File: upload jupyter notebook file

p2SemanticSegmentation.ipynb

@@ -0,0 +1,270 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {
+    "collapsed": true
+   },
+   "outputs": [],
+   "source": [
+    "### Design and Programming by Mojtaba Valipour @ Shiraz University, Udacity"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {
+    "collapsed": false
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Copyright is important @2017\n"
+     ]
+    }
+   ],
+   "source": [
+    "'''\n",
+    "References:\n",
+    "    - \n",
+    "    - https://www.tensorflow.org\n",
+    "'''\n",
+    "print('Copyright is important @2017')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {
+    "collapsed": false
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "[name: \"/cpu:0\"\n",
+      "device_type: \"CPU\"\n",
+      "memory_limit: 268435456\n",
+      "locality {\n",
+      "}\n",
+      "incarnation: 16308118712357310879\n",
+      ", name: \"/gpu:0\"\n",
+      "device_type: \"GPU\"\n",
+      "memory_limit: 12051077530\n",
+      "locality {\n",
+      "  bus_id: 1\n",
+      "}\n",
+      "incarnation: 16559124723434357649\n",
+      "physical_device_desc: \"device: 0, name: GeForce GTX TITAN X, pci bus id: 0000:0a:00.0\"\n",
+      ", name: \"/gpu:1\"\n",
+      "device_type: \"GPU\"\n",
+      "memory_limit: 12051077530\n",
+      "locality {\n",
+      "  bus_id: 1\n",
+      "}\n",
+      "incarnation: 10025864280914671720\n",
+      "physical_device_desc: \"device: 1, name: GeForce GTX TITAN X, pci bus id: 0000:09:00.0\"\n",
+      ", name: \"/gpu:2\"\n",
+      "device_type: \"GPU\"\n",
+      "memory_limit: 12049085236\n",
+      "locality {\n",
+      "  bus_id: 1\n",
+      "}\n",
+      "incarnation: 10544825921621946364\n",
+      "physical_device_desc: \"device: 2, name: GeForce GTX TITAN X, pci bus id: 0000:06:00.0\"\n",
+      ", name: \"/gpu:3\"\n",
+      "device_type: \"GPU\"\n",
+      "memory_limit: 12047777792\n",
+      "locality {\n",
+      "  bus_id: 1\n",
+      "}\n",
+      "incarnation: 9239224823666292773\n",
+      "physical_device_desc: \"device: 3, name: GeForce GTX TITAN X, pci bus id: 0000:05:00.0\"\n",
+      "]\n"
+     ]
+    }
+   ],
+   "source": [
+    "import tensorflow as tf\n",
+    "import os\n",
+    "%matplotlib inline\n",
+    "#os.environ[\"CUDA_VISIBLE_DEVICES\"] = '0,1,2'\n",
+    "from tensorflow.python.client import device_lib\n",
+    "print(device_lib.list_local_devices())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {
+    "collapsed": false
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "TensorFlow Version: 1.3.0\n",
+      "Default GPU Device: /gpu:0\n",
+      "Tests Passed\n",
+      "Tests Passed\n",
+      "Tests Passed\n",
+      "INFO:tensorflow:Restoring parameters from b'./data/vgg/variables/variables'\n"
+     ]
+    }
+   ],
+   "source": [
+    "import os.path\n",
+    "import tensorflow as tf\n",
+    "import helper\n",
+    "import warnings\n",
+    "from distutils.version import LooseVersion\n",
+    "import project_tests as tests\n",
+    "\n",
+    "\n",
+    "# Check TensorFlow Version\n",
+    "assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer.  You are using {}'.format(tf.__version__)\n",
+    "print('TensorFlow Version: {}'.format(tf.__version__))\n",
+    "\n",
+    "# Check for a GPU\n",
+    "if not tf.test.gpu_device_name():\n",
+    "    warnings.warn('No GPU found. Please use a GPU to train your neural network.')\n",
+    "else:\n",
+    "    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))\n",
+    "\n",
+    "\n",
+    "def run():\n",
+    "    num_classes = 2\n",
+    "    image_shape = (160, 576)\n",
+    "    data_dir = './data'\n",
+    "    runs_dir = './runs'\n",
+    "    tests.test_for_kitti_dataset(data_dir)\n",
+    "\n",
+    "    # Download pretrained vgg model\n",
+    "    helper.maybe_download_pretrained_vgg(data_dir)\n",
+    "\n",
+    "    # OPTIONAL: Train and Inference on the cityscapes dataset instead of the Kitti dataset.\n",
+    "    # You'll need a GPU with at least 10 teraFLOPS to train on.\n",
+    "    #  https://www.cityscapes-dataset.com/\n",
+    "    epochs = 25\n",
+    "    batch_size = 20\n",
+    "    with tf.Session() as sess:\n",
+    "        # Path to vgg model\n",
+    "        vgg_path = os.path.join(data_dir, 'vgg')\n",
+    "        # Create function to get batches\n",
+    "        get_batches_fn = helper.gen_batch_function(os.path.join(data_dir, 'data_road/training'), image_shape)\n",
