nn_model_building_evaluation_and_prediction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Introduction\n",
    "In this notebook, we'll build a ConvNet model to classify the images as either:\n",
    "* Rs 1\n",
    "* Rs 2\n",
    "* Rs 5\n",
    "* Reverse face of the coin\n",
    "* Inavlid/fake coin\n",
    "\n",
    "Note, this classification is in accordance to the classes present in the dataset.\n",
    "# Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2023-05-08T15:52:23.694710Z",
     "iopub.status.busy": "2023-05-08T15:52:23.694123Z",
     "iopub.status.idle": "2023-05-08T15:52:30.523639Z",
     "shell.execute_reply": "2023-05-08T15:52:30.522488Z",
     "shell.execute_reply.started": "2023-05-08T15:52:23.694676Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.10/site-packages/scipy/__init__.py:146: UserWarning: A NumPy version >=1.16.5 and <1.23.0 is required for this version of SciPy (detected version 1.23.5\n",
      "  warnings.warn(f\"A NumPy version >={np_minversion} and <{np_maxversion}\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensorflow Version:  2.11.0\n"
     ]
    }
   ],
   "source": [
    "import numpy as np # linear algebra\n",
    "import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n",
    "import os, warnings\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib import gridspec\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras import layers\n",
    "from tensorflow.keras.preprocessing import image_dataset_from_directory\n",
    "\n",
    "print('Tensorflow Version: ', tf.__version__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-08T15:52:30.526010Z",
     "iopub.status.busy": "2023-05-08T15:52:30.525328Z",
     "iopub.status.idle": "2023-05-08T15:52:30.532790Z",
     "shell.execute_reply": "2023-05-08T15:52:30.531929Z",
     "shell.execute_reply.started": "2023-05-08T15:52:30.525975Z"
    }
   },
   "outputs": [],
   "source": [
    "# Set path to the images folder\n",
    "\n",
    "path = 'images/processed'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-08T15:52:34.286903Z",
     "iopub.status.busy": "2023-05-08T15:52:34.286570Z",
     "iopub.status.idle": "2023-05-08T15:52:34.292732Z",
     "shell.execute_reply": "2023-05-08T15:52:34.291789Z",
     "shell.execute_reply.started": "2023-05-08T15:52:34.286876Z"
    }
   },
   "outputs": [],
   "source": [
    "# Ensure reproducability across different runs of this program.\n",
    "\n",
    "def set_seed(seed=2000): # set seed to an arbitrary integer.\n",
    "    \n",
    "    # Set seed value for the random number generators of Numpy and Tensorflow.\n",
    "    # This is to ensure same sequence of random nos. are generated by numpy and Tensorflow\n",
    "    # everytime this program is run.\n",
    "    # Same seed value everytime leads to same set of random nos. generated everytime.\n",
    "    np.random.seed(seed)\n",
    "    tf.random.set_seed(seed)\n",
    "    \n",
    "    # Set environment variable 'PYTHONHASHSEED' to the seed value.\n",
    "    # This is to ensure same hash codes are generated for accesing values with sets and\n",
    "    # dictionaries irrespective of the order in which the values are accesed.\n",
    "    os.environ['PYTHONHASHSEED'] = str(seed)\n",
    "    \n",
    "    # Set environment variable 'TF_DETERMINISTIC_OPS' to 1.\n",
    "    # This is to ensure same sequence of  are generated for accesing values with sets and\n",
    "    # dictionaries irrespective of the order in which the values are accesed.\n",
    "    os.environ['TF_DETERMINISTIC_OPS'] = '1'\n",
    "\n",
    "# Run the set_seed function to set the seed values and environment variables.\n",
    "set_seed()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-08T15:52:36.138007Z",
     "iopub.status.busy": "2023-05-08T15:52:36.137670Z",
     "iopub.status.idle": "2023-05-08T15:52:36.143998Z",
     "shell.execute_reply": "2023-05-08T15:52:36.142947Z",
     "shell.execute_reply.started": "2023-05-08T15:52:36.137979Z"
    }
   },
   "outputs": [],
   "source": [
    "# Set Matplotlib defaults\n",
    "\n",
