Fruit Classification with Transfer Learning.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import pathlib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_url = ''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "zip_dir = tf.keras.utils.get_file('fruits360.zip', origin=data_url, extract=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = pathlib.Path(zip_dir).parent / 'fruits360_filtered'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_dir = path / 'Training'\n",
    "validation_dir = path / 'Test'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "total_train = 11398\n",
    "total_val = 4752"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_img_gen = ImageDataGenerator(\n",
    "    rescale=1./255, rotation_range=40, width_shift_range=0.1,\n",
    "    height_shift_range=0.1, shear_range=0.2, zoom_range=0.2,\n",
    "    horizontal_flip=True, fill_mode='nearest'\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "validation_img_gen = ImageDataGenerator(rescale=1./255) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "batch_size = 16\n",
    "img_height = 100\n",
    "img_width = 100\n",
    "channel = 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 11398 images belonging to 120 classes.\n"
     ]
    }
   ],
   "source": [
    "train_data_gen = train_img_gen.flow_from_directory(\n",
    "    batch_size=batch_size, directory=train_dir, target_size=(\n",
    "        img_height, img_width\n",
    "    )\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 4752 images belonging to 120 classes.\n"
     ]
    }
   ],
   "source": [
    "val_data_gen = validation_img_gen.flow_from_directory(\n",
    "    batch_size=batch_size, directory=validation_dir, target_size=(\n",
    "        img_height, img_width\n",
    "    )\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.applications import VGG16\n",
    "from tensorflow.keras import layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "base_model = VGG16(\n",
    "    input_shape=(\n",
    "        img_height, img_width, channel\n",
    "    ), weights='imagenet', include_top=False\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"vgg16\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         [(None, 100, 100, 3)]     0         \n",
      "_________________________________________________________________\n",
      "block1_conv1 (Conv2D)        (None, 100, 100, 64)      1792      \n",
      "_________________________________________________________________\n",
      "block1_conv2 (Conv2D)        (None, 100, 100, 64)      36928     \n",
      "_________________________________________________________________\n",
      "block1_pool (MaxPooling2D)   (None, 50, 50, 64)        0         \n",
      "_________________________________________________________________\n",
      "block2_conv1 (Conv2D)        (None, 50, 50, 128)       73856     \n",
      "_________________________________________________________________\n",
      "block2_conv2 (Conv2D)        (None, 50, 50, 128)       147584    \n",
      "_________________________________________________________________\n",
      "block2_pool (MaxPooling2D)   (None, 25, 25, 128)       0         \n",
      "_________________________________________________________________\n",
      "block3_conv1 (Conv2D)        (None, 25, 25, 256)       295168    \n",
      "_________________________________________________________________\n",
      "block3_conv2 (Conv2D)        (None, 25, 25, 256)       590080    \n",
      "_________________________________________________________________\n",
      "block3_conv3 (Conv2D)        (None, 25, 25, 256)       590080    \n",
      "_________________________________________________________________\n",
      "block3_pool (MaxPooling2D)   (None, 12, 12, 256)       0         \n",
      "_________________________________________________________________\n",
      "block4_conv1 (Conv2D)        (None, 12, 12, 512)       1180160   \n",
      "_________________________________________________________________\n",
      "block4_conv2 (Conv2D)        (None, 12, 12, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "block4_conv3 (Conv2D)        (None, 12, 12, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "block4_pool (MaxPooling2D)   (None, 6, 6, 512)         0         \n",
      "_________________________________________________________________\n",
      "block5_conv1 (Conv2D)        (None, 6, 6, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv2 (Conv2D)        (None, 6, 6, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv3 (Conv2D)        (None, 6, 6, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_pool (MaxPooling2D)   (None, 3, 3, 512)         0         \n",
      "=================================================================\n",
      "Total params: 14,714,688\n",
      "Trainable params: 0\n",
      "Non-trainable params: 14,714,688\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "base_model.trainable = False\n",
    "base_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = tf.keras.Sequential([\n",
    "    base_model,\n",
    "    layers.Flatten(),\n",
    "    layers.Dense(1000, activation='relu'),\n",
    "    layers.Dense(120, activation='softmax')\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "optimizer = tf.keras.optimizers.Adam(0.001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(\n",
    "    loss='categorical_crossentropy',\n",
    "    optimizer=optimizer,\n",
    "    metrics=['accuracy']\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "vgg16 (Functional)           (None, 3, 3, 512)         14714688  \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 4608)              0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 1000)              4609000   \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 120)               120120    \n",
      "=================================================================\n",
      "Total params: 19,443,808\n",
      "Trainable params: 4,729,120\n",
      "Non-trainable params: 14,714,688\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\hp\\anaconda3\\lib\\site-packages\\tensorflow\\python\\keras\\engine\\training.py:1940: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
      "  warnings.warn('`Model.fit_generator` is deprecated and '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "712/712 [==============================] - 42s 59ms/step - loss: 0.2507 - accuracy: 0.9206 - val_loss: 0.3567 - val_accuracy: 0.8904\n",
      "Epoch 2/5\n",
      "712/712 [==============================] - 40s 56ms/step - loss: 0.2532 - accuracy: 0.9200 - val_loss: 0.4797 - val_accuracy: 0.8590\n",
      "Epoch 3/5\n",
      "712/712 [==============================] - 37s 52ms/step - loss: 0.2323 - accuracy: 0.9236 - val_loss: 0.3910 - val_accuracy: 0.8864\n",
      "Epoch 4/5\n",
      "712/712 [==============================] - 38s 54ms/step - loss: 0.2087 - accuracy: 0.9307 - val_loss: 0.5171 - val_accuracy: 0.8647\n",
      "Epoch 5/5\n",
      "712/712 [==============================] - 38s 54ms/step - loss: 0.1996 - accuracy: 0.9338 - val_loss: 0.4826 - val_accuracy: 0.8628\n"
     ]
    }
   ],
   "source": [
    "history = model.fit_generator(\n",
    "    train_data_gen,\n",
    "    steps_per_epoch=total_train // batch_size,\n",
    "    epochs=5,\n",
    "    validation_data=val_data_gen,\n",
    "    validation_steps=total_val // batch_size\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create a dictionary of the model history\n",
    "history_dict = hisotry.history\n",
    "loss_values = hisotry_dict['loss']\n",
    "val_loss_values = hisotry_dict['val_loss']\n",
    "acc_values = hisotry_dict['accuracy']\n",
    "val_acc_values = hisotry_dict['val_accuracy']\n",
    "epochs = range(1, len(hisotry_dict['accuracy']) + 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "# plot the training/validation loss\n",
    "plt.plot(epochs, loss_values, 'bo', label='Training loss')\n",
    "plt.plot(epochs, val_loss_values, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "# plot the training/validation accuracy\n",
    "plt.plot(epochs, acc_values, 'bo', label='Training accuracy')\n",
    "plt.plot(epochs, val_acc_values, 'b', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "297/297 [==============================] - 7s 22ms/step - loss: 0.4826 - accuracy: 0.8628\n",
      "Model testing accuracy/testing loss:  0.8627946376800537   0.48264288902282715\n"
     ]
    }
   ],
   "source": [
    "test_loss, test_acc = model.evaluate(val_data_gen)\n",
    "print('Model testing accuracy/testing loss: ', test_acc, \" \", test_loss)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}