+    "\n",
+    "        correct_label = tf.placeholder(tf.float32, name='correct_label')\n",
+    "        learning_rate = tf.placeholder(tf.float32, name='learning_rate')\n",
+    "        \n",
+    "        # OPTIONAL: Augment Images for better results\n",
+    "        #  https://datascience.stackexchange.com/questions/5224/how-to-prepare-augment-images-for-neural-network\n",
+    "\n",
+    "        # TODO: Build NN using load_vgg, layers, and optimize function\n",
+    "        input_image, keep_prob, vgg_layer3_out, vgg_layer4_out, vgg_layer7_out = load_vgg(sess, vgg_path)\n",
+    "        nn_last_layer = layers(vgg_layer3_out, vgg_layer4_out, vgg_layer7_out, num_classes)\n",
+    "        logits, train_op, cross_entropy_loss = optimize(nn_last_layer, correct_label, learning_rate, num_classes)\n",
+    "\n",
+    "        # TODO: Train NN using the train_nn function\n",
+    "        train_nn(sess, epochs, batch_size, get_batches_fn, train_op, cross_entropy_loss, input_image,\n",
+    "             correct_label, keep_prob, learning_rate)\n",
+    "\n",
+    "        # TODO: Save inference data using helper.save_inference_samples\n",
+    "        #  helper.save_inference_samples(runs_dir, data_dir, sess, image_shape, logits, keep_prob, input_image)\n",
+    "        helper.save_inference_samples(runs_dir, data_dir, sess, image_shape, logits, keep_prob, input_image)\n",
+    "        # OPTIONAL: Apply the trained model to a video\n",
+    "\n",
+    "\n",
+    "if __name__ == '__main__':\n",
+    "    run()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false
+   },
+   "outputs": [],
+   "source": [
+    "num_classes = 2\n",
+    "image_shape = (160, 576)\n",
+    "data_dir = './data'\n",
+    "runs_dir = './runs'\n",
+    "tests.test_for_kitti_dataset(data_dir)\n",
+    "\n",
+    "# Download pretrained vgg model\n",
+    "helper.maybe_download_pretrained_vgg(data_dir)\n",
+    "\n",
+    "# OPTIONAL: Train and Inference on the cityscapes dataset instead of the Kitti dataset.\n",
+    "# You'll need a GPU with at least 10 teraFLOPS to train on.\n",
+    "#  https://www.cityscapes-dataset.com/\n",
+    "epochs = 25\n",
+    "batch_size = 20\n",
+    "with tf.Session() as sess:\n",
+    "    # Path to vgg model\n",
+    "    vgg_path = os.path.join(data_dir, 'vgg')\n",
+    "    # Create function to get batches\n",
+    "    get_batches_fn = helper.gen_batch_function(os.path.join(data_dir, 'data_road/training'), image_shape)\n",
+    "\n",
+    "    correct_label = tf.placeholder(tf.float32, name='correct_label')\n",
+    "    learning_rate = tf.placeholder(tf.float32, name='learning_rate')\n",
+    "\n",
+    "    # OPTIONAL: Augment Images for better results\n",
+    "    #  https://datascience.stackexchange.com/questions/5224/how-to-prepare-augment-images-for-neural-network\n",
+    "\n",
+    "    # TODO: Build NN using load_vgg, layers, and optimize function\n",
+    "    input_image, keep_prob, vgg_layer3_out, vgg_layer4_out, vgg_layer7_out = load_vgg(sess, vgg_path)\n",
+    "    nn_last_layer = layers(vgg_layer3_out, vgg_layer4_out, vgg_layer7_out, num_classes)\n",
+    "    logits, train_op, cross_entropy_loss = optimize(nn_last_layer, correct_label, learning_rate, num_classes)\n",
+    "\n",
+    "    # TODO: Train NN using the train_nn function\n",
+    "    train_nn(sess, epochs, batch_size, get_batches_fn, train_op, cross_entropy_loss, input_image,\n",
+    "         correct_label, keep_prob, learning_rate)\n",
+    "\n",
+    "    # TODO: Save inference data using helper.save_inference_samples\n",
+    "    #  helper.save_inference_samples(runs_dir, data_dir, sess, image_shape, logits, keep_prob, input_image)\n",
+    "    helper.save_inference_samples(runs_dir, data_dir, sess, image_shape, logits, keep_prob, input_image)\n",
+    "    # OPTIONAL: Apply the trained model to a video\n",
+    "    \n",
+    "    saver = tf.train.Saver()\n",
+    "    save_path = saver.save(sess, \"/tmp/model_01_good.ckpt\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "collapsed": true
+   },
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "anaconda-cloud": {},
+  "kernelspec": {
+   "display_name": "Python [conda env:udacityProjects]",
+   "language": "python",
+   "name": "conda-env-udacityProjects-py"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.5.4"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 1
+}