Full Code of zhangruiskyline/DeepLearning for AI

master 9b8742aa03df cached
149 files
171.7 MB
4.2M tokens
130 symbols
1 requests
Copy disabled (too large) Download .txt
Showing preview only (16,951K chars total). Download the full file to get everything.
Repository: zhangruiskyline/DeepLearning
Branch: master
Commit: 9b8742aa03df
Files: 149
Total size: 171.7 MB

Directory structure:
gitextract_g80rflq0/

├── .gitignore
├── Digit_Recognizer/
│   ├── random_forrest.txt
│   ├── test.csv
│   └── train.csv
├── Keras_CNN_BitTiger.ipynb
├── Keras_FNN_BitTiger.ipynb
├── Quora_questions_similar/
│   ├── .gitignore
│   ├── CNN_Question_pair.py
│   ├── README.md
│   ├── lstm_embedding.py
│   └── similar_questions.py
├── README.md
├── TensorFlow_Examples/
│   ├── .gitignore
│   ├── LICENSE
│   ├── _config.yml
│   ├── assets/
│   │   └── css/
│   │       └── style.scss
│   ├── labs/
│   │   ├── 02_recommendation/
│   │   │   ├── Explicit_Feedback_Neural_Recommender_System.ipynb
│   │   │   ├── Implicit_Feedback_Recsys_with_the_triplet_loss.ipynb
│   │   │   └── solutions/
│   │   │       ├── classification.py
│   │   │       ├── compute_errors.py
│   │   │       ├── deep_explicit_feedback_recsys.py
│   │   │       ├── deep_implicit_feedback_recsys.py
│   │   │       ├── deep_triplet_parameter_count.py
│   │   │       ├── ml-1m.py
│   │   │       ├── similarity.py
│   │   │       └── triplet_parameter_count.py
│   │   ├── 03_LSTM_NLP/
│   │   │   └── NLP_word_vectors_classification.ipynb
│   │   ├── 03_conv_nets/
│   │   │   ├── ConvNets_with_TensorFlow_rendered.ipynb
│   │   │   ├── Pretrained_ConvNets_with_Keras.ipynb
│   │   │   ├── cifar10_input.py
│   │   │   ├── img_emb.h5
│   │   │   ├── process_images.py
│   │   │   └── solutions/
│   │   │       ├── average_as_conv.py
│   │   │       ├── dogs_vs_cats_extract_features.py
│   │   │       ├── dogs_vs_cats_worst_predictions.py
│   │   │       ├── edge_detection.py
│   │   │       ├── mnist_conv.py
│   │   │       ├── mnist_conv_dropout.py
│   │   │       ├── pooling.py
│   │   │       ├── predict_image.py
│   │   │       ├── representations.py
│   │   │       └── strides_padding.py
│   │   └── 04_seq2seq/
│   │       ├── Translation_of_Numeric_Phrases_with_Seq2Seq_rendered.ipynb
│   │       ├── data_download.ipynb
│   │       ├── french_numbers.py
│   │       ├── solutions/
│   │       │   ├── beam_search.py
│   │       │   ├── interpret_output.py
│   │       │   ├── make_input_output.py
│   │       │   └── simple_seq2seq.py
│   │       └── translate.py
│   └── slides/
│       ├── 01_intro_to_deep_learning/
│       │   ├── index.html
│       │   ├── slides.css
│       │   ├── webfont-ubuntu-400-300-100.css
│       │   └── webfont-ubuntu-mono-400-700-400italic.css
│       ├── 02_recommender_systems/
│       │   ├── images/
│       │   │   └── make_dropout_curves.py
│       │   ├── index.html
│       │   ├── slides.css
│       │   ├── webfont-ubuntu-400-300-100.css
│       │   └── webfont-ubuntu-mono-400-700-400italic.css
│       ├── 03_conv_nets/
│       │   ├── index.html
│       │   ├── slides.css
│       │   ├── webfont-ubuntu-400-300-100.css
│       │   └── webfont-ubuntu-mono-400-700-400italic.css
│       ├── 04_conv_nets_2/
│       │   ├── index.html
│       │   ├── slides.css
│       │   ├── webfont-ubuntu-400-300-100.css
│       │   └── webfont-ubuntu-mono-400-700-400italic.css
│       ├── 05_deep_nlp/
│       │   ├── index.html
│       │   ├── slides.css
│       │   ├── webfont-ubuntu-400-300-100.css
│       │   └── webfont-ubuntu-mono-400-700-400italic.css
│       └── 06_expressivity_optimization_generalization/
│           ├── images/
│           │   ├── make_mini_mlp_loss_figures.py
│           │   ├── make_relu_approx_figures.py
│           │   ├── make_relu_composition_figures.py
│           │   └── random_relu_surface.py
│           ├── index.html
│           ├── slides.css
│           ├── webfont-ubuntu-400-300-100.css
│           └── webfont-ubuntu-mono-400-700-400italic.css
├── Workshop/
│   └── sessions/
│       ├── .ipynb_checkpoints/
│       │   ├── 2-3_classify-checkpoint.ipynb
│       │   └── 4_deep-checkpoint.ipynb
│       ├── 13_lstm.ipynb
│       ├── 14_embeddings.ipynb
│       ├── 15_news20.ipynb
│       ├── 16_QA.ipynb
│       ├── 2-3_classify.ipynb
│       ├── 20news_download.sh
│       ├── 4_deep.ipynb
│       ├── 5_convnets.ipynb
│       ├── 6_visualization.ipynb
│       ├── 7_gan.ipynb
│       ├── 8_dream.ipynb
│       ├── 8_style.ipynb
│       ├── detection/
│       │   ├── README.md
│       │   ├── data/
│       │   │   ├── checkpoints/
│       │   │   │   └── README.md
│       │   │   ├── files/
│       │   │   │   ├── VOC2007.pkl
│       │   │   │   ├── gt_pascal.pkl
│       │   │   │   └── prior_boxes_ssd300.pkl
│       │   │   └── out/
│       │   │       └── README.md
│       │   └── ssd_keras/
│       │       ├── PASCAL_VOC/
│       │       │   ├── __init__.py
│       │       │   └── get_data_from_XML.py
│       │       ├── ssd.py
│       │       ├── ssd_layers.py
│       │       ├── ssd_training.py
│       │       ├── ssd_utils.py
│       │       ├── testing.ipynb
│       │       ├── testing_video.ipynb
│       │       └── training.ipynb
│       ├── dream.py
│       ├── glove_download.sh
│       ├── style.py
│       ├── tracking/
│       │   ├── README.md
│       │   ├── data/
│       │   │   ├── ADL-Rundle-6/
│       │   │   │   └── det.txt
│       │   │   ├── ADL-Rundle-8/
│       │   │   │   └── det.txt
│       │   │   ├── ETH-Bahnhof/
│       │   │   │   └── det.txt
│       │   │   ├── ETH-Pedcross2/
│       │   │   │   └── det.txt
│       │   │   ├── ETH-Sunnyday/
│       │   │   │   └── det.txt
│       │   │   ├── KITTI-13/
│       │   │   │   └── det.txt
│       │   │   ├── KITTI-17/
│       │   │   │   └── det.txt
│       │   │   ├── PETS09-S2L1/
│       │   │   │   └── det.txt
│       │   │   ├── TUD-Campus/
│       │   │   │   └── det.txt
│       │   │   ├── TUD-Stadtmitte/
│       │   │   │   └── det.txt
│       │   │   └── Venice-2/
│       │   │       └── det.txt
│       │   ├── sort.py
│       │   ├── tracking_kalman.ipynb
│       │   └── tracking_siamese.ipynb
│       └── utils.py
├── doc/
│   ├── AI_chip.md
│   ├── Basics_Machine_learning.md
│   ├── CNN.md
│   ├── Kaggle.md
│   ├── NLP.md
│   ├── RNN.md
│   ├── Recommendation.md
│   ├── SVM.md
│   ├── Search.md
│   ├── Tensorflow_Keras.md
│   ├── decision_tree.md
│   ├── industrial_design.md
│   ├── pytorch.md
│   └── system.md
├── human-activity-recognition-with-smartphones/
│   ├── .gitignore
│   └── ADL_sensor_data.ipynb
├── keras_FNN_BitTiger.py
├── sklearn_test.py
├── tensor_flow_test.py
└── xgboost/
    ├── .gitignore
    ├── tree_xgboost.ipynb
    └── xgboost_feature.ipynb

