sign-language-recognition.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "bd80fc40",
   "metadata": {},
   "source": [
    "**Problem Definition**\n",
    "\n",
    "Convert the Sign Alphabet to Text\n",
    "\n",
    "**Data**\n",
    "\n",
    "[Dataset](https://www.kaggle.com/datasets/datamunge/sign-language-mnist)\n",
    "\n",
    "1. No cases for the letters J & Z (Reason: J & Z require motion)\n",
    "2. GrayScale Images\n",
    "3. Pixel Values Range From 0 to 255\n",
    "4. Each image contains 784 Pixels\n",
    "\n",
    "\n",
    "**Evaluation**\n",
    "\n",
    "Accuracy\n",
    "\n",
    "**Modelling**\n",
    "\n",
    "CNN"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "83ec1c39",
   "metadata": {},
   "source": [
    "# Initial End-to-End Workflow (Start)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "15daf294",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importing required libraries\n",
    "\n",
    "from sklearn.preprocessing import LabelBinarizer\n",
    "from tensorflow import keras\n",
    "from keras.utils import plot_model\n",
    "\n",
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "dafa44b9",
   "metadata": {},
   "outputs": [],
   "source": [
    "# loading the training data (X+y)\n",
    "train_df = pd.read_csv('data/alphabet/sign_mnist_train.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "d7669304",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df = train_df.sample(frac=1, random_state=42) # Shuffiling the entire dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "8269e216",
   "metadata": {},
   "outputs": [],
   "source": [
    "X, y = train_df.drop('label', axis=1), train_df['label'] # Split the dataset into X, y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "8c2b3aa3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((27455, 784), (27455,))"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape, y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "da58cd49",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([dtype('int64')], dtype=object), dtype('int64'))"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.unique(X.dtypes), y.dtype"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "22b62d7a",
   "metadata": {},
   "outputs": [],
   "source": [
    "label_binarizer = LabelBinarizer() # Binarize labels in a one-vs-all fashion (return one-hot encoded vectors)\n",
    "y = label_binarizer.fit_transform(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "3cc4add2",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = X/255.0 # Normalizing the training data and converting the data type to float"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "d6700a24",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([dtype('float64')], dtype=object)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.unique(X.dtypes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "c9872eca",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Converting the 1-D array of 784 pixels to (28, 28, 1) Image\n",
    "# (28, 28) represents the spatial dimensions of the image & 1 specifies that the image is grayscale\n",
    "X = tf.reshape(X, [-1, 28, 28, 1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c6a7f19f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(TensorShape([27455, 28, 28, 1]), (27455, 24))"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape, y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "f950e21c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Generating a validation set\n",
    "\n",
    "X_train, X_valid = X[:25000], X[25000:]\n",
    "y_train, y_valid = y[:25000], y[25000:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "a51e4a80",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tf.float64"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[0].dtype"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "f567ca9b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([28, 28, 1])"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "d2533d22",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(<matplotlib.image.AxesImage at 0x2bda1b1cdc0>,\n",
       " array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "        0, 1]))"
      ]
     },
     "execution_count": 15,
     "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": [
    "plt.imshow(X[0], cmap='gray'), y[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "09046338",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Defining the CNN\n",
    "\n",
    "model = keras.models.Sequential()\n",
    "model.add(keras.layers.Conv2D(32, (5, 5), padding='same', activation='relu', input_shape=(28, 28, 1)))\n",
    "model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(keras.layers.Conv2D(64, (5, 5), padding='same', activation='relu'))\n",
    "model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(keras.layers.Conv2D(128, (5, 5), padding='same', activation='relu'))\n",
    "model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(keras.layers.Flatten())\n",
    "model.add(keras.layers.Dense(128, activation='relu'))\n",
    "model.add(keras.layers.Dense(24, activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "333da112",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d (Conv2D)             (None, 28, 28, 32)        832       \n",
      "                                                                 \n",
      " max_pooling2d (MaxPooling2D  (None, 14, 14, 32)       0         \n",
      " )                                                               \n",
      "                                                                 \n",
      " conv2d_1 (Conv2D)           (None, 14, 14, 64)        51264     \n",
      "                                                                 \n",
      " max_pooling2d_1 (MaxPooling  (None, 7, 7, 64)         0         \n",
      " 2D)                                                             \n",
      "                                                                 \n",
      " conv2d_2 (Conv2D)           (None, 7, 7, 128)         204928    \n",
      "                                                                 \n",
      " max_pooling2d_2 (MaxPooling  (None, 3, 3, 128)        0         \n",
      " 2D)                                                             \n",
      "                                                                 \n",
      " flatten (Flatten)           (None, 1152)              0         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 128)               147584    \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 24)                3096      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 407,704\n",
      "Trainable params: 407,704\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "9acfb354",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "f9ebd211",
   "metadata": {},
   "outputs": [],
   "source": [
    "save_best_cb = keras.callbacks.ModelCheckpoint('models/initial-end-to-end', save_best_only=True) # Saves the best model so far\n",
    "early_stopping_cb = keras.callbacks.EarlyStopping(patience=5) # Interupts training when there is no progress"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "4451f28a",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "782/782 [==============================] - ETA: 0s - loss: 0.8774 - accuracy: 0.7326"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 26s 21ms/step - loss: 0.8774 - accuracy: 0.7326 - val_loss: 0.0267 - val_accuracy: 0.9935\n",
      "Epoch 2/20\n",
      "782/782 [==============================] - ETA: 0s - loss: 0.0051 - accuracy: 0.9994"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 14s 18ms/step - loss: 0.0051 - accuracy: 0.9994 - val_loss: 0.0010 - val_accuracy: 1.0000\n",
      "Epoch 3/20\n",
      "778/782 [============================>.] - ETA: 0s - loss: 5.8701e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 12s 16ms/step - loss: 5.9076e-04 - accuracy: 1.0000 - val_loss: 4.7529e-04 - val_accuracy: 1.0000\n",
      "Epoch 4/20\n",
      "781/782 [============================>.] - ETA: 0s - loss: 0.0357 - accuracy: 0.9893"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 12s 15ms/step - loss: 0.0357 - accuracy: 0.9893 - val_loss: 2.4525e-04 - val_accuracy: 1.0000\n",
      "Epoch 5/20\n",
      "779/782 [============================>.] - ETA: 0s - loss: 1.2933e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 14s 18ms/step - loss: 1.2916e-04 - accuracy: 1.0000 - val_loss: 1.0227e-04 - val_accuracy: 1.0000\n",
      "Epoch 6/20\n",
      "782/782 [==============================] - ETA: 0s - loss: 6.1190e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 13s 16ms/step - loss: 6.1190e-05 - accuracy: 1.0000 - val_loss: 5.5581e-05 - val_accuracy: 1.0000\n",
      "Epoch 7/20\n",
      "779/782 [============================>.] - ETA: 0s - loss: 3.5186e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 14s 18ms/step - loss: 3.5120e-05 - accuracy: 1.0000 - val_loss: 3.2938e-05 - val_accuracy: 1.0000\n",
      "Epoch 8/20\n",
      "778/782 [============================>.] - ETA: 0s - loss: 2.1345e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 12s 16ms/step - loss: 2.1343e-05 - accuracy: 1.0000 - val_loss: 2.0919e-05 - val_accuracy: 1.0000\n",
      "Epoch 9/20\n",
      "779/782 [============================>.] - ETA: 0s - loss: 1.3634e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 14s 18ms/step - loss: 1.3638e-05 - accuracy: 1.0000 - val_loss: 1.3809e-05 - val_accuracy: 1.0000\n",
      "Epoch 10/20\n",
      "778/782 [============================>.] - ETA: 0s - loss: 9.0006e-06 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 13s 17ms/step - loss: 8.9994e-06 - accuracy: 1.0000 - val_loss: 8.8405e-06 - val_accuracy: 1.0000\n",
      "Epoch 11/20\n",
      "780/782 [============================>.] - ETA: 0s - loss: 5.8028e-06 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 12s 15ms/step - loss: 5.8089e-06 - accuracy: 1.0000 - val_loss: 5.6665e-06 - val_accuracy: 1.0000\n",
      "Epoch 12/20\n",
      "782/782 [==============================] - ETA: 0s - loss: 3.8640e-06 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 13s 16ms/step - loss: 3.8640e-06 - accuracy: 1.0000 - val_loss: 3.9938e-06 - val_accuracy: 1.0000\n",
      "Epoch 13/20\n",
      "781/782 [============================>.] - ETA: 0s - loss: 2.5768e-06 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 13s 16ms/step - loss: 2.5762e-06 - accuracy: 1.0000 - val_loss: 2.6486e-06 - val_accuracy: 1.0000\n",
      "Epoch 14/20\n",