    "plt.rc('figure', autolayout=True)\n",
    "plt.rc('axes', labelweight='bold', labelsize='medium',\n",
    "       titleweight='bold', titlesize=13, titlepad=10,\n",
    "      )\n",
    "#warnings.filterwarnings(\"ignore\") # to clean up output cells"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Create Training, Validation and Test Sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-08T15:52:49.643533Z",
     "iopub.status.busy": "2023-05-08T15:52:49.643171Z",
     "iopub.status.idle": "2023-05-08T15:52:53.639637Z",
     "shell.execute_reply": "2023-05-08T15:52:53.638618Z",
     "shell.execute_reply.started": "2023-05-08T15:52:49.643504Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 1750 files belonging to 5 classes.\n",
      "Using 1225 files for training.\n",
      "Found 1750 files belonging to 5 classes.\n",
      "Using 525 files for validation.\n",
      "Class names:  ['1_rupee', '2_rupee', '5_rupee', 'invalid', 'reverse']\n"
     ]
    }
   ],
   "source": [
    "# Parameters\n",
    "image_size = [256, 256] # For resizing images. 256px X 256px is same as the size of the images in the datset.\n",
    "val_test_split = 0.3 # Validation and test sets together take 30% of the data\n",
    "batch_size=64 # No. of samples in each batch\n",
    "seed=2000 # Seed value is a arbitary non-negative integer.\n",
    "\n",
    "# train set\n",
    "ds_train_ = image_dataset_from_directory(\n",
    "    path, # The path to the dataset\n",
    "    validation_split=val_test_split,\n",
    "    subset='training',\n",
    "    labels='inferred', # Labels infered from folder names\n",
    "    label_mode='categorical',\n",
    "    image_size=image_size,\n",
    "    interpolation='nearest', # Algorithm used for resizing images.\n",
    "    batch_size=batch_size,\n",
    "    color_mode='grayscale', # color mode of images in the dataset.\n",
    "    shuffle=True, # This will ensure shuffling of the set randomly before each epoch of training.\n",
    "    seed=seed\n",
    ")\n",
    "\n",
    "# validation set\n",
    "ds_valid_ = image_dataset_from_directory(\n",
    "    path,\n",
    "    validation_split=val_test_split,\n",
    "    subset='validation',\n",
    "    labels='inferred',\n",
    "    label_mode='categorical',\n",
    "    image_size=image_size,\n",
    "    interpolation='nearest',\n",
    "    batch_size=batch_size,\n",
    "    color_mode='grayscale',\n",
    "    shuffle=True, # Validation set shouldn't be shuffled.\n",
    "    seed=seed\n",
    ")\n",
    "\n",
    "# Get the number of batches available in the validation set.\n",
    "val_batches = tf.data.experimental.cardinality(ds_valid_)\n",
    "\n",
    "# Takeaway 50% of samples in ds_valid_ for test set.\n",
    "ds_test_ = ds_valid_.take(val_batches // 2)\n",
    "ds_valid_ = ds_valid_.skip(val_batches // 2)\n",
    "\n",
    "# Get class names\n",
    "class_names = ds_train_.class_names\n",
    "print('Class names: ', class_names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "_kg_hide-output": false,
    "execution": {
     "iopub.execute_input": "2023-05-08T15:52:53.641967Z",
     "iopub.status.busy": "2023-05-08T15:52:53.641643Z",
     "iopub.status.idle": "2023-05-08T15:52:56.617949Z",
     "shell.execute_reply": "2023-05-08T15:52:56.616924Z",
     "shell.execute_reply.started": "2023-05-08T15:52:53.641933Z"
    }
   },
   "outputs": [],
   "source": [
    "# Define a function to extract the labels from the test set\n",
    "def get_labels(image, label):\n",
    "    return label\n",
    "\n",
    "# Extract the labels from the dataset\n",
    "true_test_labels = ds_test_.map(get_labels)\n",
    "true_test_labels = np.concatenate(list(true_test_labels.as_numpy_iterator()))\n",
    "true_test_labels = np.argmax(true_test_labels, axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Setup the Data Pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-07T12:47:31.277762Z",
     "iopub.status.busy": "2023-05-07T12:47:31.277404Z",
     "iopub.status.idle": "2023-05-07T12:47:31.347189Z",
     "shell.execute_reply": "2023-05-07T12:47:31.346359Z",
