model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "python version:  3.9.13 (main, Oct 13 2022, 16:12:30) \n",
      "[Clang 12.0.0 ]\n"
     ]
    }
   ],
   "source": [
    "# check versions and devices\n",
    "import tensorflow as tf\n",
    "import os\n",
    "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '1'\n",
    "print(\"python version: \", sys.version)\n",
    "# print(\"tf version: \", tf.__version__)\n",
    "# print(\"devices: \", tf.config.list_physical_devices())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 16041 files belonging to 2 classes.\n",
      "Using 12833 files for training.\n",
      "Found 16041 files belonging to 2 classes.\n",
      "Using 3208 files for validation.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['car', 'not_car']"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# create train and validation datasets\n",
    "batch_size = 128\n",
    "\n",
    "train_ds = tf.keras.preprocessing.image_dataset_from_directory(\n",
    "    \"./data/images/\",\n",
    "    validation_split=0.2,\n",
    "    subset=\"training\",\n",
    "    seed=123,\n",
    "    batch_size=batch_size,\n",
    "    shuffle=True\n",
    ")\n",
    "val_ds = tf.keras.preprocessing.image_dataset_from_directory(\n",
    "    \"./data/images/\",\n",
    "    validation_split=0.2,\n",
    "    subset=\"validation\",\n",
    "    seed=123,\n",
    "    batch_size=batch_size,\n",
    "    shuffle=True\n",
    ")\n",
    "\n",
    "train_ds.class_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# configure performance\n",
    "autotune = tf.data.AUTOTUNE\n",
    "\n",
    "def rsize(image, label):\n",
    "    image = tf.image.resize(image, (224, 224))\n",
    "    return image, label\n",
    "\n",
    "train_ds = train_ds.map(rsize, num_parallel_calls=autotune)\n",
    "val_ds = val_ds.map(rsize, num_parallel_calls=autotune)\n",
    "\n",
    "train_ds = train_ds.cache().prefetch(buffer_size=autotune)\n",
    "val_ds = val_ds.cache().prefetch(buffer_size=autotune)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " rescaling_2 (Rescaling)     (None, 224, 224, 3)       0         \n",
      "                                                                 \n",
      " conv2d_6 (Conv2D)           (None, 222, 222, 32)      896       \n",
      "                                                                 \n",
      " max_pooling2d_6 (MaxPooling  (None, 111, 111, 32)     0         \n",
      " 2D)                                                             \n",
      "                                                                 \n",
      " conv2d_7 (Conv2D)           (None, 109, 109, 32)      9248      \n",
      "                                                                 \n",
      " max_pooling2d_7 (MaxPooling  (None, 54, 54, 32)       0         \n",
      " 2D)                                                             \n",
      "                                                                 \n",
      " conv2d_8 (Conv2D)           (None, 52, 52, 32)        9248      \n",
      "                                                                 \n",
      " max_pooling2d_8 (MaxPooling  (None, 26, 26, 32)       0         \n",
      " 2D)                                                             \n",
      "                                                                 \n",
      " flatten_2 (Flatten)         (None, 21632)             0         \n",
      "                                                                 \n",
      " dense_4 (Dense)             (None, 128)               2769024   \n",
      "                                                                 \n",
      " dense_5 (Dense)             (None, 2)                 258       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 2,788,674\n",
      "Trainable params: 2,788,674\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# model -> fine tuning pre-trained model imported from CNN-1.6 by Mario Peng Lee\n",
    "model = tf.keras.Sequential([\n",
    "  # input layer: dynamic input size\n",
    "  tf.keras.layers.InputLayer(input_shape=(224, 224, 3)),\n",
    "  # rescaling to 0-1\n",
    "  tf.keras.layers.Rescaling(1./255),\n",
    "  # hidden layers: conv2D + maxpooling (note: batchnorm not used as destroys validation accuracy)\n",
    "  tf.keras.layers.Conv2D(32, 3, activation='relu'),\n",
    "  tf.keras.layers.MaxPooling2D(),\n",
    "  tf.keras.layers.Conv2D(32, 3, activation='relu'),\n",
    "  tf.keras.layers.MaxPooling2D(),\n",
    "  tf.keras.layers.Conv2D(32, 3, activation='relu'),\n",
    "  tf.keras.layers.MaxPooling2D(),\n",
    "  # flatten and dense layers\n",
    "  tf.keras.layers.Flatten(),\n",
    "  tf.keras.layers.Dense(128, activation='relu'),\n",
    "  # output layer\n",
    "  tf.keras.layers.Dense(2)\n",
    "])\n",
    "model.summary()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# compile model\n",
    "model.compile(\n",
    "    optimizer='adam',\n",
    "    loss=tf.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
    "    metrics=['accuracy']\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/3\n",
      "101/101 [==============================] - 76s 726ms/step - loss: 0.4184 - accuracy: 0.8108 - val_loss: 0.2274 - val_accuracy: 0.9165\n",
      "Epoch 2/3\n",
      "101/101 [==============================] - 70s 692ms/step - loss: 0.2237 - accuracy: 0.9155 - val_loss: 0.1791 - val_accuracy: 0.9373\n",
      "Epoch 3/3\n",
      "101/101 [==============================] - 70s 693ms/step - loss: 0.1835 - accuracy: 0.9306 - val_loss: 0.1970 - val_accuracy: 0.9230\n"
     ]
    }
   ],
   "source": [
    "# train\n",
    "import tensorflow as tf\n",
    "history = model.fit(\n",
    "    train_ds,\n",
    "    validation_data=val_ds,\n",
    "    epochs=3\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save model\n",
    "import pandas as pd\n",
    "\n",
    "model.save('model.h5')\n",
    "hist_df = pd.DataFrame(history.history)\n",
    "with open(\"hist.json\", \"w\") as f:\n",
    "    hist_df.to_json(f)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load\n",
    "from tensorflow.keras.models import load_model\n",
    "import pandas as pd\n",
    "from tensorflow_hub import KerasLayer\n",
    "import pandas as pd\n",
    "\n",
    "model = load_model('model.h5')\n",
    "model_hist = pd.read_json(\"hist.json\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot accuracy and val_accuracy\n",
    "import matplotlib.pyplot as plt\n",
    "plt.plot(model_hist['accuracy'], label='train')\n",
    "plt.plot(model_hist['val_accuracy'], label = 'test')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend(loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot loss and val_loss\n",
    "plt.plot(model_hist['loss'], label='train')\n",
    "plt.plot(model_hist['val_loss'], label = 'test')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend(loc='upper left')\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.9.13 ('ai')",
   "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": "3.9.13"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "10b546dc9a6492740535e7e359781ac3cb076df88249adf04015d7e9d7b6abb1"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}