      "782/782 [==============================] - ETA: 0s - loss: 1.7138e-06 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 12s 16ms/step - loss: 1.7138e-06 - accuracy: 1.0000 - val_loss: 1.8131e-06 - val_accuracy: 1.0000\n",
      "Epoch 15/20\n",
      "782/782 [==============================] - ETA: 0s - loss: 1.1717e-06 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 12s 15ms/step - loss: 1.1717e-06 - accuracy: 1.0000 - val_loss: 1.2420e-06 - val_accuracy: 1.0000\n",
      "Epoch 16/20\n",
      "779/782 [============================>.] - ETA: 0s - loss: 7.6287e-07 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 12s 15ms/step - loss: 7.6705e-07 - accuracy: 1.0000 - val_loss: 8.0546e-07 - val_accuracy: 1.0000\n",
      "Epoch 17/20\n",
      "782/782 [==============================] - ETA: 0s - loss: 5.0912e-07 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 13s 16ms/step - loss: 5.0912e-07 - accuracy: 1.0000 - val_loss: 5.3233e-07 - val_accuracy: 1.0000\n",
      "Epoch 18/20\n",
      "782/782 [==============================] - ETA: 0s - loss: 3.4244e-07 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 12s 15ms/step - loss: 3.4244e-07 - accuracy: 1.0000 - val_loss: 4.0079e-07 - val_accuracy: 1.0000\n",
      "Epoch 19/20\n",
      "779/782 [============================>.] - ETA: 0s - loss: 2.2810e-07 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 11s 15ms/step - loss: 2.2815e-07 - accuracy: 1.0000 - val_loss: 2.9071e-07 - val_accuracy: 1.0000\n",
      "Epoch 20/20\n",
      "780/782 [============================>.] - ETA: 0s - loss: 1.5193e-07 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: initial-end-to-end\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
      "782/782 [==============================] - 11s 15ms/step - loss: 1.5183e-07 - accuracy: 1.0000 - val_loss: 1.6801e-07 - val_accuracy: 1.0000\n"
     ]
    }
   ],
   "source": [
    "# The model is same is 'models/initial-end-to-end'\n",
    "# The history object is 'models/initial-end-to-end-history'\n",
    "\n",
    "history = model.fit(X_train, y_train, epochs=20, validation_data=(X_valid, y_valid), callbacks=[save_best_cb, early_stopping_cb])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "c5e12a49",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'loss': [0.8774073719978333,\n",
       "  0.005074513144791126,\n",
       "  0.0005907623562961817,\n",
       "  0.035671062767505646,\n",
       "  0.00012915894330944866,\n",
       "  6.119009776739404e-05,\n",
       "  3.5120148822898045e-05,\n",
       "  2.134285750798881e-05,\n",
       "  1.3638154086947907e-05,\n",
       "  8.999433703138493e-06,\n",
       "  5.808890364278341e-06,\n",
       "  3.864049176627304e-06,\n",
       "  2.576162614786881e-06,\n",
       "  1.7137577970061102e-06,\n",
       "  1.1716660992533434e-06,\n",
       "  7.670482204957807e-07,\n",
       "  5.091210937280266e-07,\n",
       "  3.424441104016296e-07,\n",
       "  2.2814701594597864e-07,\n",
       "  1.518342855888477e-07],\n",
       " 'accuracy': [0.7325999736785889,\n",
       "  0.9993600249290466,\n",
       "  1.0,\n",
       "  0.9892799854278564,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0],\n",
       " 'val_loss': [0.02665344439446926,\n",
       "  0.0010351891396567225,\n",
       "  0.0004752879322040826,\n",
       "  0.00024525265325792134,\n",
       "  0.00010227379971183836,\n",
       "  5.5580920161446556e-05,\n",
       "  3.293823101557791e-05,\n",
       "  2.091940950776916e-05,\n",
       "  1.3808692528982647e-05,\n",
       "  8.840480404614937e-06,\n",
       "  5.666528068104526e-06,\n",
       "  3.9937822293723e-06,\n",
       "  2.64862956100842e-06,\n",
       "  1.8130501757696038e-06,\n",
       "  1.2420367738741334e-06,\n",
       "  8.05462889275077e-07,\n",
       "  5.323349228092411e-07,\n",
       "  4.007931124760944e-07,\n",
       "  2.907113696437591e-07,\n",
       "  1.6800919411252835e-07],\n",
       " 'val_accuracy': [0.9934827089309692,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0]}"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "history.history # Contains the training related information for each epoch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "c3aa7673",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Saving the history object\n",
    "\n",
    "# with open('models/intial-end-to-end-history', 'wb') as history_file:\n",
    "#     pickle.dump(history.history, history_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "acc714c0",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'loss': [0.8774073719978333,\n",
       "  0.005074513144791126,\n",
       "  0.0005907623562961817,\n",
       "  0.035671062767505646,\n",
       "  0.00012915894330944866,\n",
       "  6.119009776739404e-05,\n",
       "  3.5120148822898045e-05,\n",
       "  2.134285750798881e-05,\n",
       "  1.3638154086947907e-05,\n",
       "  8.999433703138493e-06,\n",
       "  5.808890364278341e-06,\n",
       "  3.864049176627304e-06,\n",
       "  2.576162614786881e-06,\n",
       "  1.7137577970061102e-06,\n",
       "  1.1716660992533434e-06,\n",
       "  7.670482204957807e-07,\n",
       "  5.091210937280266e-07,\n",
       "  3.424441104016296e-07,\n",
       "  2.2814701594597864e-07,\n",
       "  1.518342855888477e-07],\n",
       " 'accuracy': [0.7325999736785889,\n",
       "  0.9993600249290466,\n",
       "  1.0,\n",
       "  0.9892799854278564,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0],\n",
       " 'val_loss': [0.02665344439446926,\n",
       "  0.0010351891396567225,\n",
       "  0.0004752879322040826,\n",
       "  0.00024525265325792134,\n",
       "  0.00010227379971183836,\n",
       "  5.5580920161446556e-05,\n",
       "  3.293823101557791e-05,\n",
       "  2.091940950776916e-05,\n",
       "  1.3808692528982647e-05,\n",
       "  8.840480404614937e-06,\n",
       "  5.666528068104526e-06,\n",
       "  3.9937822293723e-06,\n",
       "  2.64862956100842e-06,\n",
       "  1.8130501757696038e-06,\n",
       "  1.2420367738741334e-06,\n",
       "  8.05462889275077e-07,\n",
       "  5.323349228092411e-07,\n",
       "  4.007931124760944e-07,\n",
       "  2.907113696437591e-07,\n",
       "  1.6800919411252835e-07],\n",
       " 'val_accuracy': [0.9934827089309692,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0,\n",
       "  1.0]}"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "h = np.load('models/intial-end-to-end-history', allow_pickle=True)\n",
    "h"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "96517c3b",
   "metadata": {},
   "outputs": [],
   "source": [
    "best_model = keras.models.load_model('models/initial-end-to-end') # Model with best set of parameters not necessarily the model at the last epoch of training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "94301555",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x16c5dc26580>"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(10, 5))\n",
    "n_epochs = len(h['loss'])\n",
    "ax.plot(range(1, n_epochs+1), h['loss'], color='b', label='train_loss')\n",
    "ax.plot(range(1, n_epochs+1), h['val_loss'], color='c', label='val_loss')\n",
    "ax.plot(range(1, n_epochs+1), h['accuracy'], color='b', label='train_accuracy', linestyle='--')\n",
    "ax.plot(range(1, n_epochs+1), h['val_accuracy'], color='c', label='val_accuracy', linestyle='--')\n",
    "ax.set_xticks(range(1, n_epochs+1))\n",
    "ax.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "1cceeb24",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x16c5db98fa0>"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Training Loss Correction\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(10, 5))\n",
    "n_epochs = len(h['loss'])\n",
    "\n",
    "# Shift training loss by 0.5 as training loss is measured during the epoch and validation loss is measured after the epoch\n",
    "\n",
    "x_loss = np.arange(n_epochs+1)-0.5\n",
    "ax.plot(x_loss[x_loss >= 0], h['loss'], color='b', label='train_loss')\n",
    "ax.plot(range(1, n_epochs+1), h['val_loss'], color='r', label='val_loss')\n",
    "ax.plot(range(1, n_epochs+1), h['accuracy'], color='b', label='train_accuracy', linestyle='--')\n",
    "ax.plot(range(1, n_epochs+1), h['val_accuracy'], color='r', label='val_accuracy', linestyle='--')\n",
    "ax.set_xlim(0, n_epochs)\n",
    "ax.set_xticks(range(1, n_epochs+1))\n",
    "ax.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "b7e5d4b0",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Defining a function the get the training and validation plots representing the accuracy and loss at each epoch\n",
    "\n",
    "def get_train_val_plots(h, yticks=None, figsize=(10, 5)):\n",
    "    # h: Any dictionary like the history.history\n",
    "    \n",
    "    fig, ax = plt.subplots(figsize=figsize)\n",
    "    n_epochs = len(h['loss'])\n",
    "    x_loss = np.arange(n_epochs+1)-0.5\n",
    "    \n",
    "    ax.plot(x_loss[x_loss >= 0], h['loss'], color='b', label='train_loss')\n",
    "    ax.plot(range(1, n_epochs+1), h['val_loss'], color='r', label='val_loss')\n",
    "    ax.plot(range(1, n_epochs+1), h['accuracy'], color='b', label='train_accuracy', linestyle='--')\n",
    "    ax.plot(range(1, n_epochs+1), h['val_accuracy'], color='r', label='val_accuracy', linestyle='--')\n",
    "    ax.set_xlim(0, n_epochs)\n",
    "    ax.set_xticks(range(1, n_epochs+1))\n",
    "    if yticks is not None:\n",
    "        ax.set_yticks(yticks)\n",
    "    ax.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "6f1885f7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "get_train_val_plots(h)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "aadf9622",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_df = pd.read_csv('data/alphabet/sign_mnist_test.csv') # Load the test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "c22ee404",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test, y_test = test_df.drop('label', axis=1), test_df['label']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "50fcf741",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = tf.reshape(X_test, [-1, 28, 28, 1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "04d31bf4",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test = label_binarizer.transform(y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "4c761190",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "225/225 [==============================] - 2s 8ms/step - loss: 87.9655 - accuracy: 0.9405\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[87.96553802490234, 0.9404628872871399]"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best_model.evaluate(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9085b7ff",