     "shell.execute_reply.started": "2023-05-07T12:47:31.277732Z"
    }
   },
   "outputs": [],
   "source": [
    "# Convert the image tensor to float32 tensor \n",
    "def convert_to_float(image, label):\n",
    "    image = tf.image.convert_image_dtype(image, dtype=tf.float32)\n",
    "    return image, label\n",
    "\n",
    "AUTOTUNE = tf.data.experimental.AUTOTUNE\n",
    "# AUTOTUNE enables automatic tuning of prefetch buffer size dynamically during runtime\n",
    "# Such a tuned buffer size ensures proper allocation of hardware resources for optimal performance.\n",
    "\n",
    "ds_train = (\n",
    "    ds_train_\n",
    "    .map(convert_to_float) # Applies convert_to_float function to each image tensor\n",
    "    .cache() # Caches the dataset in memory to improve performance during training.\n",
    "    .prefetch(buffer_size=AUTOTUNE) # Prefetches dataset elements in the background during training.\n",
    "                                    # buffer_size specifies max. no. of elements to prefetch.\n",
    ")\n",
    "\n",
    "ds_valid = (\n",
    "    ds_valid_\n",
    "    .map(convert_to_float)\n",
    "    .cache()\n",
    "    .prefetch(buffer_size=AUTOTUNE)\n",
    ")\n",
    "\n",
    "ds_test = (\n",
    "    ds_test_\n",
    "    .map(convert_to_float)\n",
    "    .cache()\n",
    "    .prefetch(buffer_size=AUTOTUNE)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-07T12:48:16.248433Z",
     "iopub.status.busy": "2023-05-07T12:48:16.247766Z",
     "iopub.status.idle": "2023-05-07T12:48:16.253041Z",
     "shell.execute_reply": "2023-05-07T12:48:16.251951Z",
     "shell.execute_reply.started": "2023-05-07T12:48:16.248398Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<PrefetchDataset element_spec=(TensorSpec(shape=(None, 256, 256, 1), dtype=tf.float32, name=None), TensorSpec(shape=(None, 5), dtype=tf.float32, name=None))>\n"
     ]
    }
   ],
   "source": [
    "print(ds_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Define the NN Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-07T10:34:17.417142Z",
     "iopub.status.busy": "2023-05-07T10:34:17.416744Z",
     "iopub.status.idle": "2023-05-07T10:34:17.543639Z",
     "shell.execute_reply": "2023-05-07T10:34:17.542953Z",
     "shell.execute_reply.started": "2023-05-07T10:34:17.417106Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d (Conv2D)             (None, 256, 256, 32)      832       \n",
      "                                                                 \n",
      " max_pooling2d (MaxPooling2D  (None, 128, 128, 32)     0         \n",
      " )                                                               \n",
      "                                                                 \n",
      " conv2d_1 (Conv2D)           (None, 128, 128, 64)      18496     \n",
      "                                                                 \n",
      " max_pooling2d_1 (MaxPooling  (None, 64, 64, 64)       0         \n",
      " 2D)                                                             \n",
      "                                                                 \n",
      " conv2d_2 (Conv2D)           (None, 64, 64, 128)       73856     \n",
      "                                                                 \n",
      " max_pooling2d_2 (MaxPooling  (None, 32, 32, 128)      0         \n",
      " 2D)                                                             \n",
      "                                                                 \n",
      " flatten (Flatten)           (None, 131072)            0         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 128)               16777344  \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 5)                 645       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 16,871,173\n",
      "Trainable params: 16,871,173\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = keras.Sequential([\n",
    "    layers.Conv2D(filters=32,\n",
    "                  kernel_size=5,\n",
    "                  activation=\"relu\",\n",
    "                  padding='same',\n",
