train-mobilenet.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Udacity Lyft Challenge\n",
    "## Training script"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import time\n",
    "import glob\n",
    "import random\n",
    "import re\n",
    "import cv2\n",
    "import math\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from IPython.display import HTML"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Force CPU: uncomment to force CPU - usefull for editing model architecture while another one is training\n",
    "# Must be done before tensorflow is imported\n",
    "# os.environ[\"CUDA_DEVICE_ORDER\"] = \"PCI_BUS_ID\"   # see issue #152\n",
    "# os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/anand/anaconda3/lib/python3.6/site-packages/h5py/__init__.py:34: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "from keras.models import load_model\n",
    "from keras.models import model_from_yaml\n",
    "from keras.applications.mobilenet import MobileNet\n",
    "from keras.applications import mobilenet\n",
    "from keras.preprocessing import image\n",
    "from keras.models import Model\n",
    "from keras.layers import Activation, Dropout, Add, Flatten, Dense, Reshape, Conv2D, Conv2DTranspose, GlobalAveragePooling2D, GlobalMaxPooling2D\n",
    "from keras.optimizers import Adam, RMSprop, SGD\n",
    "from keras import backend as K\n",
    "from keras import regularizers\n",
    "from keras import losses\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.initializers import TruncatedNormal\n",
    "from keras.callbacks import LambdaCallback, ModelCheckpoint, ReduceLROnPlateau, LearningRateScheduler, TensorBoard, EarlyStopping\n",
    "from keras.utils import Sequence\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[name: \"/device:CPU:0\"\n",
      "device_type: \"CPU\"\n",
      "memory_limit: 268435456\n",
      "locality {\n",
      "}\n",
      "incarnation: 14980066422001819313\n",
      ", name: \"/device:GPU:0\"\n",
      "device_type: \"GPU\"\n",
      "memory_limit: 7563123098\n",
      "locality {\n",
      "  bus_id: 1\n",
      "  links {\n",
      "  }\n",
      "}\n",
      "incarnation: 3476406077072319794\n",
      "physical_device_desc: \"device: 0, name: GeForce GTX 1080, pci bus id: 0000:01:00.0, compute capability: 6.1\"\n",
      "]\n"
     ]
    }
   ],
   "source": [
    "# To ensure GPU is being used\n",
    "from tensorflow.python.client import device_lib\n",
    "print(device_lib.list_local_devices())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Basic Definitions\n",
    "\n",
    "- defines carla's hood shape for filling in labels\n",
    "- define base_model_file if training from existing model; set to None to start training from scratch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "IMAGE_HT = 576\n",
    "IMAGE_WD = 800\n",
    "data_dir = '/home/anand/lyft/CARLA_0.8.2/PythonClient/'\n",
    "CarlaLabels = {\n",
    "    \"None\":0,\n",
    "    \"Buildings\":1,\n",
    "    \"Fences\":2,\n",
    "    \"Other\":3,\n",
    "    \"Pedestrians\":4,\n",
    "    \"Poles\":5,\n",
    "    \"RoadLines\":6,\n",
    "    \"Roads\":7,\n",
    "    \"Sidewalks\":8,\n",
    "    \"Vegetation\":9,\n",
    "    \"Vehicles\":10,\n",
    "    \"Walls\":11,\n",
    "    \"TrafficSigns\":12,\n",
    "}\n",
    "CarlaHood = np.array([\n",
    "        (0,519),\n",
    "        (110,521),\n",
    "        (135, 508),\n",
    "        (383, 498),\n",
    "        (417,498),\n",
    "        (665, 508),\n",
    "        (687, 521),\n",
    "        (IMAGE_WD,519),\n",
    "        (IMAGE_WD, IMAGE_HT), \n",
    "        (0, IMAGE_HT)])\n",
    "h = np.ones((IMAGE_HT, IMAGE_WD), dtype=np.uint8)\n",
    "cv2.fillPoly(h, [CarlaHood], 0)\n",
    "CarlaHood = h==0\n",
    "\n",
    "num_classes = 2\n",
    "model_name = 'mn12_%s' % time.strftime('%m%d')\n",
    "base_model_file = os.path.join(data_dir, '0_mn6_0601_weights_2.hdf5')\n",
    "trainingMode = True"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Define model\n",
    "\n",
    "- Load existing model or define new with pre-trained imagenet weights\n",
    "- Can specify layer name to freeze till; set FROZEN_LAYERS=0 to train all layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Trainable after conv_pw_13_relu\n"
     ]
    }
   ],
   "source": [
    "from keras.utils.data_utils import get_file\n",
    "\n",
    "# Load model if available on disk\n",
    "if base_model_file != None:\n",
    "    if not os.path.exists(os.path.join(base_model_file)):\n",
    "        raise Exception('base model not found')\n",
    "    model = load_model(base_model_file, compile=False, custom_objects={\n",
    "                   'relu6': mobilenet.relu6,\n",
    "                   'DepthwiseConv2D': mobilenet.DepthwiseConv2D})\n",
    "else:\n",
    "    trainingMode = True\n",
    "    base_model = MobileNet(include_top=False, weights=None, input_shape=(IMAGE_HT, IMAGE_WD, 3))\n",
    "    \n",
    "    BASE_WEIGHT_PATH = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.6/'\n",
    "    root_model_name = 'mobilenet_1_0_224_tf_no_top.h5'\n",
    "    weights_path = BASE_WEIGHT_PATH + root_model_name\n",
    "    weights_path = get_file(root_model_name,\n",
    "                            weights_path,\n",
    "                            cache_subdir='models')\n",
    "    base_model.load_weights(weights_path)\n",
    "    base_model.summary()\n",
    "    \n",
    "    regularizer = regularizers.l2(0.01)\n",
    "    x = base_model.output\n",
    "    x = Conv2D(filters=num_classes, kernel_size=1, strides=1, padding='SAME',\n",
    "               kernel_regularizer=regularizer, kernel_initializer='random_uniform', bias_initializer='zeros')(x)\n",
    "\n",
    "    #conv_pw_5_relu (Activation)  (None, 72, 100, 256)\n",
    "    x = Conv2DTranspose(filters=256, kernel_size=8, strides=4, padding='SAME',\n",
    "               kernel_regularizer=regularizer, kernel_initializer='random_uniform', bias_initializer='zeros')(x)\n",
    "    y = Dropout(0.5)(base_model.get_layer('conv_pw_5_relu').output)\n",
    "    x = Add()([x, y])\n",
    "    \n",
    "    #conv_pw_3_relu (Activation)  (None, 144, 200, 128)\n",
    "    x = Conv2DTranspose(filters=128, kernel_size=4, strides=2, padding='SAME',\n",
    "               kernel_regularizer=regularizer, kernel_initializer='random_uniform', bias_initializer='zeros')(x)\n",
