Skip to content
Snippets Groups Projects
03-Tracking-and-visualizing.ipynb 91.7 KiB
Newer Older
{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "German Traffic Sign Recognition Benchmark (GTSRB)\n",
    "=================================================\n",
    "---\n",
    "Introduction au Deep Learning  (IDLE) - S. Aria, E. Maldonado, JL. Parouty - CNRS/SARI/DEVLOG - 2020\n",
    "\n",
    "## Episode 3 : Tracking and visualizing\n",
    "\n",
    "Our main steps:\n",
    " - Monitoring and understanding our model training\n",
    " - Analyze the results \n",
    " - Improving our model\n",
    " - Add recovery points\n",
    "\n",
    "\n",
    "## 1/ Import and init"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "IDLE 2020 - Practical Work Module\n",
      "  Version            : 0.1.1\n",
      "  Run time           : Monday 6 January 2020, 14:49:33\n",
      "  Matplotlib style   : idle/talk.mplstyle\n",
      "  TensorFlow version : 2.0.0\n",
      "  Keras version      : 2.2.4-tf\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras.callbacks import TensorBoard\n",
    "\n",
    "import numpy as np\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "import time\n",
    "\n",
    "import idle.pwk as ooo\n",
    "\n",
    "ooo.init()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2/ Reload dataset\n",
    "Dataset is one of the saved dataset: RGB25, RGB35, L25, L35, etc.  \n",
    "First of all, we're going to use the dataset : **L25**  \n",
    "(with a GPU, it only takes 35'' compared to more than 5' with a CPU !)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset loaded, size=247.6 Mo\n",
      "\n",
      "CPU times: user 0 ns, sys: 120 ms, total: 120 ms\n",
      "Wall time: 120 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "dataset ='L25'\n",
    "img_lx  = 25\n",
    "img_ly  = 25\n",
    "img_lz  = 1\n",
    "\n",
    "# ---- Read dataset\n",
    "x_train = np.load('./data/{}/x_train.npy'.format(dataset))\n",
    "y_train = np.load('./data/{}/y_train.npy'.format(dataset))\n",
    "\n",
    "x_test  = np.load('./data/{}/x_test.npy'.format(dataset))\n",
    "y_test  = np.load('./data/{}/y_test.npy'.format(dataset))\n",
    "\n",
    "# ---- Reshape data\n",
    "x_train = x_train.reshape( x_train.shape[0], img_lx, img_ly, img_lz)\n",
    "x_test  = x_test.reshape(  x_test.shape[0],  img_lx, img_ly, img_lz)\n",
    "\n",
    "input_shape = (img_lx, img_ly, img_lz)\n",
    "\n",
    "print(\"Dataset loaded, size={:.1f} Mo\\n\".format(ooo.get_directory_size('./data/'+dataset)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3/ Have a look to the dataset\n",
    "Note: Data must be reshape for matplotlib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x_train :  (39209, 25, 25, 1)\n",
      "y_train :  (39209,)\n",
      "x_test  :  (12630, 25, 25, 1)\n",
      "y_test  :  (12630,)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x169.2 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x291.6 with 36 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(\"x_train : \", x_train.shape)\n",
    "print(\"y_train : \", y_train.shape)\n",
    "print(\"x_test  : \", x_test.shape)\n",
    "print(\"y_test  : \", y_test.shape)\n",
    "\n",
    "if img_lz>1:\n",
    "    ooo.plot_images(x_train.reshape(-1,img_lx,img_ly,img_lz), y_train, range(6),  columns=3,  x_size=4, y_size=3)\n",
    "    ooo.plot_images(x_train.reshape(-1,img_lx,img_ly,img_lz), y_train, range(36), columns=12, x_size=1, y_size=1)\n",
    "else:\n",
    "    ooo.plot_images(x_train.reshape(-1,img_lx,img_ly), y_train, range(6),  columns=6,  x_size=2, y_size=2)\n",
    "    ooo.plot_images(x_train.reshape(-1,img_lx,img_ly), y_train, range(36), columns=12, x_size=1, y_size=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4/ Create model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "batch_size  =  64\n",
    "num_classes =  43\n",
    "epochs      =  5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d (Conv2D)              (None, 23, 23, 96)        960       \n",
      "_________________________________________________________________\n",
      "max_pooling2d (MaxPooling2D) (None, 11, 11, 96)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 9, 9, 192)         166080    \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 4, 4, 192)         0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 3072)              0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 3072)              9440256   \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 500)               1536500   \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 500)               250500    \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 43)                21543     \n",
      "=================================================================\n",
      "Total params: 11,415,839\n",
      "Trainable params: 11,415,839\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = keras.models.Sequential()\n",
    "model.add( keras.layers.Conv2D(96, (3,3), activation='relu', input_shape=(img_lx, img_ly, img_lz)))\n",
    "model.add( keras.layers.MaxPooling2D((2, 2)))\n",
    "model.add( keras.layers.Conv2D(192, (3, 3), activation='relu'))\n",
    "model.add( keras.layers.MaxPooling2D((2, 2)))\n",
    "model.add( keras.layers.Flatten()) \n",
    "model.add( keras.layers.Dense(3072, activation='relu'))\n",
    "model.add( keras.layers.Dense(500, activation='relu'))\n",
    "model.add( keras.layers.Dense(500, activation='relu'))\n",
    "model.add( keras.layers.Dense(43, activation='softmax'))\n",
    "model.summary()\n",
    "\n",
    "model.compile(optimizer='adam',\n",
    "              loss='sparse_categorical_crossentropy',\n",
    "              metrics=['accuracy'])\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5/ Run model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 39209 samples, validate on 12630 samples\n",
      "Epoch 1/5\n",
      "39209/39209 [==============================] - 14s 354us/sample - loss: 0.9577 - accuracy: 0.7231 - val_loss: 0.4159 - val_accuracy: 0.8985\n",
      "Epoch 2/5\n",
      "39209/39209 [==============================] - 5s 128us/sample - loss: 0.1073 - accuracy: 0.9688 - val_loss: 0.2749 - val_accuracy: 0.9319\n",
      "Epoch 3/5\n",
      "39209/39209 [==============================] - 5s 132us/sample - loss: 0.0546 - accuracy: 0.9834 - val_loss: 0.2435 - val_accuracy: 0.9407\n",
      "Epoch 4/5\n",
      "39209/39209 [==============================] - 5s 131us/sample - loss: 0.0380 - accuracy: 0.9888 - val_loss: 0.3128 - val_accuracy: 0.9365\n",
      "Epoch 5/5\n",
      "39209/39209 [==============================] - 5s 132us/sample - loss: 0.0335 - accuracy: 0.9898 - val_loss: 0.3526 - val_accuracy: 0.9315\n",
      "CPU times: user 40.1 s, sys: 5.03 s, total: 45.1 s\n",
      "Wall time: 34.5 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "history = model.fit(  x_train, y_train,\n",
    "                      batch_size=batch_size,\n",
    "                      epochs=epochs,\n",
    "                      verbose=1,\n",
    "                      validation_data=(x_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6/ Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss      : 0.3526\n",
      "Test accuracy  : 0.9315\n"
     ]
    }
   ],
   "source": [
    "score = model.evaluate(x_test, y_test, verbose=0)\n",
    "\n",
    "print('Test loss      : {:5.4f}'.format(score[0]))\n",
    "print('Test accuracy  : {:5.4f}'.format(score[1]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "### Results :  \n",
    "L25 : size=250 Mo, 93.15%  \n",
    "..."
   ]
  },
  {
   "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": "3.7.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}