   "metadata": {},
   "source": [
    "Accuracy: 94%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "685402ff",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preprocesses the input and evaluates the model\n",
    "\n",
    "def evaluate_model(model, X_test, y_test, label_binarizer):\n",
    "    X_test_reshape = tf.reshape(X_test, [-1, 28, 28, 1])\n",
    "    y_test_labels = label_binarizer.transform(y_test)\n",
    "    results = model.evaluate(X_test_reshape, y_test_labels)\n",
    "    print(f'Loss: {results[0]:.3f} Accuracy: {results[1]:.3f}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "af272eed",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "225/225 [==============================] - 1s 6ms/step - loss: 87.9655 - accuracy: 0.9405\n",
      "Loss: 87.966 Accuracy: 0.940\n"
     ]
    }
   ],
   "source": [
    "results = evaluate_model(best_model, test_df.drop('label', axis=1), test_df['label'], label_binarizer)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3edb545c",
   "metadata": {},
   "source": [
    "## Converting Images to a Sentence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "70c6d7cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_df = pd.read_csv('data/alphabet/sign_mnist_test.csv') # Load the test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "fcf80126",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test, y_test = test_df.drop('label', axis=1), test_df['label']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "c01b2e70",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = tf.reshape(X_test, [-1, 28, 28, 1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "92beca5f",
   "metadata": {},
   "outputs": [],
   "source": [
    "d = {chr(ord('a') + i):i for i in range(26)}\n",
    "d_rev = {i:chr(ord('a') + i) for i in range(26)}\n",
    "d[' '] = d_rev[' '] = ' '"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "4e7518e4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18 8 6 13   11 0 13 6 20 0 6 4 "
     ]
    }
   ],
   "source": [
    "sentence = 'sign language'\n",
    "\n",
    "for i in sentence:\n",
    "    print(d[i], end=' ')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "4f429ea0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 288ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
       "        0., 0., 0., 0., 0., 0., 0., 0.]], dtype=float32)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best_model.predict(tf.reshape(X_test[0], [-1, 28, 28, 1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "2f09fb2d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 48ms/step\n",
      "1/1 [==============================] - 0s 39ms/step\n",
      "1/1 [==============================] - 0s 39ms/step\n",
      "1/1 [==============================] - 0s 40ms/step\n",
      "1/1 [==============================] - 0s 38ms/step\n",
      "1/1 [==============================] - 0s 37ms/step\n",
      "1/1 [==============================] - 0s 44ms/step\n",
      "1/1 [==============================] - 0s 58ms/step\n",
      "1/1 [==============================] - 0s 67ms/step\n",
      "1/1 [==============================] - 0s 38ms/step\n",
      "1/1 [==============================] - 0s 47ms/step\n",
      "1/1 [==============================] - 0s 38ms/step\n",
      "sign language\n"
     ]
    }
   ],
   "source": [
    "images_taken = []\n",
    "result = ''\n",
    "\n",
    "for i in sentence:\n",
    "    if i != ' ':\n",
    "        char_index = np.random.choice(y_test[y_test==ord(i)-ord('a')].index)\n",
    "        images_taken.append(char_index)\n",
    "        y_pred = best_model.predict(tf.reshape(X_test[char_index], [-1, 28, 28, 1]))\n",
    "        result += d_rev[label_binarizer.inverse_transform(y_pred)[0]]\n",
    "    else:\n",
    "        result += ' '\n",
    "print(result)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "f19190d7",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 8 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualizing the test images\n",
    "images_taken_dup = list(reversed(images_taken))\n",
    "for word in sentence.split():\n",
    "    fig, ax = plt.subplots(1, len(word), figsize=(20, 20))\n",
    "    for i in range(len(word)):\n",
    "        ax[i].imshow(X_test[images_taken_dup.pop()], cmap='gray')\n",
    "        ax[i].set_title(word[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "fd0be37f",
   "metadata": {},
   "outputs": [],
   "source": [
    "def test_on_sentence(model, sentence, X_test, y_test, label_binarizer, figsize=(20, 20)):\n",
    "    # Random images are taken from X_test along with the corresponding labels in y_test\n",
    "    # based on the letters in the sentence.\n",
    "    # These images are fed to the model and its output is printed\n",
    "    \n",
    "    sentence = sentence.lower()\n",
    "    \n",
    "    d = {chr(ord('a') + i):i for i in range(26)}\n",
    "    d_rev = {i:chr(ord('a') + i) for i in range(26)}\n",
    "    d[' '] = d_rev[' '] = ' '\n",
    "    \n",
    "    \n",
    "    images_taken = []\n",
    "    result = ''\n",
    "    \n",
    "    X_test_reshape = tf.reshape(X_test, [-1, 28, 28, 1])\n",
    "    \n",
    "\n",
    "    for i in sentence:\n",
    "        if i != ' ':\n",
    "            char_index = np.random.choice(y_test[y_test==ord(i)-ord('a')].index)\n",
    "            images_taken.append(char_index)\n",
    "            y_pred = model.predict(tf.reshape(X_test_reshape[char_index], [1, 28, 28, 1]))\n",
    "            result += d_rev[label_binarizer.inverse_transform(y_pred)[0]]\n",
    "        else:\n",
    "            result += ' '\n",
    "            \n",
    "    print(f'The actual sentence is \"{sentence}\"')\n",
    "    print(f'The predicted sentence is \"{result}\"')\n",
    "        \n",
    "    images_taken.reverse()\n",
    "    for word in sentence.split():\n",
    "        fig, ax = plt.subplots(1, len(word), figsize=figsize)\n",
    "        for i in range(len(word)):\n",
    "            ax[i].imshow(X_test_reshape[images_taken.pop()], cmap='gray')\n",
    "            ax[i].set_title(word[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "d8c07946",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 37ms/step\n",
      "1/1 [==============================] - 0s 40ms/step\n",
      "1/1 [==============================] - 0s 49ms/step\n",
      "1/1 [==============================] - 0s 48ms/step\n",
      "1/1 [==============================] - 0s 35ms/step\n",
      "1/1 [==============================] - 0s 29ms/step\n",
      "1/1 [==============================] - 0s 40ms/step\n",
      "1/1 [==============================] - 0s 40ms/step\n",
      "1/1 [==============================] - 0s 36ms/step\n",
      "1/1 [==============================] - 0s 76ms/step\n",
      "1/1 [==============================] - 0s 54ms/step\n",
      "1/1 [==============================] - 0s 44ms/step\n",
      "The actual sentence is \"sign language\"\n",
      "The predicted sentence is \"sign language\"\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 8 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_on_sentence(best_model, 'sign language', test_df.drop('label', axis=1), test_df['label'], label_binarizer)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d9b0ecea",
   "metadata": {},
   "source": [
    "# Initial End-to-End Workflow (End)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b127a589",
   "metadata": {},
   "source": [
    "# Hyperparameter Tuning"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2d26fb0e",
   "metadata": {},
   "source": [
    "1. Convolution and Max Pooling Pairs\n",
    "2. Filters in the convolution layers\n",
    "3. Filter Shape\n",
    "4. Dropout"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "393e2a7f",
   "metadata": {},
   "source": [
    "## Convolution and Max Pooling Pairs"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "640290aa",
   "metadata": {},
   "source": [
    "Before flattening\n",
    "\n",
    "\n",
    "1. For pair = 1 -> Output to the dense layer will be of the shape `(None, 14, 14, 32)`\n",
    "2. For pair = 2 -> Output to the dense layer will be of the shape `(None, 7, 7, 64)`\n",
    "3. For pair = 3 -> Output to the dense layer will be of the shape `(None, 3, 3, 96)`\n",
    "4. For pair = 4 -> Output to the dense layer will be of the shape `(None, 1, 1, 128)`\n",
    "\n",
    "As the output shape rapidly decreases for the pair = 4 it is better to choose among the pair = 1, 2 or 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "1f392588",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 20% Validation Split\n",
    "\n",
    "X_train, X_valid = X[:19500], X[19500:]\n",
    "y_train, y_valid = y[:19500], y[19500:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "e5583291",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "598/610 [============================>.] - ETA: 0s - loss: 1.2888 - accuracy: 0.6378"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 4s 6ms/step - loss: 1.2717 - accuracy: 0.6428 - val_loss: 0.3912 - val_accuracy: 0.9062\n",
      "Epoch 2/15\n",
      "605/610 [============================>.] - ETA: 0s - loss: 0.2074 - accuracy: 0.9560"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 4s 6ms/step - loss: 0.2066 - accuracy: 0.9562 - val_loss: 0.0849 - val_accuracy: 0.9912\n",
      "Epoch 3/15\n",
      "608/610 [============================>.] - ETA: 0s - loss: 0.0515 - accuracy: 0.9958"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 5s 9ms/step - loss: 0.0514 - accuracy: 0.9958 - val_loss: 0.0299 - val_accuracy: 0.9982\n",
      "Epoch 4/15\n",
      "609/610 [============================>.] - ETA: 0s - loss: 0.0165 - accuracy: 0.9997"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 4s 6ms/step - loss: 0.0165 - accuracy: 0.9997 - val_loss: 0.0115 - val_accuracy: 1.0000\n",
      "Epoch 5/15\n",
      "606/610 [============================>.] - ETA: 0s - loss: 0.0073 - accuracy: 0.9999"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 4s 6ms/step - loss: 0.0073 - accuracy: 0.9999 - val_loss: 0.0049 - val_accuracy: 1.0000\n",
      "Epoch 6/15\n",
      "609/610 [============================>.] - ETA: 0s - loss: 0.0066 - accuracy: 0.9993"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 4s 6ms/step - loss: 0.0066 - accuracy: 0.9993 - val_loss: 0.0044 - val_accuracy: 1.0000\n",