    "    y = Dropout(0.5)(base_model.get_layer('conv_pw_3_relu').output)\n",
    "    x = Add()([x, y])\n",
    "\n",
    "    #conv_pw_1_relu (Activation)  (None, 288, 400, 64) \n",
    "    x = Conv2DTranspose(filters=64, kernel_size=4, strides=2, padding='SAME',\n",
    "               kernel_regularizer=regularizer, kernel_initializer='random_uniform', bias_initializer='zeros')(x)\n",
    "    y = Dropout(0.5)(base_model.get_layer('conv_pw_1_relu').output)\n",
    "    x = Add()([x, y])\n",
    "\n",
    "    logits = Conv2DTranspose(filters=num_classes, kernel_size=4, strides=2, padding='SAME', activation='sigmoid',\n",
    "               kernel_regularizer=regularizer, kernel_initializer='random_uniform', bias_initializer='zeros')(x)\n",
    "\n",
    "    #logits = Reshape((-1, num_classes))(x)\n",
    "    model = Model(inputs=base_model.input, outputs=logits)\n",
    "\n",
    "# Freeze layers\n",
    "FROZEN_LAYERS = [idx for idx,x in enumerate(model.layers) if x.name=='conv_pw_13_relu'][0]\n",
    "print('Trainable after %s' % model.layers[FROZEN_LAYERS].name)\n",
    "for idx,layer in enumerate(model.layers):\n",
    "    layer.trainable = (idx > FROZEN_LAYERS)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dice Coefficient and Loss function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def dice_layer(y_true_layer, y_pred_layer):\n",
    "    intersection = K.sum(y_true_layer * y_pred_layer)\n",
    "    return (2. * intersection + 1.) / (K.sum(y_true_layer) + K.sum(y_pred_layer) + 1.)\n",
    "    \n",
    "def dice_coef0(y_true, y_pred):\n",
    "    return dice_layer(y_true[:,:,:,0], y_pred[:,:,:,0])\n",
    "\n",
    "def dice_coef1(y_true, y_pred):\n",
    "    return dice_layer(y_true[:,:,:,1], y_pred[:,:,:,1])\n",
    "\n",
    "def dice_coef_loss(y_true, y_pred):\n",
    "    return 3.0 - (dice_coef0(y_true, y_pred) + (2.0 * dice_coef1(y_true, y_pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## IOU metric and Loss function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def iou_layer(y_true_layer, y_pred_layer):\n",
    "    intersection = K.sum(y_true_layer * y_pred_layer)\n",
    "    return (intersection + 1.) / (K.sum(y_true_layer) + K.sum(y_pred_layer) - intersection + 1.)\n",
    "\n",
    "def iou0(y_true, y_pred):\n",
    "    return iou_layer(y_true[:,:,:,0], y_pred[:,:,:,0])\n",
    "\n",
    "def iou1(y_true, y_pred):\n",
    "    return iou_layer(y_true[:,:,:,1], y_pred[:,:,:,1])\n",
    "\n",
    "def iou_loss(y_true, y_pred):\n",
    "    return 3.0 - (iou0(y_true, y_pred) + (2.0 * iou1(y_true, y_pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Metrics and Loss function\n",
    "\n",
    "- False positive, negative and F1 score metrics and loss functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Metrics\n",
    "def fp(y_true_layer, y_pred_layer):\n",
    "    intersection = K.sum(y_true_layer * y_pred_layer)\n",
    "    return K.sum(y_pred_layer) - intersection\n",
    "\n",
    "def fn(y_true_layer, y_pred_layer):\n",
    "    intersection = K.sum(y_true_layer * y_pred_layer)\n",
    "    return K.sum(y_true_layer) - intersection\n",
    "\n",
    "def fp0(y_true, y_pred):\n",
    "    return fp(y_true[:,:,:,0], y_pred[:,:,:,0])\n",
    "\n",
    "def fp1(y_true, y_pred):\n",
    "    return fp(y_true[:,:,:,1], y_pred[:,:,:,1])\n",
    "\n",
    "def fn0(y_true, y_pred):\n",
    "    return fn(y_true[:,:,:,0], y_pred[:,:,:,0])\n",
    "\n",
    "def fn1(y_true, y_pred):\n",
    "    return fn(y_true[:,:,:,1], y_pred[:,:,:,1])\n",
    "\n",
    "def precision_layer(y_true_layer, y_pred_layer):\n",
    "    intersection = K.sum(y_true_layer * y_pred_layer)\n",
    "    return intersection / (K.sum(y_pred_layer) + 1.)\n",
    "\n",
    "def recall_layer(y_true_layer, y_pred_layer):\n",
    "    intersection = K.sum(y_true_layer * y_pred_layer)\n",
    "    return intersection / (K.sum(y_true_layer) + 1.)\n",
    "\n",
    "def precision0(y_true, y_pred):\n",
    "    return precision_layer(y_true[:,:,:,0], y_pred[:,:,:,0])\n",
    "\n",
    "def precision1(y_true, y_pred):\n",
    "    return precision_layer(y_true[:,:,:,1], y_pred[:,:,:,1])\n",
    "\n",
    "def recall0(y_true, y_pred):\n",
    "    return recall_layer(y_true[:,:,:,0], y_pred[:,:,:,0])\n",
    "\n",
    "def recall1(y_true, y_pred):\n",
    "    return recall_layer(y_true[:,:,:,1], y_pred[:,:,:,1])\n",
    "\n",
    "def carf(y_true, y_pred):\n",
    "    presci = precision1(y_true, y_pred)\n",
    "    recall = recall1(y_true, y_pred)\n",
    "    return (1+4)*(presci * recall)/(4*presci + recall)\n",
    "\n",
    "def roadf(y_true, y_pred):\n",
    "    presci = precision0(y_true, y_pred)\n",
    "    recall = recall0(y_true, y_pred)\n",
    "    return (1+0.25)*(presci * recall)/(0.25*presci + recall)\n",
    "\n",
    "def avgf(y_true, y_pred):\n",
    "    return (carf(y_true, y_pred) + roadf(y_true, y_pred)) / 2.0\n",
    "\n",
    "def avgf_loss(y_true, y_pred):\n",
    "    return 10.0 * (1 - roadf(y_true, y_pred)) +  30.0 * (1 - carf(y_true, y_pred))\n",
    "\n",
    "# Loss\n",
    "def fpfn_layer(y_true_layer, y_pred_layer, fpnw=[1.,1.]):\n",
    "    intersection = K.sum(y_true_layer * y_pred_layer)\n",
    "    return (fpnw[0] * (K.sum(y_pred_layer) - intersection)) + (fpnw[1] * (K.sum(y_true_layer) - intersection))\n",
    "\n",
    "def fpfn0(y_true, y_pred, fpnw=[1.,1.]):\n",
    "    return fpfn_layer(y_true[:,:,:,0], y_pred[:,:,:,0], fpnw)\n",
    "\n",
    "def fpfn1(y_true, y_pred, fpnw=[1.,1.]):\n",
    "    return fpfn_layer(y_true[:,:,:,1], y_pred[:,:,:,1], fpnw)\n",
    "\n",
    "def fpfn_loss(y_true, y_pred):\n",