    "                  input_shape=[256, 256, 1]),\n",
    "    layers.MaxPool2D(),\n",
    "    layers.Conv2D(filters=64, kernel_size=3, activation=\"relu\", padding='same'),\n",
    "    layers.MaxPool2D(),\n",
    "    layers.Conv2D(filters=128, kernel_size=3, activation=\"relu\", padding='same'),\n",
    "    layers.MaxPool2D(),\n",
    "    layers.Flatten(),\n",
    "    layers.Dense(units=128, activation=\"relu\"),\n",
    "    layers.Dense(units=5, activation=\"softmax\"),\n",
    "])\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Compile, Build and Evaluate the Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-07T10:34:17.545168Z",
     "iopub.status.busy": "2023-05-07T10:34:17.544768Z",
     "iopub.status.idle": "2023-05-07T10:37:17.967586Z",
     "shell.execute_reply": "2023-05-07T10:37:17.966697Z",
     "shell.execute_reply.started": "2023-05-07T10:34:17.545135Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "20/20 [==============================] - 12s 309ms/step - loss: 1.5991 - accuracy: 0.2392 - val_loss: 1.5703 - val_accuracy: 0.2825\n",
      "Epoch 2/50\n",
      "20/20 [==============================] - 3s 156ms/step - loss: 1.5462 - accuracy: 0.3404 - val_loss: 1.4900 - val_accuracy: 0.3829\n",
      "Epoch 3/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 1.4290 - accuracy: 0.4327 - val_loss: 1.4067 - val_accuracy: 0.3978\n",
      "Epoch 4/50\n",
      "20/20 [==============================] - 3s 154ms/step - loss: 1.2963 - accuracy: 0.4735 - val_loss: 1.2967 - val_accuracy: 0.4238\n",
      "Epoch 5/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 1.1506 - accuracy: 0.5371 - val_loss: 1.0218 - val_accuracy: 0.6059\n",
      "Epoch 6/50\n",
      "20/20 [==============================] - 3s 157ms/step - loss: 0.9735 - accuracy: 0.6253 - val_loss: 0.9935 - val_accuracy: 0.5874\n",
      "Epoch 7/50\n",
      "20/20 [==============================] - 3s 156ms/step - loss: 0.8770 - accuracy: 0.6702 - val_loss: 0.8100 - val_accuracy: 0.6952\n",
      "Epoch 8/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 0.6737 - accuracy: 0.7714 - val_loss: 0.5954 - val_accuracy: 0.8216\n",
      "Epoch 9/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 0.5402 - accuracy: 0.8220 - val_loss: 0.5288 - val_accuracy: 0.8476\n",
      "Epoch 10/50\n",
      "20/20 [==============================] - 3s 156ms/step - loss: 0.4386 - accuracy: 0.8580 - val_loss: 0.5101 - val_accuracy: 0.8401\n",
      "Epoch 11/50\n",
      "20/20 [==============================] - 3s 154ms/step - loss: 0.3565 - accuracy: 0.8931 - val_loss: 0.5014 - val_accuracy: 0.8587\n",
      "Epoch 12/50\n",
      "20/20 [==============================] - 3s 154ms/step - loss: 0.2695 - accuracy: 0.9273 - val_loss: 0.4606 - val_accuracy: 0.8736\n",
      "Epoch 13/50\n",
      "20/20 [==============================] - 3s 156ms/step - loss: 0.2025 - accuracy: 0.9494 - val_loss: 0.5642 - val_accuracy: 0.8364\n",
      "Epoch 14/50\n",
      "20/20 [==============================] - 3s 159ms/step - loss: 0.1937 - accuracy: 0.9527 - val_loss: 0.2928 - val_accuracy: 0.9182\n",
      "Epoch 15/50\n",
      "20/20 [==============================] - 3s 153ms/step - loss: 0.1221 - accuracy: 0.9739 - val_loss: 0.1423 - val_accuracy: 0.9703\n",
      "Epoch 16/50\n",
      "20/20 [==============================] - 3s 156ms/step - loss: 0.0826 - accuracy: 0.9861 - val_loss: 0.1251 - val_accuracy: 0.9740\n",
      "Epoch 17/50\n",
      "20/20 [==============================] - 3s 154ms/step - loss: 0.0744 - accuracy: 0.9861 - val_loss: 0.1184 - val_accuracy: 0.9665\n",
      "Epoch 18/50\n",
      "20/20 [==============================] - 3s 156ms/step - loss: 0.0526 - accuracy: 0.9918 - val_loss: 0.1007 - val_accuracy: 0.9777\n",
      "Epoch 19/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 0.0543 - accuracy: 0.9886 - val_loss: 0.1289 - val_accuracy: 0.9628\n",
      "Epoch 20/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 0.0496 - accuracy: 0.9861 - val_loss: 0.0928 - val_accuracy: 0.9703\n",
      "Epoch 21/50\n",