      "Epoch 7/15\n",
      "610/610 [==============================] - ETA: 0s - loss: 0.0022 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 4s 7ms/step - loss: 0.0022 - accuracy: 1.0000 - val_loss: 0.0017 - val_accuracy: 1.0000\n",
      "Epoch 8/15\n",
      "610/610 [==============================] - ETA: 0s - loss: 0.0013 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 5s 8ms/step - loss: 0.0013 - accuracy: 1.0000 - val_loss: 0.0012 - val_accuracy: 1.0000\n",
      "Epoch 9/15\n",
      "609/610 [============================>.] - ETA: 0s - loss: 8.6837e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 4s 7ms/step - loss: 8.6799e-04 - accuracy: 1.0000 - val_loss: 7.7809e-04 - val_accuracy: 1.0000\n",
      "Epoch 10/15\n",
      "605/610 [============================>.] - ETA: 0s - loss: 5.8230e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 5s 9ms/step - loss: 5.8304e-04 - accuracy: 1.0000 - val_loss: 5.4940e-04 - val_accuracy: 1.0000\n",
      "Epoch 11/15\n",
      "610/610 [==============================] - 4s 6ms/step - loss: 0.0314 - accuracy: 0.9908 - val_loss: 0.0015 - val_accuracy: 1.0000\n",
      "Epoch 12/15\n",
      "610/610 [==============================] - 4s 6ms/step - loss: 8.4445e-04 - accuracy: 1.0000 - val_loss: 6.3095e-04 - val_accuracy: 1.0000\n",
      "Epoch 13/15\n",
      "608/610 [============================>.] - ETA: 0s - loss: 4.8544e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 4s 6ms/step - loss: 4.8522e-04 - accuracy: 1.0000 - val_loss: 4.7309e-04 - val_accuracy: 1.0000\n",
      "Epoch 14/15\n",
      "607/610 [============================>.] - ETA: 0s - loss: 3.5215e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 4s 7ms/step - loss: 3.5329e-04 - accuracy: 1.0000 - val_loss: 3.3240e-04 - val_accuracy: 1.0000\n",
      "Epoch 15/15\n",
      "605/610 [============================>.] - ETA: 0s - loss: 2.7148e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 4s 7ms/step - loss: 2.7154e-04 - accuracy: 1.0000 - val_loss: 2.5948e-04 - val_accuracy: 1.0000\n",
      "Epoch 1/15\n",
      "608/610 [============================>.] - ETA: 0s - loss: 0.9688 - accuracy: 0.7083"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 10ms/step - loss: 0.9668 - accuracy: 0.7090 - val_loss: 0.1467 - val_accuracy: 0.9531\n",
      "Epoch 2/15\n",
      "606/610 [============================>.] - ETA: 0s - loss: 0.0457 - accuracy: 0.9910"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 11ms/step - loss: 0.0455 - accuracy: 0.9910 - val_loss: 0.0085 - val_accuracy: 0.9997\n",
      "Epoch 3/15\n",
      "605/610 [============================>.] - ETA: 0s - loss: 0.0043 - accuracy: 0.9999"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 6s 9ms/step - loss: 0.0043 - accuracy: 0.9999 - val_loss: 0.0021 - val_accuracy: 1.0000\n",
      "Epoch 4/15\n",
      "610/610 [==============================] - 6s 9ms/step - loss: 0.0376 - accuracy: 0.9897 - val_loss: 0.0031 - val_accuracy: 1.0000\n",
      "Epoch 5/15\n",
      "605/610 [============================>.] - ETA: 0s - loss: 9.2055e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 6s 10ms/step - loss: 9.1743e-04 - accuracy: 1.0000 - val_loss: 4.9071e-04 - val_accuracy: 1.0000\n",
      "Epoch 6/15\n",
      "603/610 [============================>.] - ETA: 0s - loss: 3.3840e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 5s 9ms/step - loss: 3.3749e-04 - accuracy: 1.0000 - val_loss: 2.7421e-04 - val_accuracy: 1.0000\n",
      "Epoch 7/15\n",
      "600/610 [============================>.] - ETA: 0s - loss: 1.9964e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 6s 10ms/step - loss: 1.9890e-04 - accuracy: 1.0000 - val_loss: 1.8108e-04 - val_accuracy: 1.0000\n",
      "Epoch 8/15\n",
      "605/610 [============================>.] - ETA: 0s - loss: 1.3324e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 6s 9ms/step - loss: 1.3388e-04 - accuracy: 1.0000 - val_loss: 1.2290e-04 - val_accuracy: 1.0000\n",
      "Epoch 9/15\n",
      "609/610 [============================>.] - ETA: 0s - loss: 9.4069e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 6s 10ms/step - loss: 9.4024e-05 - accuracy: 1.0000 - val_loss: 8.7907e-05 - val_accuracy: 1.0000\n",
      "Epoch 10/15\n",
      "599/610 [============================>.] - ETA: 0s - loss: 6.4930e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 5s 8ms/step - loss: 6.5206e-05 - accuracy: 1.0000 - val_loss: 6.4753e-05 - val_accuracy: 1.0000\n",
      "Epoch 11/15\n",
      "604/610 [============================>.] - ETA: 0s - loss: 4.5556e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 6s 10ms/step - loss: 4.5509e-05 - accuracy: 1.0000 - val_loss: 3.9519e-05 - val_accuracy: 1.0000\n",
      "Epoch 12/15\n",
      "608/610 [============================>.] - ETA: 0s - loss: 3.0262e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 6s 10ms/step - loss: 3.0241e-05 - accuracy: 1.0000 - val_loss: 2.7037e-05 - val_accuracy: 1.0000\n",
      "Epoch 13/15\n",
      "608/610 [============================>.] - ETA: 0s - loss: 2.0512e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 12ms/step - loss: 2.0501e-05 - accuracy: 1.0000 - val_loss: 1.9547e-05 - val_accuracy: 1.0000\n",
      "Epoch 14/15\n",
      "610/610 [==============================] - ETA: 0s - loss: 1.4227e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 5s 9ms/step - loss: 1.4227e-05 - accuracy: 1.0000 - val_loss: 1.3458e-05 - val_accuracy: 1.0000\n",
      "Epoch 15/15\n",
      "610/610 [==============================] - ETA: 0s - loss: 9.9346e-06 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 6s 10ms/step - loss: 9.9346e-06 - accuracy: 1.0000 - val_loss: 9.3704e-06 - val_accuracy: 1.0000\n",
      "Epoch 1/15\n",
      "610/610 [==============================] - ETA: 0s - loss: 0.8947 - accuracy: 0.7210"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 9s 13ms/step - loss: 0.8947 - accuracy: 0.7210 - val_loss: 0.0764 - val_accuracy: 0.9777\n",
      "Epoch 2/15\n",
      "605/610 [============================>.] - ETA: 0s - loss: 0.0376 - accuracy: 0.9903"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 8s 13ms/step - loss: 0.0374 - accuracy: 0.9904 - val_loss: 0.0031 - val_accuracy: 1.0000\n",
      "Epoch 3/15\n",
      "610/610 [==============================] - 6s 10ms/step - loss: 0.0161 - accuracy: 0.9950 - val_loss: 0.1017 - val_accuracy: 0.9635\n",
      "Epoch 4/15\n",
      "610/610 [==============================] - 7s 11ms/step - loss: 0.0285 - accuracy: 0.9910 - val_loss: 0.0222 - val_accuracy: 0.9937\n",
      "Epoch 5/15\n",
      "606/610 [============================>.] - ETA: 0s - loss: 0.0072 - accuracy: 0.9979"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 8s 13ms/step - loss: 0.0072 - accuracy: 0.9979 - val_loss: 1.4488e-04 - val_accuracy: 1.0000\n",
      "Epoch 6/15\n",
      "610/610 [==============================] - ETA: 0s - loss: 7.7258e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 12ms/step - loss: 7.7258e-05 - accuracy: 1.0000 - val_loss: 5.4070e-05 - val_accuracy: 1.0000\n",
      "Epoch 7/15\n",
      "606/610 [============================>.] - ETA: 0s - loss: 3.4622e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 8s 12ms/step - loss: 3.4583e-05 - accuracy: 1.0000 - val_loss: 3.0925e-05 - val_accuracy: 1.0000\n",
      "Epoch 8/15\n",
      "608/610 [============================>.] - ETA: 0s - loss: 2.0418e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 6s 10ms/step - loss: 2.0388e-05 - accuracy: 1.0000 - val_loss: 1.9665e-05 - val_accuracy: 1.0000\n",
      "Epoch 9/15\n",
      "606/610 [============================>.] - ETA: 0s - loss: 1.3087e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 12ms/step - loss: 1.3067e-05 - accuracy: 1.0000 - val_loss: 1.3650e-05 - val_accuracy: 1.0000\n",
      "Epoch 10/15\n",
      "605/610 [============================>.] - ETA: 0s - loss: 8.9536e-06 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 12ms/step - loss: 8.9197e-06 - accuracy: 1.0000 - val_loss: 1.0479e-05 - val_accuracy: 1.0000\n",
      "Epoch 11/15\n",
      "604/610 [============================>.] - ETA: 0s - loss: 6.4773e-06 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 12ms/step - loss: 6.4600e-06 - accuracy: 1.0000 - val_loss: 7.0963e-06 - val_accuracy: 1.0000\n",
      "Epoch 12/15\n",
      "605/610 [============================>.] - ETA: 0s - loss: 4.6440e-06 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 12ms/step - loss: 4.6478e-06 - accuracy: 1.0000 - val_loss: 5.8179e-06 - val_accuracy: 1.0000\n",
      "Epoch 13/15\n",
      "610/610 [==============================] - ETA: 0s - loss: 3.4811e-06 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 12ms/step - loss: 3.4811e-06 - accuracy: 1.0000 - val_loss: 4.2011e-06 - val_accuracy: 1.0000\n",
      "Epoch 14/15\n",
      "603/610 [============================>.] - ETA: 0s - loss: 2.4777e-06 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 12ms/step - loss: 2.4738e-06 - accuracy: 1.0000 - val_loss: 2.9150e-06 - val_accuracy: 1.0000\n",
      "Epoch 15/15\n",
      "604/610 [============================>.] - ETA: 0s - loss: 1.8221e-06 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-1-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
      "610/610 [==============================] - 8s 12ms/step - loss: 1.8241e-06 - accuracy: 1.0000 - val_loss: 2.5653e-06 - val_accuracy: 1.0000\n"
     ]
    }
   ],
   "source": [
    "# Models\n",
    "# 'models/experiment-1-1'\n",
    "# 'models/experiment-1-2'\n",
    "# 'models/experiment-1-3'\n",
    "\n",
    "# History objects\n",
    "# 'models/experiment-1-1-history'\n",
    "# 'models/experiment-1-2-history'\n",
    "# 'models/experiment-1-3-history'\n",
    "\n",
    "n_pairs = 3\n",
    "models_pairs = [keras.models.Sequential() for i in range(n_pairs)]\n",
    "early_stopping_cb = keras.callbacks.EarlyStopping(patience=5)\n",
    "\n",
    "for n in range(1, n_pairs+1):\n",
    "    models_pairs[n-1].add(keras.layers.Conv2D(32, (5, 5), padding='same', activation='relu', input_shape=(28, 28, 1)))\n",
    "    models_pairs[n-1].add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "    for i in range(1, n):\n",
    "        models_pairs[n-1].add(keras.layers.Conv2D(32*(i+1), (5, 5), padding='same', activation='relu'))\n",
    "        models_pairs[n-1].add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "    models_pairs[n-1].add(keras.layers.Flatten())\n",