    "    return fpfn0(y_true, y_pred, [1., 2.]) + (3.0 * fpfn1(y_true, y_pred, [1.25, 1.]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Compile model with optimizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_1 (InputLayer)            (None, 576, 800, 3)  0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv1_pad (ZeroPadding2D)       (None, 578, 802, 3)  0           input_1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "conv1 (Conv2D)                  (None, 288, 400, 32) 864         conv1_pad[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv1_bn (BatchNormalization)   (None, 288, 400, 32) 128         conv1[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "conv1_relu (Activation)         (None, 288, 400, 32) 0           conv1_bn[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "conv_pad_1 (ZeroPadding2D)      (None, 290, 402, 32) 0           conv1_relu[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_1 (DepthwiseConv2D)     (None, 288, 400, 32) 288         conv_pad_1[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_1_bn (BatchNormalizatio (None, 288, 400, 32) 128         conv_dw_1[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_1_relu (Activation)     (None, 288, 400, 32) 0           conv_dw_1_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_1 (Conv2D)              (None, 288, 400, 64) 2048        conv_dw_1_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_1_bn (BatchNormalizatio (None, 288, 400, 64) 256         conv_pw_1[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_1_relu (Activation)     (None, 288, 400, 64) 0           conv_pw_1_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pad_2 (ZeroPadding2D)      (None, 290, 402, 64) 0           conv_pw_1_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_2 (DepthwiseConv2D)     (None, 144, 200, 64) 576         conv_pad_2[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_2_bn (BatchNormalizatio (None, 144, 200, 64) 256         conv_dw_2[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_2_relu (Activation)     (None, 144, 200, 64) 0           conv_dw_2_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_2 (Conv2D)              (None, 144, 200, 128 8192        conv_dw_2_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_2_bn (BatchNormalizatio (None, 144, 200, 128 512         conv_pw_2[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_2_relu (Activation)     (None, 144, 200, 128 0           conv_pw_2_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pad_3 (ZeroPadding2D)      (None, 146, 202, 128 0           conv_pw_2_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_3 (DepthwiseConv2D)     (None, 144, 200, 128 1152        conv_pad_3[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_3_bn (BatchNormalizatio (None, 144, 200, 128 512         conv_dw_3[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_3_relu (Activation)     (None, 144, 200, 128 0           conv_dw_3_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_3 (Conv2D)              (None, 144, 200, 128 16384       conv_dw_3_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_3_bn (BatchNormalizatio (None, 144, 200, 128 512         conv_pw_3[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_3_relu (Activation)     (None, 144, 200, 128 0           conv_pw_3_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pad_4 (ZeroPadding2D)      (None, 146, 202, 128 0           conv_pw_3_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_4 (DepthwiseConv2D)     (None, 72, 100, 128) 1152        conv_pad_4[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_4_bn (BatchNormalizatio (None, 72, 100, 128) 512         conv_dw_4[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_4_relu (Activation)     (None, 72, 100, 128) 0           conv_dw_4_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_4 (Conv2D)              (None, 72, 100, 256) 32768       conv_dw_4_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_4_bn (BatchNormalizatio (None, 72, 100, 256) 1024        conv_pw_4[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_4_relu (Activation)     (None, 72, 100, 256) 0           conv_pw_4_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pad_5 (ZeroPadding2D)      (None, 74, 102, 256) 0           conv_pw_4_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_5 (DepthwiseConv2D)     (None, 72, 100, 256) 2304        conv_pad_5[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_5_bn (BatchNormalizatio (None, 72, 100, 256) 1024        conv_dw_5[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_5_relu (Activation)     (None, 72, 100, 256) 0           conv_dw_5_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_5 (Conv2D)              (None, 72, 100, 256) 65536       conv_dw_5_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_5_bn (BatchNormalizatio (None, 72, 100, 256) 1024        conv_pw_5[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_5_relu (Activation)     (None, 72, 100, 256) 0           conv_pw_5_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pad_6 (ZeroPadding2D)      (None, 74, 102, 256) 0           conv_pw_5_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_6 (DepthwiseConv2D)     (None, 36, 50, 256)  2304        conv_pad_6[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_6_bn (BatchNormalizatio (None, 36, 50, 256)  1024        conv_dw_6[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_6_relu (Activation)     (None, 36, 50, 256)  0           conv_dw_6_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_6 (Conv2D)              (None, 36, 50, 512)  131072      conv_dw_6_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_6_bn (BatchNormalizatio (None, 36, 50, 512)  2048        conv_pw_6[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_6_relu (Activation)     (None, 36, 50, 512)  0           conv_pw_6_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pad_7 (ZeroPadding2D)      (None, 38, 52, 512)  0           conv_pw_6_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_7 (DepthwiseConv2D)     (None, 36, 50, 512)  4608        conv_pad_7[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_7_bn (BatchNormalizatio (None, 36, 50, 512)  2048        conv_dw_7[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_7_relu (Activation)     (None, 36, 50, 512)  0           conv_dw_7_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_7 (Conv2D)              (None, 36, 50, 512)  262144      conv_dw_7_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_7_bn (BatchNormalizatio (None, 36, 50, 512)  2048        conv_pw_7[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_7_relu (Activation)     (None, 36, 50, 512)  0           conv_pw_7_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pad_8 (ZeroPadding2D)      (None, 38, 52, 512)  0           conv_pw_7_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_8 (DepthwiseConv2D)     (None, 36, 50, 512)  4608        conv_pad_8[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_8_bn (BatchNormalizatio (None, 36, 50, 512)  2048        conv_dw_8[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_8_relu (Activation)     (None, 36, 50, 512)  0           