      "20/20 [==============================] - 3s 154ms/step - loss: 0.0423 - accuracy: 0.9935 - val_loss: 0.1033 - val_accuracy: 0.9665\n",
      "Epoch 22/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 0.0460 - accuracy: 0.9853 - val_loss: 0.1217 - val_accuracy: 0.9628\n",
      "Epoch 23/50\n",
      "20/20 [==============================] - 3s 154ms/step - loss: 0.0681 - accuracy: 0.9780 - val_loss: 0.1418 - val_accuracy: 0.9703\n",
      "Epoch 24/50\n",
      "20/20 [==============================] - 3s 157ms/step - loss: 0.0426 - accuracy: 0.9910 - val_loss: 0.1471 - val_accuracy: 0.9442\n",
      "Epoch 25/50\n",
      "20/20 [==============================] - 3s 154ms/step - loss: 0.0571 - accuracy: 0.9820 - val_loss: 0.2121 - val_accuracy: 0.9219\n",
      "Epoch 26/50\n",
      "20/20 [==============================] - 3s 156ms/step - loss: 0.0410 - accuracy: 0.9910 - val_loss: 0.1552 - val_accuracy: 0.9517\n",
      "Epoch 27/50\n",
      "20/20 [==============================] - 3s 154ms/step - loss: 0.0090 - accuracy: 1.0000 - val_loss: 0.1173 - val_accuracy: 0.9703\n",
      "Epoch 28/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 0.0125 - accuracy: 0.9976 - val_loss: 0.2361 - val_accuracy: 0.9294\n",
      "Epoch 29/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 0.0120 - accuracy: 0.9992 - val_loss: 0.1081 - val_accuracy: 0.9703\n",
      "Epoch 30/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 0.0054 - accuracy: 1.0000 - val_loss: 0.0769 - val_accuracy: 0.9703\n",
      "Epoch 31/50\n",
      "20/20 [==============================] - 3s 156ms/step - loss: 0.0082 - accuracy: 0.9992 - val_loss: 0.0713 - val_accuracy: 0.9814\n",
      "Epoch 32/50\n",
      "20/20 [==============================] - 3s 154ms/step - loss: 0.0068 - accuracy: 1.0000 - val_loss: 0.0743 - val_accuracy: 0.9814\n",
      "Epoch 33/50\n",
      "20/20 [==============================] - 3s 156ms/step - loss: 0.0044 - accuracy: 1.0000 - val_loss: 0.0806 - val_accuracy: 0.9814\n",
      "Epoch 34/50\n",
      "20/20 [==============================] - 3s 154ms/step - loss: 0.0043 - accuracy: 1.0000 - val_loss: 0.0805 - val_accuracy: 0.9740\n",
      "Epoch 35/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 0.0049 - accuracy: 1.0000 - val_loss: 0.0783 - val_accuracy: 0.9740\n",
      "Epoch 36/50\n",
      "20/20 [==============================] - 3s 154ms/step - loss: 0.0047 - accuracy: 1.0000 - val_loss: 0.0763 - val_accuracy: 0.9740\n",
      "Epoch 37/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 0.0039 - accuracy: 1.0000 - val_loss: 0.0813 - val_accuracy: 0.9814\n",
      "Epoch 38/50\n",
      "20/20 [==============================] - 3s 154ms/step - loss: 0.0020 - accuracy: 1.0000 - val_loss: 0.0852 - val_accuracy: 0.9777\n",
      "Epoch 39/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 0.0014 - accuracy: 1.0000 - val_loss: 0.0766 - val_accuracy: 0.9740\n",
      "Epoch 40/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 0.0011 - accuracy: 1.0000 - val_loss: 0.0798 - val_accuracy: 0.9777\n",
      "Epoch 41/50\n",
      "20/20 [==============================] - 3s 156ms/step - loss: 0.0010 - accuracy: 1.0000 - val_loss: 0.0778 - val_accuracy: 0.9777\n",
      "Epoch 42/50\n",
      "20/20 [==============================] - 3s 156ms/step - loss: 9.8214e-04 - accuracy: 1.0000 - val_loss: 0.0788 - val_accuracy: 0.9777\n",
      "Epoch 43/50\n",
      "20/20 [==============================] - 3s 153ms/step - loss: 9.2279e-04 - accuracy: 1.0000 - val_loss: 0.0786 - val_accuracy: 0.9777\n",
      "Epoch 44/50\n",
      "20/20 [==============================] - 3s 156ms/step - loss: 8.7765e-04 - accuracy: 1.0000 - val_loss: 0.0787 - val_accuracy: 0.9777\n",
      "Epoch 45/50\n",
      "20/20 [==============================] - 3s 154ms/step - loss: 8.3493e-04 - accuracy: 1.0000 - val_loss: 0.0787 - val_accuracy: 0.9777\n",
      "Epoch 46/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 7.9702e-04 - accuracy: 1.0000 - val_loss: 0.0788 - val_accuracy: 0.9777\n",