    "    models_pairs[n-1].add(keras.layers.Dense(128, activation='relu'))\n",
    "    models_pairs[n-1].add(keras.layers.Dense(24, activation='softmax'))\n",
    "    models_pairs[n-1].compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "    save_best_cb = keras.callbacks.ModelCheckpoint(f'models/experiment-1-{n}', save_best_only=True)\n",
    "    history = models_pairs[n-1].fit(X_train, y_train, epochs=15, validation_data=(X_valid, y_valid), callbacks=[save_best_cb, early_stopping_cb])\n",
    "    with open(f'models/experiment-1-{n}-history', 'wb') as history_file:\n",
    "        pickle.dump(history.history, history_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "fe1939d8",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_8\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d_7 (Conv2D)           (None, 28, 28, 32)        832       \n",
      "                                                                 \n",
      " max_pooling2d_7 (MaxPooling  (None, 14, 14, 32)       0         \n",
      " 2D)                                                             \n",
      "                                                                 \n",
      " flatten_4 (Flatten)         (None, 6272)              0         \n",
      "                                                                 \n",
      " dense_8 (Dense)             (None, 128)               802944    \n",
      "                                                                 \n",
      " dense_9 (Dense)             (None, 24)                3096      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 806,872\n",
      "Trainable params: 806,872\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "models_pairs[0].summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "8945d1b6",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_9\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d_8 (Conv2D)           (None, 28, 28, 32)        832       \n",
      "                                                                 \n",
      " max_pooling2d_8 (MaxPooling  (None, 14, 14, 32)       0         \n",
      " 2D)                                                             \n",
      "                                                                 \n",
      " conv2d_9 (Conv2D)           (None, 14, 14, 64)        51264     \n",
      "                                                                 \n",
      " max_pooling2d_9 (MaxPooling  (None, 7, 7, 64)         0         \n",
      " 2D)                                                             \n",
      "                                                                 \n",
      " flatten_5 (Flatten)         (None, 3136)              0         \n",
      "                                                                 \n",
      " dense_10 (Dense)            (None, 128)               401536    \n",
      "                                                                 \n",
      " dense_11 (Dense)            (None, 24)                3096      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 456,728\n",
      "Trainable params: 456,728\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "models_pairs[1].summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "c7a0122c",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_10\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d_10 (Conv2D)          (None, 28, 28, 32)        832       \n",
      "                                                                 \n",
      " max_pooling2d_10 (MaxPoolin  (None, 14, 14, 32)       0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " conv2d_11 (Conv2D)          (None, 14, 14, 64)        51264     \n",
      "                                                                 \n",
      " max_pooling2d_11 (MaxPoolin  (None, 7, 7, 64)         0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " conv2d_12 (Conv2D)          (None, 7, 7, 96)          153696    \n",
      "                                                                 \n",
      " max_pooling2d_12 (MaxPoolin  (None, 3, 3, 96)         0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " flatten_6 (Flatten)         (None, 864)               0         \n",
      "                                                                 \n",
      " dense_12 (Dense)            (None, 128)               110720    \n",
      "                                                                 \n",
      " dense_13 (Dense)            (None, 24)                3096      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 319,608\n",
      "Trainable params: 319,608\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "models_pairs[2].summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "f5e21d6c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "249/249 [==============================] - 3s 10ms/step - loss: 2.5948e-04 - accuracy: 1.0000\n",
      "249/249 [==============================] - 2s 8ms/step - loss: 9.3704e-06 - accuracy: 1.0000\n",
      "249/249 [==============================] - 3s 8ms/step - loss: 2.5653e-06 - accuracy: 1.0000\n"
     ]
    }
   ],
   "source": [
    "for index in range(n_pairs):\n",
    "    model = keras.models.load_model(f'models/experiment-1-{index+1}')\n",
    "    model.evaluate(X_valid, y_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "83c8007b",
   "metadata": {},
   "outputs": [],
   "source": [
    "h_1_1 = np.load('models/experiment-1-1-history', allow_pickle=True)\n",
    "h_1_2 = np.load('models/experiment-1-2-history', allow_pickle=True)\n",
    "h_1_3 = np.load('models/experiment-1-3-history', allow_pickle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "b62ddb38",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "get_train_val_plots(h_1_1, yticks=np.arange(0, 1.2, 0.1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "32fe1158",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "get_train_val_plots(h_1_2, yticks=np.arange(0, 1.2, 0.1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "30310268",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "get_train_val_plots(h_1_3, yticks=np.arange(0, 1.2, 0.1))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "28be53a8",
   "metadata": {},
   "source": [
    "Model 3 has the least loss on the validation data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1ebbbd65",
   "metadata": {},
   "source": [
    "## Filters\n",
    "\n",
    "Our Models now contains 3 pairs of Convolution and Pooling layers\n",
    "\n",
    "Number of filters maps can be\n",
    "\n",
    "1. 8 - 16 - 32\n",
    "2. 16 - 32 - 64\n",
    "3. 24 - 48 - 96"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "51ff2cb0",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "610/610 [==============================] - ETA: 0s - loss: 1.2862 - accuracy: 0.6032"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 13s 16ms/step - loss: 1.2862 - accuracy: 0.6032 - val_loss: 0.2935 - val_accuracy: 0.9141\n",
      "Epoch 2/10\n",
      "609/610 [============================>.] - ETA: 0s - loss: 0.1381 - accuracy: 0.9593"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 9s 16ms/step - loss: 0.1380 - accuracy: 0.9593 - val_loss: 0.0743 - val_accuracy: 0.9759\n",
      "Epoch 3/10\n",
      "607/610 [============================>.] - ETA: 0s - loss: 0.0472 - accuracy: 0.9872"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 9s 15ms/step - loss: 0.0470 - accuracy: 0.9872 - val_loss: 0.0509 - val_accuracy: 0.9871\n",
      "Epoch 4/10\n",
      "609/610 [============================>.] - ETA: 0s - loss: 0.0094 - accuracy: 0.9979"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 11s 18ms/step - loss: 0.0094 - accuracy: 0.9979 - val_loss: 0.0010 - val_accuracy: 1.0000\n",
      "Epoch 5/10\n",
      "604/610 [============================>.] - ETA: 0s - loss: 6.8334e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 10s 17ms/step - loss: 6.8063e-04 - accuracy: 1.0000 - val_loss: 5.1600e-04 - val_accuracy: 1.0000\n",
      "Epoch 6/10\n",
      "604/610 [============================>.] - ETA: 0s - loss: 3.9389e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 8s 13ms/step - loss: 3.9279e-04 - accuracy: 1.0000 - val_loss: 3.6458e-04 - val_accuracy: 1.0000\n",
      "Epoch 7/10\n",
      "610/610 [==============================] - ETA: 0s - loss: 2.3906e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 9s 14ms/step - loss: 2.3906e-04 - accuracy: 1.0000 - val_loss: 2.3030e-04 - val_accuracy: 1.0000\n",
      "Epoch 8/10\n",
      "605/610 [============================>.] - ETA: 0s - loss: 1.6341e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 12ms/step - loss: 1.6373e-04 - accuracy: 1.0000 - val_loss: 1.6579e-04 - val_accuracy: 1.0000\n",
      "Epoch 9/10\n",
      "610/610 [==============================] - ETA: 0s - loss: 1.1511e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 8s 14ms/step - loss: 1.1511e-04 - accuracy: 1.0000 - val_loss: 1.1253e-04 - val_accuracy: 1.0000\n",
      "Epoch 10/10\n",
      "610/610 [==============================] - 7s 11ms/step - loss: 0.1020 - accuracy: 0.9696 - val_loss: 0.0568 - val_accuracy: 0.9823\n",
      "Epoch 1/10\n",
      "610/610 [==============================] - ETA: 0s - loss: 1.1441 - accuracy: 0.6457"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 12s 17ms/step - loss: 1.1441 - accuracy: 0.6457 - val_loss: 0.2042 - val_accuracy: 0.9358\n",
      "Epoch 2/10\n",
      "606/610 [============================>.] - ETA: 0s - loss: 0.0792 - accuracy: 0.9764"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 8s 13ms/step - loss: 0.0788 - accuracy: 0.9766 - val_loss: 0.0106 - val_accuracy: 0.9997\n",
      "Epoch 3/10\n",
      "605/610 [============================>.] - ETA: 0s - loss: 0.0050 - accuracy: 0.9996"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 8s 14ms/step - loss: 0.0049 - accuracy: 0.9996 - val_loss: 0.0015 - val_accuracy: 1.0000\n",
      "Epoch 4/10\n",
      "608/610 [============================>.] - ETA: 0s - loss: 8.6171e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 9s 15ms/step - loss: 8.6873e-04 - accuracy: 1.0000 - val_loss: 0.0012 - val_accuracy: 1.0000\n",
      "Epoch 5/10\n",
      "604/610 [============================>.] - ETA: 0s - loss: 0.0439 - accuracy: 0.9863"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 9s 15ms/step - loss: 0.0435 - accuracy: 0.9864 - val_loss: 7.8805e-04 - val_accuracy: 1.0000\n",
      "Epoch 6/10\n",