conv_dw_8_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_8 (Conv2D)              (None, 36, 50, 512)  262144      conv_dw_8_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_8_bn (BatchNormalizatio (None, 36, 50, 512)  2048        conv_pw_8[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_8_relu (Activation)     (None, 36, 50, 512)  0           conv_pw_8_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pad_9 (ZeroPadding2D)      (None, 38, 52, 512)  0           conv_pw_8_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_9 (DepthwiseConv2D)     (None, 36, 50, 512)  4608        conv_pad_9[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_9_bn (BatchNormalizatio (None, 36, 50, 512)  2048        conv_dw_9[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_9_relu (Activation)     (None, 36, 50, 512)  0           conv_dw_9_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_9 (Conv2D)              (None, 36, 50, 512)  262144      conv_dw_9_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_9_bn (BatchNormalizatio (None, 36, 50, 512)  2048        conv_pw_9[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_9_relu (Activation)     (None, 36, 50, 512)  0           conv_pw_9_bn[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv_pad_10 (ZeroPadding2D)     (None, 38, 52, 512)  0           conv_pw_9_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_10 (DepthwiseConv2D)    (None, 36, 50, 512)  4608        conv_pad_10[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_10_bn (BatchNormalizati (None, 36, 50, 512)  2048        conv_dw_10[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_10_relu (Activation)    (None, 36, 50, 512)  0           conv_dw_10_bn[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_10 (Conv2D)             (None, 36, 50, 512)  262144      conv_dw_10_relu[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_10_bn (BatchNormalizati (None, 36, 50, 512)  2048        conv_pw_10[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_10_relu (Activation)    (None, 36, 50, 512)  0           conv_pw_10_bn[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv_pad_11 (ZeroPadding2D)     (None, 38, 52, 512)  0           conv_pw_10_relu[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_11 (DepthwiseConv2D)    (None, 36, 50, 512)  4608        conv_pad_11[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_11_bn (BatchNormalizati (None, 36, 50, 512)  2048        conv_dw_11[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_11_relu (Activation)    (None, 36, 50, 512)  0           conv_dw_11_bn[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_11 (Conv2D)             (None, 36, 50, 512)  262144      conv_dw_11_relu[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_11_bn (BatchNormalizati (None, 36, 50, 512)  2048        conv_pw_11[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_11_relu (Activation)    (None, 36, 50, 512)  0           conv_pw_11_bn[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv_pad_12 (ZeroPadding2D)     (None, 38, 52, 512)  0           conv_pw_11_relu[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_12 (DepthwiseConv2D)    (None, 18, 25, 512)  4608        conv_pad_12[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_12_bn (BatchNormalizati (None, 18, 25, 512)  2048        conv_dw_12[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_12_relu (Activation)    (None, 18, 25, 512)  0           conv_dw_12_bn[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_12 (Conv2D)             (None, 18, 25, 1024) 524288      conv_dw_12_relu[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_12_bn (BatchNormalizati (None, 18, 25, 1024) 4096        conv_pw_12[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_12_relu (Activation)    (None, 18, 25, 1024) 0           conv_pw_12_bn[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv_pad_13 (ZeroPadding2D)     (None, 20, 27, 1024) 0           conv_pw_12_relu[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_13 (DepthwiseConv2D)    (None, 18, 25, 1024) 9216        conv_pad_13[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_13_bn (BatchNormalizati (None, 18, 25, 1024) 4096        conv_dw_13[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_dw_13_relu (Activation)    (None, 18, 25, 1024) 0           conv_dw_13_bn[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_13 (Conv2D)             (None, 18, 25, 1024) 1048576     conv_dw_13_relu[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_13_bn (BatchNormalizati (None, 18, 25, 1024) 4096        conv_pw_13[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv_pw_13_relu (Activation)    (None, 18, 25, 1024) 0           conv_pw_13_bn[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_1 (Conv2D)               (None, 18, 25, 1)    1025        conv_pw_13_relu[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_transpose_1 (Conv2DTrans (None, 72, 100, 256) 16640       conv2d_1[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "add_1 (Add)                     (None, 72, 100, 256) 0           conv2d_transpose_1[0][0]         \n",
      "                                                                 conv_pw_5_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_transpose_2 (Conv2DTrans (None, 144, 200, 128 524416      add_1[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "add_2 (Add)                     (None, 144, 200, 128 0           conv2d_transpose_2[0][0]         \n",
      "                                                                 conv_pw_3_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_transpose_3 (Conv2DTrans (None, 288, 400, 64) 131136      add_2[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "add_3 (Add)                     (None, 288, 400, 64) 0           conv2d_transpose_3[0][0]         \n",
      "                                                                 conv_pw_1_relu[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_transpose_4 (Conv2DTrans (None, 576, 800, 2)  2050        add_3[0][0]                      \n",
      "==================================================================================================\n",
      "Total params: 3,904,131\n",
      "Trainable params: 675,267\n",
      "Non-trainable params: 3,228,864\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# compile the model (should be done *after* setting layers to non-trainable)\n",
    "learning_rate = 1e-3\n",
