/step - loss: 0.5313 - categorical_accuracy: 0.8100 - val_loss: 0.8214 - val_categorical_accuracy: 0.7600 - lr: 1.5625e-05\n",
      "Epoch 19/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5131 - categorical_accuracy: 0.8115\n",
      "Epoch 00019: saving model to model_init_2023-09-0703_30_43.005262/model-00019-0.51308-0.81146-0.79912-0.78000.h5\n",
      "34/34 [==============================] - 28s 827ms/step - loss: 0.5131 - categorical_accuracy: 0.8115 - val_loss: 0.7991 - val_categorical_accuracy: 0.7800 - lr: 1.5625e-05\n",
      "Epoch 20/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5326 - categorical_accuracy: 0.8100\n",
      "Epoch 00020: saving model to model_init_2023-09-0703_30_43.005262/model-00020-0.53264-0.80995-0.71600-0.77000.h5\n",
      "34/34 [==============================] - 28s 847ms/step - loss: 0.5326 - categorical_accuracy: 0.8100 - val_loss: 0.7160 - val_categorical_accuracy: 0.7700 - lr: 1.5625e-05\n",
      "Epoch 21/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5113 - categorical_accuracy: 0.8175\n",
      "Epoch 00021: saving model to model_init_2023-09-0703_30_43.005262/model-00021-0.51129-0.81750-0.69340-0.77000.h5\n",
      "34/34 [==============================] - 28s 826ms/step - loss: 0.5113 - categorical_accuracy: 0.8175 - val_loss: 0.6934 - val_categorical_accuracy: 0.7700 - lr: 1.5625e-05\n",
      "Epoch 22/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5253 - categorical_accuracy: 0.8175\n",
      "Epoch 00022: saving model to model_init_2023-09-0703_30_43.005262/model-00022-0.52532-0.81750-0.53214-0.81000.h5\n",
      "34/34 [==============================] - 28s 827ms/step - loss: 0.5253 - categorical_accuracy: 0.8175 - val_loss: 0.5321 - val_categorical_accuracy: 0.8100 - lr: 1.5625e-05\n",
      "Epoch 23/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5586 - categorical_accuracy: 0.7949\n",
      "Epoch 00023: saving model to model_init_2023-09-0703_30_43.005262/model-00023-0.55862-0.79487-0.72529-0.78000.h5\n",
      "34/34 [==============================] - 27s 808ms/step - loss: 0.5586 - categorical_accuracy: 0.7949 - val_loss: 0.7253 - val_categorical_accuracy: 0.7800 - lr: 1.5625e-05\n",
      "Epoch 24/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.5266 - categorical_accuracy: 0.8115\n",
      "Epoch 00024: saving model to model_init_2023-09-0703_30_43.005262/model-00024-0.52657-0.81146-0.76087-0.75000.h5\n",
      "\n",
      "Epoch 00024: ReduceLROnPlateau reducing learning rate to 1e-05.\n",
      "34/34 [==============================] - 28s 838ms/step - loss: 0.5266 - categorical_accuracy: 0.8115 - val_loss: 0.7609 - val_categorical_accuracy: 0.7500 - lr: 1.5625e-05\n",
      "Epoch 25/25\n",
      "34/34 [==============================] - ETA: 0s - loss: 0.4995 - categorical_accuracy: 0.8265\n",
      "Epoch 00025: saving model to model_init_2023-09-0703_30_43.005262/model-00025-0.49948-0.82655-0.52212-0.82000.h5\n",
      "34/34 [==============================] - 28s 826ms/step - loss: 0.4995 - categorical_accuracy: 0.8265 - val_loss: 0.5221 - val_categorical_accuracy: 0.8200 - lr: 1.0000e-05\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f908814ee50>"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit_generator(train_generator, steps_per_epoch=steps_per_epoch, epochs=num_epochs, verbose=1, \n",
    "                    callbacks=callbacks_list, validation_data=val_generator, \n",
    "                    validation_steps=validation_steps, class_weight=None, workers=1, initial_epoch=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4ffe54a7",
   "metadata": {},
   "source": [
    "# After doing all the experiments, we finalized Experiment 3 - CNN + RNN with Trainable VGGnet layers for 20 epochs\n",
    "\n",
    "\n",
    "\n",
    "- (Accuracy: 0.8500)\n",
    "\n",
    "The best weights of CNN-LSTM:model-00020-0.42394-0.95023-0.54457-0.85000.h5 Experiment 3 - CNN + RNN with Trainable VGGnet layers for 20 epochs. we considered this weight for model testing."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}