      "605/610 [============================>.] - ETA: 0s - loss: 3.1687e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 8s 13ms/step - loss: 3.1626e-04 - accuracy: 1.0000 - val_loss: 2.6880e-04 - val_accuracy: 1.0000\n",
      "Epoch 7/10\n",
      "608/610 [============================>.] - ETA: 0s - loss: 1.5618e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 9s 14ms/step - loss: 1.5616e-04 - accuracy: 1.0000 - val_loss: 1.6740e-04 - val_accuracy: 1.0000\n",
      "Epoch 8/10\n",
      "605/610 [============================>.] - ETA: 0s - loss: 9.4205e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 8s 13ms/step - loss: 9.4174e-05 - accuracy: 1.0000 - val_loss: 1.0578e-04 - val_accuracy: 1.0000\n",
      "Epoch 9/10\n",
      "605/610 [============================>.] - ETA: 0s - loss: 6.3606e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 8s 12ms/step - loss: 6.3824e-05 - accuracy: 1.0000 - val_loss: 7.2740e-05 - val_accuracy: 1.0000\n",
      "Epoch 10/10\n",
      "609/610 [============================>.] - ETA: 0s - loss: 4.3726e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 6s 11ms/step - loss: 4.3781e-05 - accuracy: 1.0000 - val_loss: 5.2467e-05 - val_accuracy: 1.0000\n",
      "Epoch 1/10\n",
      "610/610 [==============================] - ETA: 0s - loss: 0.9440 - accuracy: 0.7072"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 13s 20ms/step - loss: 0.9440 - accuracy: 0.7072 - val_loss: 0.1007 - val_accuracy: 0.9661\n",
      "Epoch 2/10\n",
      "609/610 [============================>.] - ETA: 0s - loss: 0.0560 - accuracy: 0.9842"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 12ms/step - loss: 0.0560 - accuracy: 0.9843 - val_loss: 0.0209 - val_accuracy: 0.9960\n",
      "Epoch 3/10\n",
      "608/610 [============================>.] - ETA: 0s - loss: 0.0023 - accuracy: 0.9999"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 12ms/step - loss: 0.0023 - accuracy: 0.9999 - val_loss: 7.2013e-04 - val_accuracy: 1.0000\n",
      "Epoch 4/10\n",
      "605/610 [============================>.] - ETA: 0s - loss: 4.6963e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 8s 13ms/step - loss: 4.6816e-04 - accuracy: 1.0000 - val_loss: 3.4561e-04 - val_accuracy: 1.0000\n",
      "Epoch 5/10\n",
      "607/610 [============================>.] - ETA: 0s - loss: 2.5383e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 8s 12ms/step - loss: 2.5440e-04 - accuracy: 1.0000 - val_loss: 2.1135e-04 - val_accuracy: 1.0000\n",
      "Epoch 6/10\n",
      "604/610 [============================>.] - ETA: 0s - loss: 1.5354e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 8s 12ms/step - loss: 1.5301e-04 - accuracy: 1.0000 - val_loss: 1.5294e-04 - val_accuracy: 1.0000\n",
      "Epoch 7/10\n",
      "605/610 [============================>.] - ETA: 0s - loss: 1.0028e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 12ms/step - loss: 1.0031e-04 - accuracy: 1.0000 - val_loss: 1.0363e-04 - val_accuracy: 1.0000\n",
      "Epoch 8/10\n",
      "605/610 [============================>.] - ETA: 0s - loss: 6.5328e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 8s 13ms/step - loss: 6.5049e-05 - accuracy: 1.0000 - val_loss: 6.1038e-05 - val_accuracy: 1.0000\n",
      "Epoch 9/10\n",
      "610/610 [==============================] - ETA: 0s - loss: 4.3536e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 7s 12ms/step - loss: 4.3536e-05 - accuracy: 1.0000 - val_loss: 4.2570e-05 - val_accuracy: 1.0000\n",
      "Epoch 10/10\n",
      "607/610 [============================>.] - ETA: 0s - loss: 2.9967e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fmaps-3\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
      "610/610 [==============================] - 7s 12ms/step - loss: 2.9911e-05 - accuracy: 1.0000 - val_loss: 3.0291e-05 - val_accuracy: 1.0000\n"
     ]
    }
   ],
   "source": [
    "# Models\n",
    "# 'models/experiment-fmaps-1'\n",
    "# 'models/experiment-fmaps-2'\n",
    "# 'models/experiment-fmaps-3'\n",
    "\n",
    "# History objects\n",
    "# 'models/experiment-fmaps-1-history'\n",
    "# 'models/experiment-fmaps-2-history'\n",
    "# 'models/experiment-fmaps-3-history'\n",
    "\n",
    "n_tests = 3\n",
    "\n",
    "models = []\n",
    "early_stopping_cb = keras.callbacks.EarlyStopping(patience=5)\n",
    "\n",
    "\n",
    "for i in range(n_tests):\n",
    "    model = keras.models.Sequential()\n",
    "    models.append(model)\n",
    "    model.add(keras.layers.Input(shape=(28, 28, 1)))\n",
    "    for pairs in range(3):\n",
    "        model.add(keras.layers.Conv2D((8*(i+1))*(2**pairs), (5, 5), padding='same', activation='relu'))\n",
    "        model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "    model.add(keras.layers.Flatten())\n",
    "    model.add(keras.layers.Dense(128, activation='relu'))\n",
    "    model.add(keras.layers.Dense(24, activation='softmax'))\n",
    "    model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "    save_best_cb = keras.callbacks.ModelCheckpoint(f'models/experiment-fmaps-{i+1}', save_best_only=True)\n",
    "    history = model.fit(X_train, y_train, epochs=10, validation_data=(X_valid, y_valid), callbacks=[save_best_cb, early_stopping_cb])\n",
    "    with open(f'models/experiment-fmaps-{i+1}-history', 'wb') as history_file:\n",
    "        pickle.dump(history.history, history_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "4c22de89",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_36\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d_88 (Conv2D)          (None, 28, 28, 8)         208       \n",
      "                                                                 \n",
      " max_pooling2d_88 (MaxPoolin  (None, 14, 14, 8)        0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " conv2d_89 (Conv2D)          (None, 14, 14, 16)        3216      \n",
      "                                                                 \n",
      " max_pooling2d_89 (MaxPoolin  (None, 7, 7, 16)         0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " conv2d_90 (Conv2D)          (None, 7, 7, 32)          12832     \n",
      "                                                                 \n",
      " max_pooling2d_90 (MaxPoolin  (None, 3, 3, 32)         0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " flatten_17 (Flatten)        (None, 288)               0         \n",
      "                                                                 \n",
      " dense_34 (Dense)            (None, 128)               36992     \n",
      "                                                                 \n",
      " dense_35 (Dense)            (None, 24)                3096      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 56,344\n",
      "Trainable params: 56,344\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "models[0].summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "27f5faf6",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_37\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d_91 (Conv2D)          (None, 28, 28, 16)        416       \n",
      "                                                                 \n",
      " max_pooling2d_91 (MaxPoolin  (None, 14, 14, 16)       0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " conv2d_92 (Conv2D)          (None, 14, 14, 32)        12832     \n",
      "                                                                 \n",
      " max_pooling2d_92 (MaxPoolin  (None, 7, 7, 32)         0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " conv2d_93 (Conv2D)          (None, 7, 7, 64)          51264     \n",
      "                                                                 \n",
      " max_pooling2d_93 (MaxPoolin  (None, 3, 3, 64)         0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " flatten_18 (Flatten)        (None, 576)               0         \n",
      "                                                                 \n",
      " dense_36 (Dense)            (None, 128)               73856     \n",
      "                                                                 \n",
      " dense_37 (Dense)            (None, 24)                3096      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 141,464\n",
      "Trainable params: 141,464\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "models[1].summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "89b7bbaf",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_38\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d_94 (Conv2D)          (None, 28, 28, 24)        624       \n",
      "                                                                 \n",
      " max_pooling2d_94 (MaxPoolin  (None, 14, 14, 24)       0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " conv2d_95 (Conv2D)          (None, 14, 14, 48)        28848     \n",
      "                                                                 \n",
      " max_pooling2d_95 (MaxPoolin  (None, 7, 7, 48)         0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " conv2d_96 (Conv2D)          (None, 7, 7, 96)          115296    \n",
      "                                                                 \n",
      " max_pooling2d_96 (MaxPoolin  (None, 3, 3, 96)         0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " flatten_19 (Flatten)        (None, 864)               0         \n",
      "                                                                 \n",
      " dense_38 (Dense)            (None, 128)               110720    \n",
      "                                                                 \n",
      " dense_39 (Dense)            (None, 24)                3096      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 258,584\n",
      "Trainable params: 258,584\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "models[2].summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "d7b2bf2d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "249/249 [==============================] - 1s 5ms/step - loss: 1.1253e-04 - accuracy: 1.0000\n",
      "249/249 [==============================] - 1s 5ms/step - loss: 5.2467e-05 - accuracy: 1.0000\n",
      "249/249 [==============================] - 1s 5ms/step - loss: 3.0291e-05 - accuracy: 1.0000\n"
     ]
    }
   ],
   "source": [
    "for index in range(n_tests):\n",
    "    model = keras.models.load_model(f'models/experiment-fmaps-{index+1}')\n",
    "    model.evaluate(X_valid, y_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "c6af5d12",
   "metadata": {},
   "outputs": [],
   "source": [
    "h_2_1 = np.load('models/experiment-fmaps-1-history', allow_pickle=True)\n",