    "model.compile(optimizer=SGD(lr=learning_rate, momentum=0.9, nesterov=True), loss=avgf_loss, metrics=[precision0, recall0, precision1, recall1, roadf, carf, avgf])\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Image data accessors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def getLabelFrames(seg_file):\n",
    "    seg = cv2.imread(seg_file)[0:IMAGE_HT, 0:IMAGE_WD, :]\n",
    "    if np.any(seg[:,:,0:2] != 0):\n",
    "        raise Exception('bad seg data: non zero values in channel 1 & 2')\n",
    "    seg = seg[:,:,2]\n",
    "    vehicle_frame = np.zeros(seg.shape)\n",
    "    road_frame = np.zeros(seg.shape)\n",
    "\n",
    "    road = np.logical_or(seg==CarlaLabels['Roads'], seg==CarlaLabels['RoadLines'])\n",
    "    vehicles = seg==CarlaLabels['Vehicles']\n",
    "    road_frame[road] = 1\n",
    "    vehicle_frame[vehicles] = 1\n",
    "    vehicle_frame[CarlaHood] = 0\n",
    "    return (road_frame, vehicle_frame)\n",
    "\n",
    "def getImage(img_file):\n",
    "    x = cv2.imread(img_file)[0:IMAGE_HT, 0:IMAGE_WD, :]\n",
    "    cv2.cvtColor(x, cv2.COLOR_BGR2RGB)\n",
    "    return x"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training and Validation Sets\n",
    "\n",
    "- Sequence class to wrap file list and serve out image & label data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training data size: 83130 \n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABqCAYAAACyAZq/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec9099908>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABqCAYAAACyAZq/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec52a6198>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABqCAYAAACyAZq/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n8/8BKT4a9vl3cogAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ebf9d4c50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABqCAYAAACyAZq/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nnhXHC73H4/GsOF7oPR6PZ8X5f1tUlaWwZPTFAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8f5354a160>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABqCAYAAACyAZq/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n0DscDseG44Te4XA4Npz/B40oUvZIn+E+AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ecbfb1940>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABqCAYAAACyAZq/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n3sHBwWHC8T8BdS+DUFsTccsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ef80ff8d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABqCAYAAACyAZq/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n8VfmOd2LKwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec7b7b908>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABqCAYAAACyAZq/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec52a6cc0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABqCAYAAACyAZq/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/wcE4d3KiXGPBQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ef83a4080>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABqCAYAAACyAZq/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n+S/+H1YrBIPz1QAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ef84195c0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "label_list1 = [] #glob.glob(os.path.join(data_dir, 'Train-old/CameraSeg/*0.png')) + glob.glob(os.path.join(data_dir, 'Train-old/CameraSeg/*5.png'))\n",
    "\n",
    "with open('75000.txt', 'r') as train_list:\n",
    "    label_list = [x.strip() for x in train_list if x.strip() not in label_list1]\n",
    "\n",
    "label_list = label_list1 + label_list\n",
    "print(\"Training data size: %d \" % len(label_list))\n",
    "num_labels = len(label_list)\n",
    "np.random.shuffle(label_list)\n",
    "\n",
    "#val_size = num_labels*20//100\n",
    "# if val_size > 3000:\n",
    "#     val_size = 3000\n",
    "# train_y = label_list[val_size:]\n",
    "# test_y = label_list[0:val_size]\n",
    "# print(\"Training set: {0}, Test set: {1}\".format(len(train_y), len(test_y)))\n",
    "# train_x = [re.sub('CameraSeg', 'CameraRGB', x) for x in train_y]\n",
    "# test_x = [re.sub('CameraSeg', 'CameraRGB', x) for x in test_y]\n",
    "\n",
    "train_y = label_list\n",
    "train_x = [re.sub('CameraSeg', 'CameraRGB', x) for x in train_y]\n",
    "\n",
    "test_y = glob.glob(os.path.join('./data/', 'Train/CameraSeg/*.png'))\n",
    "np.random.shuffle(test_y)\n",
    "test_x = [re.sub('CameraSeg', 'CameraRGB', x) for x in test_y]\n",
    "\n",
    "for i in range(10):\n",
    "    x = getImage(test_x[i])\n",
    "    road_frame, vehicle_frame = getLabelFrames(test_y[i])\n",
    "    plt.figure()\n",
    "    plt.subplot(1,3,1)\n",
    "    plt.imshow(x)\n",
    "    plt.subplot(1,3,2)\n",
    "    plt.imshow(road_frame, cmap='gray')\n",
    "    plt.subplot(1,3,3)\n",
    "    plt.imshow(vehicle_frame, cmap='gray')\n",
    "    plt.show()\n",
    "\n",
    "class DataGenerator(Sequence):\n",
    "\n",
    "    def __init__(self, x, y, batch_size, input_preprocessor):\n",
    "        self.y = y\n",
    "        self.x = x\n",
    "        self.batch_size = batch_size\n",
    "        self.input_preprocessor = input_preprocessor\n",
    "        self.minidx = 1e+10\n",
    "        self.maxidx = -1e+10\n",
    "\n",
    "    def getResetIdx(self):\n",
    "        min_idx = self.minidx\n",
    "        max_idx = self.maxidx\n",
    "        self.minidx = 1e+10\n",
    "        self.maxidx = -1e+10\n",
    "        return (min_idx,max_idx)       \n",
    "\n",
    "    def __len__(self):\n",
    "        return math.ceil(len(self.x) / float(self.batch_size))\n",
    "\n",
    "    def inputimage(self, file_name):\n",
    "        return self.input_preprocessor(getImage(file_name).astype(np.float32))\n",
    "\n",
    "    def labelread(self, file_name):\n",
    "        tup = getLabelFrames(file_name)\n",
    "        if num_classes > 1:\n",
    "            label = np.dstack(tup)\n",
    "        else:\n",
    "            label = tup[1]\n",
    "        return label#.reshape(-1, num_classes)\n",
    "\n",
    "    def __getitem__(self, idx):\n",
    "        if idx < self.minidx:\n",
    "            self.minidx = idx\n",
    "        if idx > self.maxidx:\n",
    "            self.maxidx = idx\n",
    "        batch_x = self.x[idx * self.batch_size: (idx + 1) * self.batch_size]\n",
    "        batch_y = self.y[idx * self.batch_size: (idx + 1) * self.batch_size]\n",