      "Epoch 47/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 7.6211e-04 - accuracy: 1.0000 - val_loss: 0.0788 - val_accuracy: 0.9777\n",
      "Epoch 48/50\n",
      "20/20 [==============================] - 3s 156ms/step - loss: 7.2996e-04 - accuracy: 1.0000 - val_loss: 0.0789 - val_accuracy: 0.9777\n",
      "Epoch 49/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 7.0016e-04 - accuracy: 1.0000 - val_loss: 0.0789 - val_accuracy: 0.9777\n",
      "Epoch 50/50\n",
      "20/20 [==============================] - 3s 155ms/step - loss: 6.7244e-04 - accuracy: 1.0000 - val_loss: 0.0790 - val_accuracy: 0.9777\n",
      "\n",
      "Test set Evaluation:\n",
      "\n",
      "4/4 [==============================] - 2s 26ms/step - loss: 0.0636 - accuracy: 0.9844\n",
      "\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.compile(\n",
    "    optimizer=tf.keras.optimizers.Adam(epsilon=0.01),\n",
    "    loss='categorical_crossentropy',\n",
    "    metrics=['accuracy']\n",
    ")\n",
    "\n",
    "history = model.fit(\n",
    "    ds_train,\n",
    "    validation_data=ds_valid,\n",
    "    epochs=50\n",
    ")\n",
    "\n",
    "# Evaluate the model on the test set\n",
    "print('\\nTest set Evaluation:\\n')\n",
    "test_loss, test_accuracy = model.evaluate(ds_test)\n",
    "print('\\n')\n",
    "\n",
    "# Plot loss and accuracy curves\n",
    "history_frame = pd.DataFrame(history.history)\n",
    "history_frame.loc[:, ['loss', 'val_loss']].plot()\n",
    "plt.title('Loss Curves')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Loss')\n",
    "history_frame.loc[:, ['accuracy', 'val_accuracy']].plot()\n",
    "plt.title('Accuracy Curves')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Save the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-07T10:37:17.969355Z",
     "iopub.status.busy": "2023-05-07T10:37:17.969035Z",
     "iopub.status.idle": "2023-05-07T10:37:18.559888Z",
     "shell.execute_reply": "2023-05-07T10:37:18.558906Z",
     "shell.execute_reply.started": "2023-05-07T10:37:17.969325Z"
    }
   },
   "outputs": [],
   "source": [
    "model.save('model.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Make a Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-07T10:37:18.561521Z",
     "iopub.status.busy": "2023-05-07T10:37:18.561176Z",
     "iopub.status.idle": "2023-05-07T10:37:19.041941Z",
     "shell.execute_reply": "2023-05-07T10:37:19.040966Z",
     "shell.execute_reply.started": "2023-05-07T10:37:18.561491Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 200x200 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 146ms/step\n",
      "Prediction:  [[0. 1. 0. 0. 0.]]\n",
      "Predicted class:  2_rupee\n"
     ]
    }
   ],
   "source": [
    "from PIL import Image\n",
    "\n",
    "# Select a file from the dataset to predict\n",
    "image_file_path = 'images/processed/2_rupee/IMG_20230501_200352912.jpg' \n",
    "\n",
    "# Get the image from the file.\n",
    "image = Image.open(image_file_path)\n",
    "image = image.convert('L') # Convert image to grayscale\n",
    "\n",
    "# Plot the image\n",
    "plt.figure(figsize=(2,2))\n",
    "plt.imshow(image, 'gray')\n",
    "plt.title('Image')\n",
    "plt.show()\n",
    "\n",
    "image_array = np.asarray(image) # convert PIL image to np.ndarray\n",
    "\n",
    "# image_array has a shape of (256,256) but the prediction method of NN model requires an input\n",
    "# shape of (w,x,y,z) where:\n",
    "# w is batch size\n",
    "# x and y are image dimentions\n",
    "# z is the channel depth\n",
    "# So we have to add an extra dimention at the begining and end of image_array\n",
    "image_array = image_array[np.newaxis, :, :, np.newaxis]\n",
    "\n",
    "preditcion = model.predict(image_array)\n",
    "print('Prediction: ', preditcion)\n",
    "\n",
    "predicted_class = np.argmax(preditcion)\n",
    "print('Predicted class: ', class_names[predicted_class])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Done**"
   ]
  }
 ],
 "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": "3.11.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}