    "h_2_2 = np.load('models/experiment-fmaps-2-history', allow_pickle=True)\n",
    "h_2_3 = np.load('models/experiment-fmaps-3-history', allow_pickle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "dac13522",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "get_train_val_plots(h_2_1, yticks=np.arange(0, 1.2, 0.1))\n",
    "get_train_val_plots(h_2_2, yticks=np.arange(0, 1.2, 0.1))\n",
    "get_train_val_plots(h_2_3, yticks=np.arange(0, 1.2, 0.1))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "74dc1a09",
   "metadata": {},
   "source": [
    "Choose Model 3"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "68a857fd",
   "metadata": {},
   "source": [
    "## Filter Size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "4ca42854",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "782/782 [==============================] - ETA: 0s - loss: 0.9397 - accuracy: 0.7093"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 46s 18ms/step - loss: 0.9397 - accuracy: 0.7093 - val_loss: 0.0811 - val_accuracy: 0.9841\n",
      "Epoch 2/10\n",
      "782/782 [==============================] - ETA: 0s - loss: 0.0391 - accuracy: 0.9906"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 11s 14ms/step - loss: 0.0391 - accuracy: 0.9906 - val_loss: 0.0061 - val_accuracy: 0.9992\n",
      "Epoch 3/10\n",
      "782/782 [==============================] - 8s 10ms/step - loss: 0.0023 - accuracy: 0.9998 - val_loss: 0.0307 - val_accuracy: 0.9910\n",
      "Epoch 4/10\n",
      "779/782 [============================>.] - ETA: 0s - loss: 0.0295 - accuracy: 0.9917"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 10s 13ms/step - loss: 0.0294 - accuracy: 0.9917 - val_loss: 7.3830e-04 - val_accuracy: 1.0000\n",
      "Epoch 5/10\n",
      "780/782 [============================>.] - ETA: 0s - loss: 4.3767e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 10s 13ms/step - loss: 4.3740e-04 - accuracy: 1.0000 - val_loss: 3.3426e-04 - val_accuracy: 1.0000\n",
      "Epoch 6/10\n",
      "776/782 [============================>.] - ETA: 0s - loss: 2.0925e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 10s 13ms/step - loss: 2.0886e-04 - accuracy: 1.0000 - val_loss: 1.9949e-04 - val_accuracy: 1.0000\n",
      "Epoch 7/10\n",
      "776/782 [============================>.] - ETA: 0s - loss: 1.1790e-04 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 10s 13ms/step - loss: 1.1779e-04 - accuracy: 1.0000 - val_loss: 1.2782e-04 - val_accuracy: 1.0000\n",
      "Epoch 8/10\n",
      "779/782 [============================>.] - ETA: 0s - loss: 7.4937e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 10s 13ms/step - loss: 7.4947e-05 - accuracy: 1.0000 - val_loss: 7.9334e-05 - val_accuracy: 1.0000\n",
      "Epoch 9/10\n",
      "782/782 [==============================] - ETA: 0s - loss: 4.5664e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "782/782 [==============================] - 12s 16ms/step - loss: 4.5664e-05 - accuracy: 1.0000 - val_loss: 4.8071e-05 - val_accuracy: 1.0000\n",
      "Epoch 10/10\n",
      "780/782 [============================>.] - ETA: 0s - loss: 2.7893e-05 - accuracy: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-fiters-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
      "782/782 [==============================] - 10s 13ms/step - loss: 2.7887e-05 - accuracy: 1.0000 - val_loss: 3.0051e-05 - val_accuracy: 1.0000\n"
     ]
    }
   ],
   "source": [
    "# Models\n",
    "# 'models/experiment-fiters-1'\n",
    "\n",
    "# History objects\n",
    "# 'models/experiment-filters-1-history'\n",
    "\n",
    "save_best_cb = keras.callbacks.ModelCheckpoint(f'models/experiment-fiters-1', save_best_only=True)\n",
    "early_stopping_cb = keras.callbacks.EarlyStopping(patience=5)\n",
    "\n",
    "model = keras.models.Sequential()\n",
    "model.add(keras.layers.Conv2D(24, (3, 3), padding='same', activation='relu'))\n",
    "model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(keras.layers.Conv2D(48, (3, 3), padding='same', activation='relu'))\n",
    "model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(keras.layers.Conv2D(96, (3, 3), padding='same', activation='relu'))\n",
    "model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(keras.layers.Flatten())\n",
    "model.add(keras.layers.Dense(128, activation='relu'))\n",
    "model.add(keras.layers.Dense(24, activation='softmax'))\n",
    "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "history = model.fit(X_train, y_train, epochs=10, validation_data=(X_valid, y_valid), callbacks=[save_best_cb, early_stopping_cb])\n",
    "with open(f'models/experiment-filters-1-history', 'wb') as history_file:\n",
    "    pickle.dump(history.history, history_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "816df03f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "249/249 [==============================] - 2s 7ms/step - loss: 2.3939e-05 - accuracy: 1.0000\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[2.393894646957051e-05, 1.0]"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = keras.models.load_model('models/experiment-fiters-1/')\n",
    "model.evaluate(X_valid, y_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "1163b70d",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "h_2_3 = np.load('models/experiment-fmaps-3-history', allow_pickle=True)\n",
    "h = np.load('models/experiment-filters-1-history', allow_pickle=True)\n",
    "get_train_val_plots(h, yticks=np.arange(0, 1.2, 0.1))\n",
    "get_train_val_plots(h_2_3, yticks=np.arange(0, 1.2, 0.1))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "41197ef4",
   "metadata": {},
   "source": [
    "We choose to use filter sizes of (5x5)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d6ab0734",
   "metadata": {},
   "source": [
    "## Dropout"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "39ea34d6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "610/610 [==============================] - ETA: 0s - loss: 2.0785 - accuracy: 0.3484"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 22s 20ms/step - loss: 2.0785 - accuracy: 0.3484 - val_loss: 0.4270 - val_accuracy: 0.8683\n",
      "Epoch 2/10\n",
      "609/610 [============================>.] - ETA: 0s - loss: 0.4214 - accuracy: 0.8547"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 10s 17ms/step - loss: 0.4212 - accuracy: 0.8548 - val_loss: 0.0518 - val_accuracy: 0.9961\n",
      "Epoch 3/10\n",
      "610/610 [==============================] - ETA: 0s - loss: 0.1852 - accuracy: 0.9407"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 11s 19ms/step - loss: 0.1852 - accuracy: 0.9407 - val_loss: 0.0151 - val_accuracy: 0.9971\n",
      "Epoch 4/10\n",
      "607/610 [============================>.] - ETA: 0s - loss: 0.1114 - accuracy: 0.9627"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 12s 19ms/step - loss: 0.1115 - accuracy: 0.9626 - val_loss: 0.0081 - val_accuracy: 0.9992\n",
      "Epoch 5/10\n",
      "610/610 [==============================] - ETA: 0s - loss: 0.0822 - accuracy: 0.9737"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 11s 18ms/step - loss: 0.0822 - accuracy: 0.9737 - val_loss: 0.0020 - val_accuracy: 0.9999\n",
      "Epoch 6/10\n",
      "607/610 [============================>.] - ETA: 0s - loss: 0.0685 - accuracy: 0.9777"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 11s 18ms/step - loss: 0.0688 - accuracy: 0.9776 - val_loss: 0.0016 - val_accuracy: 1.0000\n",
      "Epoch 7/10\n",
      "604/610 [============================>.] - ETA: 0s - loss: 0.0576 - accuracy: 0.9817"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 10s 17ms/step - loss: 0.0579 - accuracy: 0.9817 - val_loss: 5.7927e-04 - val_accuracy: 1.0000\n",
      "Epoch 8/10\n",
      "606/610 [============================>.] - ETA: 0s - loss: 0.0561 - accuracy: 0.9811"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-0\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 10s 16ms/step - loss: 0.0560 - accuracy: 0.9812 - val_loss: 3.6957e-04 - val_accuracy: 1.0000\n",
      "Epoch 9/10\n",
      "610/610 [==============================] - 7s 11ms/step - loss: 0.0523 - accuracy: 0.9824 - val_loss: 0.0011 - val_accuracy: 0.9999\n",
      "Epoch 10/10\n",
      "610/610 [==============================] - 6s 11ms/step - loss: 0.0512 - accuracy: 0.9834 - val_loss: 3.9864e-04 - val_accuracy: 1.0000\n",
      "Epoch 1/10\n",
      "606/610 [============================>.] - ETA: 0s - loss: 2.3799 - accuracy: 0.2601"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 10s 15ms/step - loss: 2.3726 - accuracy: 0.2620 - val_loss: 0.7182 - val_accuracy: 0.7989\n",
      "Epoch 2/10\n",
      "607/610 [============================>.] - ETA: 0s - loss: 0.7228 - accuracy: 0.7477"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 13s 21ms/step - loss: 0.7215 - accuracy: 0.7481 - val_loss: 0.1625 - val_accuracy: 0.9703\n",
      "Epoch 3/10\n",
      "609/610 [============================>.] - ETA: 0s - loss: 0.3544 - accuracy: 0.8783"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 11s 18ms/step - loss: 0.3548 - accuracy: 0.8783 - val_loss: 0.0479 - val_accuracy: 0.9906\n",
      "Epoch 4/10\n",
      "608/610 [============================>.] - ETA: 0s - loss: 0.2359 - accuracy: 0.9175"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 12s 20ms/step - loss: 0.2357 - accuracy: 0.9175 - val_loss: 0.0210 - val_accuracy: 0.9975\n",
      "Epoch 5/10\n",
      "609/610 [============================>.] - ETA: 0s - loss: 0.1672 - accuracy: 0.9420"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 10s 16ms/step - loss: 0.1672 - accuracy: 0.9420 - val_loss: 0.0074 - val_accuracy: 0.9989\n",
      "Epoch 6/10\n",
      "609/610 [============================>.] - ETA: 0s - loss: 0.1402 - accuracy: 0.9527"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 9s 15ms/step - loss: 0.1402 - accuracy: 0.9528 - val_loss: 0.0044 - val_accuracy: 0.9999\n",
      "Epoch 7/10\n",
      "608/610 [============================>.] - ETA: 0s - loss: 0.1163 - accuracy: 0.9603"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 9s 15ms/step - loss: 0.1161 - accuracy: 0.9604 - val_loss: 0.0039 - val_accuracy: 0.9999\n",
      "Epoch 8/10\n",
      "610/610 [==============================] - ETA: 0s - loss: 0.1072 - accuracy: 0.9634"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 9s 15ms/step - loss: 0.1072 - accuracy: 0.9634 - val_loss: 0.0021 - val_accuracy: 0.9999\n",
      "Epoch 9/10\n",