    "        return np.array([self.inputimage(file_name) for file_name in batch_x]), np.array([self.labelread(file_name) for file_name in batch_y])\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Pre-process image data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def preprocess_input(x):\n",
    "    return (x / 127.5) - 1.\n",
    "\n",
    "batch_size = 30\n",
    "train_gen = DataGenerator(train_x, train_y, batch_size, input_preprocessor = preprocess_input)\n",
    "val_gen = DataGenerator(test_x, test_y, batch_size, input_preprocessor = preprocess_input)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training\n",
    "\n",
    "- Collects metrics\n",
    "- Logs to tensorboard\n",
    "- Checkpoints weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training for: 20 epochs, 133 steps each, 30 images per step\n",
      "Epoch 1/20\n",
      "133/133 [==============================] - 343s 3s/step - loss: 1.5354 - precision0: 0.9978 - recall0: 0.9703 - precision1: 0.9091 - recall1: 0.9697 - roadf: 0.9922 - carf: 0.9569 - avgf: 0.9745 - val_loss: 2.1791 - val_precision0: 0.9975 - val_recall0: 0.9735 - val_precision1: 0.8777 - val_recall1: 0.9509 - val_roadf: 0.9926 - val_carf: 0.9353 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.17914, saving model to /home/anand/lyft/CARLA_0.8.2/PythonClient/mn12_0604-01-0.97.hdf5\n",
      "Epoch 2/20\n",
      "133/133 [==============================] - 272s 2s/step - loss: 1.5389 - precision0: 0.9978 - recall0: 0.9701 - precision1: 0.9088 - recall1: 0.9696 - roadf: 0.9921 - carf: 0.9568 - avgf: 0.9745 - val_loss: 2.1802 - val_precision0: 0.9975 - val_recall0: 0.9736 - val_precision1: 0.8790 - val_recall1: 0.9505 - val_roadf: 0.9926 - val_carf: 0.9353 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00002: val_loss did not improve from 2.17914\n",
      "Epoch 3/20\n",
      "133/133 [==============================] - 264s 2s/step - loss: 1.5121 - precision0: 0.9977 - recall0: 0.9703 - precision1: 0.9100 - recall1: 0.9704 - roadf: 0.9921 - carf: 0.9577 - avgf: 0.9749 - val_loss: 2.1812 - val_precision0: 0.9974 - val_recall0: 0.9738 - val_precision1: 0.8813 - val_recall1: 0.9498 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00003: val_loss did not improve from 2.17914\n",
      "Epoch 4/20\n",
      "133/133 [==============================] - 338s 3s/step - loss: 1.5635 - precision0: 0.9977 - recall0: 0.9703 - precision1: 0.9071 - recall1: 0.9690 - roadf: 0.9921 - carf: 0.9560 - avgf: 0.9740 - val_loss: 2.1808 - val_precision0: 0.9975 - val_recall0: 0.9734 - val_precision1: 0.8788 - val_recall1: 0.9505 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00004: val_loss did not improve from 2.17914\n",
      "Epoch 5/20\n",
      "133/133 [==============================] - 274s 2s/step - loss: 1.5172 - precision0: 0.9977 - recall0: 0.9701 - precision1: 0.9100 - recall1: 0.9702 - roadf: 0.9921 - carf: 0.9575 - avgf: 0.9748 - val_loss: 2.1795 - val_precision0: 0.9975 - val_recall0: 0.9734 - val_precision1: 0.8812 - val_recall1: 0.9498 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00005: val_loss did not improve from 2.17914\n",
      "Epoch 6/20\n",
      "133/133 [==============================] - 263s 2s/step - loss: 1.5401 - precision0: 0.9977 - recall0: 0.9702 - precision1: 0.9076 - recall1: 0.9699 - roadf: 0.9921 - carf: 0.9567 - avgf: 0.9744 - val_loss: 2.1820 - val_precision0: 0.9974 - val_recall0: 0.9737 - val_precision1: 0.8836 - val_recall1: 0.9490 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00006: val_loss did not improve from 2.17914\n",
      "\n",
      "Epoch 00006: ReduceLROnPlateau reducing learning rate to 0.0003000000142492354.\n",
      "Epoch 7/20\n",
      "133/133 [==============================] - 265s 2s/step - loss: 1.5453 - precision0: 0.9978 - recall0: 0.9702 - precision1: 0.9071 - recall1: 0.9697 - roadf: 0.9922 - carf: 0.9565 - avgf: 0.9743 - val_loss: 2.1798 - val_precision0: 0.9974 - val_recall0: 0.9737 - val_precision1: 0.8800 - val_recall1: 0.9502 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00007: val_loss did not improve from 2.17914\n",
      "Epoch 8/20\n",
      "133/133 [==============================] - 266s 2s/step - loss: 1.5317 - precision0: 0.9978 - recall0: 0.9703 - precision1: 0.9092 - recall1: 0.9698 - roadf: 0.9922 - carf: 0.9570 - avgf: 0.9746 - val_loss: 2.1788 - val_precision0: 0.9974 - val_recall0: 0.9739 - val_precision1: 0.8787 - val_recall1: 0.9506 - val_roadf: 0.9926 - val_carf: 0.9353 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00008: val_loss improved from 2.17914 to 2.17882, saving model to /home/anand/lyft/CARLA_0.8.2/PythonClient/mn12_0604-08-0.97.hdf5\n",
      "Epoch 9/20\n",
      "133/133 [==============================] - 263s 2s/step - loss: 1.5556 - precision0: 0.9978 - recall0: 0.9703 - precision1: 0.9078 - recall1: 0.9691 - roadf: 0.9922 - carf: 0.9562 - avgf: 0.9742 - val_loss: 2.1800 - val_precision0: 0.9974 - val_recall0: 0.9737 - val_precision1: 0.8818 - val_recall1: 0.9496 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00009: val_loss did not improve from 2.17882\n",
      "Epoch 10/20\n",
      "133/133 [==============================] - 263s 2s/step - loss: 1.5089 - precision0: 0.9977 - recall0: 0.9698 - precision1: 0.9098 - recall1: 0.9706 - roadf: 0.9920 - carf: 0.9578 - avgf: 0.9749 - val_loss: 2.1793 - val_precision0: 0.9974 - val_recall0: 0.9737 - val_precision1: 0.8805 - val_recall1: 0.9500 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00010: val_loss did not improve from 2.17882\n",
      "Epoch 11/20\n",
      "133/133 [==============================] - 254s 2s/step - loss: 1.5533 - precision0: 0.9978 - recall0: 0.9702 - precision1: 0.9068 - recall1: 0.9695 - roadf: 0.9921 - carf: 0.9563 - avgf: 0.9742 - val_loss: 2.1785 - val_precision0: 0.9975 - val_recall0: 0.9736 - val_precision1: 0.8778 - val_recall1: 0.9509 - val_roadf: 0.9926 - val_carf: 0.9353 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00011: val_loss improved from 2.17882 to 2.17854, saving model to /home/anand/lyft/CARLA_0.8.2/PythonClient/mn12_0604-11-0.97.hdf5\n",
      "Epoch 12/20\n",
      "133/133 [==============================] - 271s 2s/step - loss: 1.5583 - precision0: 0.9978 - recall0: 0.9702 - precision1: 0.9062 - recall1: 0.9695 - roadf: 0.9921 - carf: 0.9561 - avgf: 0.9741 - val_loss: 2.1792 - val_precision0: 0.9975 - val_recall0: 0.9736 - val_precision1: 0.8799 - val_recall1: 0.9502 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00012: val_loss did not improve from 2.17854\n",