      "609/610 [============================>.] - ETA: 0s - loss: 0.0953 - accuracy: 0.9693"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 9s 15ms/step - loss: 0.0953 - accuracy: 0.9693 - val_loss: 9.3107e-04 - val_accuracy: 1.0000\n",
      "Epoch 10/10\n",
      "604/610 [============================>.] - ETA: 0s - loss: 0.0882 - accuracy: 0.9720"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-1\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 9s 15ms/step - loss: 0.0887 - accuracy: 0.9720 - val_loss: 5.2711e-04 - val_accuracy: 1.0000\n",
      "Epoch 1/10\n",
      "607/610 [============================>.] - ETA: 0s - loss: 2.9102 - accuracy: 0.1143"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 11s 16ms/step - loss: 2.9072 - accuracy: 0.1150 - val_loss: 1.6103 - val_accuracy: 0.5564\n",
      "Epoch 2/10\n",
      "605/610 [============================>.] - ETA: 0s - loss: 1.2602 - accuracy: 0.5728"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 10s 16ms/step - loss: 1.2587 - accuracy: 0.5734 - val_loss: 0.4016 - val_accuracy: 0.9031\n",
      "Epoch 3/10\n",
      "609/610 [============================>.] - ETA: 0s - loss: 0.7070 - accuracy: 0.7565"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 9s 15ms/step - loss: 0.7069 - accuracy: 0.7565 - val_loss: 0.2071 - val_accuracy: 0.9622\n",
      "Epoch 4/10\n",
      "609/610 [============================>.] - ETA: 0s - loss: 0.5019 - accuracy: 0.8265"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 10s 16ms/step - loss: 0.5019 - accuracy: 0.8265 - val_loss: 0.1040 - val_accuracy: 0.9774\n",
      "Epoch 5/10\n",
      "609/610 [============================>.] - ETA: 0s - loss: 0.3981 - accuracy: 0.8618"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 9s 15ms/step - loss: 0.3982 - accuracy: 0.8617 - val_loss: 0.0525 - val_accuracy: 0.9950\n",
      "Epoch 6/10\n",
      "607/610 [============================>.] - ETA: 0s - loss: 0.3182 - accuracy: 0.8896"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 11s 18ms/step - loss: 0.3180 - accuracy: 0.8896 - val_loss: 0.0371 - val_accuracy: 0.9937\n",
      "Epoch 7/10\n",
      "609/610 [============================>.] - ETA: 0s - loss: 0.2721 - accuracy: 0.9069"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 10s 16ms/step - loss: 0.2720 - accuracy: 0.9070 - val_loss: 0.0245 - val_accuracy: 0.9972\n",
      "Epoch 8/10\n",
      "610/610 [==============================] - ETA: 0s - loss: 0.2528 - accuracy: 0.9168"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 10s 16ms/step - loss: 0.2528 - accuracy: 0.9168 - val_loss: 0.0124 - val_accuracy: 0.9987\n",
      "Epoch 9/10\n",
      "610/610 [==============================] - ETA: 0s - loss: 0.2299 - accuracy: 0.9224"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "610/610 [==============================] - 10s 16ms/step - loss: 0.2299 - accuracy: 0.9224 - val_loss: 0.0106 - val_accuracy: 0.9997\n",
      "Epoch 10/10\n",
      "610/610 [==============================] - ETA: 0s - loss: 0.2004 - accuracy: 0.9347"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: models\\experiment-dropout-2\\assets\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r",
      "610/610 [==============================] - 9s 15ms/step - loss: 0.2004 - accuracy: 0.9347 - val_loss: 0.0070 - val_accuracy: 0.9992\n"
     ]
    }
   ],
   "source": [
    "# Models\n",
    "# 'models/experiment-dropout-0'\n",
    "# 'models/experiment-dropout-1'\n",
    "# 'models/experiment-dropout-2'\n",
    "\n",
    "# History objects\n",
    "# 'models/experiment-dropout-0-history'\n",
    "# 'models/experiment-dropout-1-history'\n",
    "# 'models/experiment-dropout-2-history'\n",
    "\n",
    "early_stopping_cb = keras.callbacks.EarlyStopping(patience=5)\n",
    "\n",
    "dropout_rates = [0.3, 0.4, 0.5]\n",
    "\n",
    "for index, i in enumerate(dropout_rates):\n",
    "    model = keras.models.Sequential()\n",
    "    model.add(keras.layers.Conv2D(24, (5, 5), padding='same', activation='relu'))\n",
    "    model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "    model.add(keras.layers.Dropout(i))\n",
    "    model.add(keras.layers.Conv2D(48, (5, 5), padding='same', activation='relu'))\n",
    "    model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "    model.add(keras.layers.Dropout(i))\n",
    "    model.add(keras.layers.Conv2D(96, (5, 5), padding='same', activation='relu'))\n",
    "    model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
    "    model.add(keras.layers.Dropout(i))\n",
    "    model.add(keras.layers.Flatten())\n",
    "    model.add(keras.layers.Dense(128, activation='relu'))\n",
    "    model.add(keras.layers.Dropout(i))\n",
    "    model.add(keras.layers.Dense(24, activation='softmax'))\n",
    "    model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "    save_best_cb = keras.callbacks.ModelCheckpoint(f'models/experiment-dropout-{index}', save_best_only=True)\n",
    "    history = model.fit(X_train, y_train, epochs=10, validation_data=(X_valid, y_valid), callbacks=[save_best_cb, early_stopping_cb])\n",
    "    with open(f'models/experiment-dropout-{index}-history', 'wb') as history_file:\n",
    "        pickle.dump(history.history, history_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "97243155",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "249/249 [==============================] - 1s 4ms/step - loss: 3.6957e-04 - accuracy: 1.0000\n",
      "249/249 [==============================] - 1s 4ms/step - loss: 5.2711e-04 - accuracy: 1.0000\n",
      "249/249 [==============================] - 1s 3ms/step - loss: 0.0070 - accuracy: 0.9992\n"
     ]
    }
   ],
   "source": [
    "for index in range(3):\n",
    "    model = keras.models.load_model(f'models/experiment-dropout-{index}')\n",
    "    model.evaluate(X_valid, y_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "e3b5e4d1",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "h_0 = np.load('models/experiment-dropout-0-history', allow_pickle=True)\n",
    "h_1 = np.load('models/experiment-dropout-1-history', allow_pickle=True)\n",
    "h_2 = np.load('models/experiment-dropout-2-history', allow_pickle=True)\n",
    "\n",
    "get_train_val_plots(h_0)\n",
    "get_train_val_plots(h_1)\n",
    "get_train_val_plots(h_2)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "12fda6b1",
   "metadata": {},
   "source": [
    "Select Model 1"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2a324af8",
   "metadata": {},
   "source": [
    "## Data Augmentation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a8248a34",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Sample Code\n",
    "\n",
    "data_augmentation = keras.models.Sequential()\n",
    "data_augmentation.add(keras.layers.RandomRotation(0.1, fill_mode='nearest', input_shape=(28, 28, 1)))\n",
    "data_augmentation.add(keras.layers.RandomZoom((0.15, 0.2), fill_mode='nearest'))\n",
    "data_augmentation.add(keras.layers.RandomTranslation(0.1, 0.1, fill_mode='nearest'))\n",
    "\n",
    "model = keras.models.Sequential()\n",
    "# Add the layers from the above model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4d16aecf",
   "metadata": {},
   "source": [
    "Improves the models performance by decreasing the tendency to overfit"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5a43ebe6",
   "metadata": {},
   "source": [
    "## Other Hyperparameters to try\n",
    "\n",
    "1. Batch Normalization - It normalizes the layer inputs\n",
    "2. Deeper networks work well - Replacing the single convolution layer of filter size (5X5) with two successive consecutive convolution layers of filter size (3X3)\n",
    "3. Number of units in the dense layer and number of dense layers\n",
    "4. Replacing the MaxPooling Layer with a convolution layer having a stride > 1\n",
    "5. Optimizers\n",
    "6. Learning rate of the optimizer"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec6be927",
   "metadata": {},
   "source": [
    "## Visualising the final model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "753b2d78",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best_model = keras.models.load_model('models/experiment-dropout-0/')\n",
    "plot_model(best_model, to_file='model.png', show_shapes=True, show_layer_names=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b04bca5e",
   "metadata": {},
   "source": [
    "# Performance on the Test Set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "90fee9da",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_df = pd.read_csv('data/alphabet/sign_mnist_test.csv')\n",
    "X_test, y_test = test_df.drop('label', axis=1), test_df['label']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "588e05b8",
   "metadata": {},
   "outputs": [],
   "source": [
    "best_model = keras.models.load_model('models/experiment-dropout-0/')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "666818aa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "225/225 [==============================] - 25s 5ms/step - loss: 29.5216 - accuracy: 0.9631\n",
      "Loss: 29.522 Accuracy: 0.963\n"
     ]
    }
   ],
   "source": [
    "evaluate_model(best_model, X_test, y_test, label_binarizer)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a8090486",
   "metadata": {},
   "source": [
    "Accuracy: 96%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "82643927",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_df = pd.read_csv('data/alphabet/sign_mnist_test.csv')\n",
    "X_test, y_test = test_df.drop('label', axis=1), test_df['label']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "5d02d431",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Applying normalisation which is applied for X_train\n",
    "X_test /= 255.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "7332b757",
   "metadata": {},
   "outputs": [],
   "source": [
    "best_model = keras.models.load_model('models/experiment-dropout-0/')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "83bb851a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "225/225 [==============================] - 1s 4ms/step - loss: 0.1247 - accuracy: 0.9632\n",
      "Loss: 0.125 Accuracy: 0.963\n"
     ]
    }
   ],
   "source": [
    "evaluate_model(best_model, X_test, y_test, label_binarizer)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "97e05057",
   "metadata": {},
   "source": [
    "Accuracy: 96%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6d8a4735",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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
}