      "Epoch 13/20\n",
      "133/133 [==============================] - 272s 2s/step - loss: 1.5135 - precision0: 0.9978 - recall0: 0.9706 - precision1: 0.9096 - recall1: 0.9704 - roadf: 0.9922 - carf: 0.9576 - avgf: 0.9749 - val_loss: 2.1801 - val_precision0: 0.9974 - val_recall0: 0.9739 - val_precision1: 0.8810 - val_recall1: 0.9498 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00013: val_loss did not improve from 2.17854\n",
      "Epoch 14/20\n",
      "133/133 [==============================] - 261s 2s/step - loss: 1.5514 - precision0: 0.9977 - recall0: 0.9705 - precision1: 0.9079 - recall1: 0.9693 - roadf: 0.9922 - carf: 0.9563 - avgf: 0.9742 - val_loss: 2.1797 - val_precision0: 0.9973 - val_recall0: 0.9740 - val_precision1: 0.8796 - val_recall1: 0.9503 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00014: val_loss did not improve from 2.17854\n",
      "Epoch 15/20\n",
      "133/133 [==============================] - 258s 2s/step - loss: 1.5255 - precision0: 0.9978 - recall0: 0.9706 - precision1: 0.9091 - recall1: 0.9700 - roadf: 0.9922 - carf: 0.9572 - avgf: 0.9747 - val_loss: 2.1793 - val_precision0: 0.9975 - val_recall0: 0.9736 - val_precision1: 0.8798 - val_recall1: 0.9503 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00015: val_loss did not improve from 2.17854\n",
      "Epoch 16/20\n",
      "133/133 [==============================] - 261s 2s/step - loss: 1.5706 - precision0: 0.9977 - recall0: 0.9702 - precision1: 0.9066 - recall1: 0.9688 - roadf: 0.9921 - carf: 0.9557 - avgf: 0.9739 - val_loss: 2.1789 - val_precision0: 0.9975 - val_recall0: 0.9736 - val_precision1: 0.8785 - val_recall1: 0.9506 - val_roadf: 0.9926 - val_carf: 0.9353 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00016: val_loss did not improve from 2.17854\n",
      "\n",
      "Epoch 00016: ReduceLROnPlateau reducing learning rate to 9.000000427477062e-05.\n",
      "Epoch 17/20\n",
      "133/133 [==============================] - 262s 2s/step - loss: 1.5547 - precision0: 0.9977 - recall0: 0.9703 - precision1: 0.9062 - recall1: 0.9696 - roadf: 0.9921 - carf: 0.9562 - avgf: 0.9742 - val_loss: 2.1792 - val_precision0: 0.9974 - val_recall0: 0.9737 - val_precision1: 0.8805 - val_recall1: 0.9500 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00017: val_loss did not improve from 2.17854\n",
      "Epoch 18/20\n",
      "133/133 [==============================] - 260s 2s/step - loss: 1.5705 - precision0: 0.9978 - recall0: 0.9705 - precision1: 0.9058 - recall1: 0.9690 - roadf: 0.9922 - carf: 0.9557 - avgf: 0.9739 - val_loss: 2.1793 - val_precision0: 0.9974 - val_recall0: 0.9737 - val_precision1: 0.8800 - val_recall1: 0.9502 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00018: val_loss did not improve from 2.17854\n",
      "Epoch 19/20\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "133/133 [==============================] - 264s 2s/step - loss: 1.5571 - precision0: 0.9977 - recall0: 0.9699 - precision1: 0.9065 - recall1: 0.9695 - roadf: 0.9920 - carf: 0.9562 - avgf: 0.9741 - val_loss: 2.1796 - val_precision0: 0.9974 - val_recall0: 0.9737 - val_precision1: 0.8806 - val_recall1: 0.9500 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00019: val_loss did not improve from 2.17854\n",
      "Epoch 20/20\n",
      "133/133 [==============================] - 262s 2s/step - loss: 1.5474 - precision0: 0.9976 - recall0: 0.9708 - precision1: 0.9079 - recall1: 0.9694 - roadf: 0.9922 - carf: 0.9565 - avgf: 0.9743 - val_loss: 2.1801 - val_precision0: 0.9975 - val_recall0: 0.9735 - val_precision1: 0.8808 - val_recall1: 0.9499 - val_roadf: 0.9926 - val_carf: 0.9352 - val_avgf: 0.9639\n",
      "\n",
      "Epoch 00020: val_loss did not improve from 2.17854\n"
     ]
    }
   ],
   "source": [
    "if trainingMode:\n",
    "    steps_per_epoch = 4000 // batch_size\n",
    "    epochs_for_1pass = len(train_gen) // steps_per_epoch\n",
    "    epochs = 1 * epochs_for_1pass\n",
    "    print(\"Training for: {0} epochs, {1} steps each, {2} images per step\".format(epochs, steps_per_epoch, batch_size))\n",
    "    checkpoint_file = os.path.join(data_dir, '%s-{epoch:02d}-{avgf:.2f}.hdf5' % (model_name))\n",
    "    saveWeights = ModelCheckpoint(filepath=checkpoint_file, verbose=1, save_best_only=True)\n",
    "    tensorboardLog = TensorBoard(log_dir='./logs/%s' % model_name)\n",
    "    #train_gen_indexes = LambdaCallback(on_epoch_end=lambda batch,logs: print(\"trained on index range: {}\".format(train_gen.getResetIdx())))\n",
    "    lradjust = ReduceLROnPlateau(monitor='val_loss', mode='min', factor=0.3, patience=5, verbose=1, cooldown=2, min_lr=1e-10)\n",
    "    #lradjust = LearningRateScheduler(schedule=lambda e: 1e-3 * (0.1**(e // epochs_for_1pass)), verbose=1 )\n",
    "    model.fit_generator(train_gen, validation_data = val_gen, \n",
    "                        epochs=epochs, steps_per_epoch=steps_per_epoch,\n",
    "                        callbacks=[saveWeights, tensorboardLog, lradjust],\n",
    "                        shuffle=True, use_multiprocessing=True, workers = 3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test trained model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1000/1000 [==============================] - 37s 37ms/step\n",
      "Loss: 5.4275, preci0:0.9975 recall0: 0.9733, preci1: 0.7917, recall1:0.8388, roadf:0.9925, carf:0.8270, avgf:0.8270\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nypijQq8oijLmqNAriqKMOf8f81IeSzyc8pkAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ef81cbe80>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8eb5aa5ba8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8eb56288d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nLJY1zv8Dsokf5RfbMD0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8eb5381a20>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ncILQBwKBwAYnCH0gEAhscP4/ULY9hDlf3ZsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8eb53d2128>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nmKNCryiKMuao0CuKoow5KvSKoihjzv8HgOw9vDn/s84AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8eb5a381d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZkmrcYNvwgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8eb50e2240>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec521c160>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nFjhe6D0ej2eB8/8BTVkLMZCYmzAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec514cbe0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec4fd3860>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nvSRJ0gQnE70kSdIEJxO9JEnSBPf/AbgSKtYcgFFiAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec4ef8a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nOV7oPR6PZ5njhd7j8XiWOf8f5CXYmY3Vf3EAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec4d9bb70>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec4c47be0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nVqfSkVgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec4bfae10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJxO9JEnSFCcTvSRJ0hT3/wEc7wMQE60qUQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec47c8e10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nmJMnX/uGXgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec46f3080>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nNU3T5jmd6DVN0+Y5neg1TdPmuf8P4hXlGFSPElEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec4598240>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nlryACQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec44431d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAoF1zv8Hsi1C8DecwycAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec4381f98>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAABYCAYAAAD7szwZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n0Wuapi1xOtFrmqYtcTrRa5qmLXH/Hxa4gFEhWzAIAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ec426a710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_y = glob.glob(os.path.join('./data/', 'Train/CameraSeg/*.png'))\n",
    "np.random.shuffle(test_y)\n",
    "test_x = [re.sub('CameraSeg', 'CameraRGB', x) for x in test_y]\n",
    "test_gen = DataGenerator(test_x, test_y, 1, input_preprocessor = preprocess_input)\n",
    "\n",
    "result = model.evaluate_generator(test_gen, max_queue_size=10, workers=3, use_multiprocessing=True, verbose=1)\n",
    "print(\"Loss: {0:.4f}, preci0:{1:.4f} recall0: {2:.4f}, preci1: {3:.4f}, recall1:{4:.4f}, roadf:{5:.4f}, carf:{6:.4f}, avgf:{6:.4f}\".format(*result))\n",
    "\n",
    "avg_loss = 0\n",
    "avg_dc0 = 0\n",
    "avg_dc1 = 0\n",
    "for idx in range(0,20):\n",
    "    X, label = test_gen[idx]\n",
    "    pred = model.predict(X)\n",
    "    pred = np.squeeze(pred)\n",
    "    pred = np.rint(pred).astype(np.float32)\n",
    "    label = np.squeeze(label).astype(np.float32)\n",
    "\n",
    "    plt.figure()\n",
    "    num_imgs = 1 + (2*num_classes)\n",
    "    \n",
    "    img = getImage(test_gen.x[idx])\n",
    "    ax = plt.subplot(1, num_imgs, 1)\n",
    "    plt.axis('off')\n",
    "    plt.imshow(img)\n",
    "    ax.set_title('image')\n",
    "    \n",
    "    for lidx in range(num_classes):\n",
    "        ax=plt.subplot(1, num_imgs, 2 + lidx)\n",
    "        plt.axis('off')\n",
    "        plt.imshow(label[:,:,lidx], cmap='gray')\n",
    "        ax.set_title('label')\n",
    "\n",
    "    for pidx in range(num_classes):\n",
    "        ax=plt.subplot(1, num_imgs, 2 + num_classes + pidx)\n",
    "        plt.axis('off')\n",
    "        plt.imshow(pred[:,:,pidx], cmap='gray')\n",
    "        ax.set_title('pred')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Export final model\n",
    "\n",
    "- HDF5 for Keras\n",
    "- Protobuf pb,pbtxt for TensorFlow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "if trainingMode:\n",
    "#     train_y = glob.glob(os.path.join('./data/', 'Train/CameraSeg/*.png'))\n",
    "#     np.random.shuffle(train_y)\n",
    "#     train_x = [re.sub('CameraSeg', 'CameraRGB', x) for x in train_y]\n",
    "#     train_gen = DataGenerator(train_x, train_y, batch_size, input_preprocessor = preprocess_input)\n",
    "#     model.fit_generator(train_gen, \n",
    "#                         epochs=1, callbacks=[saveWeights], \n",
    "#                         shuffle=True, use_multiprocessing=True, workers = 3)\n",
    "    iteration = 0\n",
    "    final_model_file = os.path.join(data_dir, '%d_%s_weights_%d.hdf5' % (iteration, model_name, num_classes))\n",
    "    while os.path.exists(final_model_file):\n",
    "        iteration += 1\n",
    "        final_model_file = os.path.join(data_dir, '%d_%s_weights_%d.hdf5' % (iteration, model_name, num_classes))\n",
    "    model.save(final_model_file, overwrite=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "output nodes names are:  ['output_node0']\n",
      "INFO:tensorflow:Froze 145 variables.\n",
      "Converted 145 variables to const ops.\n",
      "INFO:tensorflow:Froze 145 variables.\n",
      "Converted 145 variables to const ops.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'/home/anand/lyft/CARLA_0.8.2/PythonClient/models/mn12_0604_saved_model_quantized.pb'"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow.python.framework import graph_util\n",
    "from tensorflow.python.framework import graph_io\n",
    "from tensorflow.tools.graph_transforms import TransformGraph\n",
    "\n",
    "num_output = 1\n",
    "output_node_prefix = 'output_node'\n",
    "\n",
    "pred = [None]*num_output\n",
    "pred_node_names = [None]*num_output\n",
    "for i in range(num_output):\n",
    "    pred_node_names[i] = output_node_prefix+str(i)\n",
    "    pred[i] = tf.identity(model.outputs[i], name=pred_node_names[i])\n",
    "print('output nodes names are: ', pred_node_names)\n",
    "\n",
    "sess = K.get_session()\n",
    "out_fldr = os.path.join(data_dir, 'models')\n",
    "tf.train.write_graph(sess.graph.as_graph_def(), out_fldr, '%s_graph_def.pbtxt' % model_name, as_text=True)\n",
    "constant_graph = graph_util.convert_variables_to_constants(sess, sess.graph.as_graph_def(), pred_node_names)\n",
    "graph_io.write_graph(constant_graph, out_fldr, '%s_saved_model.pb' % model_name, as_text=False)\n",
    "\n",
    "transforms = [\"quantize_weights\", \"quantize_nodes\"]\n",
    "transformed_graph_def = TransformGraph(sess.graph.as_graph_def(), [], pred_node_names, transforms)\n",
    "constant_graph = graph_util.convert_variables_to_constants(sess, transformed_graph_def, pred_node_names)\n",
    "graph_io.write_graph(constant_graph, out_fldr, '%s_saved_model_quantized.pb' % model_name, as_text=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}