================================================
FILE CONTENTS
================================================

================================================
FILE: .gitignore
================================================
*.docx
*.pdf
/book
.idea/*
.ipynb_checkpoints/*
xgboost/.ipynb_checkpoints/*

/TensorFlow_Examples/labs/03_conv_nets/images_resize/*


================================================
FILE: Digit_Recognizer/random_forrest.txt
================================================
from sklearn.ensemble import RandomForestClassifier
from numpy import genfromtxt, savetxt

def main():
    #create the training & test sets, skipping the header row with [1:]
    dataset = genfromtxt(open('Data/train.csv','r'), delimiter=',', dtype='f8')[1:]    
    target = [x[0] for x in dataset]
    train = [x[1:] for x in dataset]
    test = genfromtxt(open('Data/test.csv','r'), delimiter=',', dtype='f8')[1:]
    
    #create and train the random forest
    #multi-core CPUs can use: rf = RandomForestClassifier(n_estimators=100, n_jobs=2)
    rf = RandomForestClassifier(n_estimators=100)
    rf.fit(train, target)

    savetxt('Data/submission2.csv', rf.predict(test), delimiter=',', fmt='%f')

if __name__=="__main__":
    main()

================================================
FILE: Digit_Recognizer/test.csv
================================================
[File too large to display: 48.8 MB]

================================================
FILE: Digit_Recognizer/train.csv
================================================
[File too large to display: 73.2 MB]

================================================
FILE: Keras_CNN_BitTiger.ipynb
================================================
{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Codelab for Convolutional Neural Net\n",
    "\n",
    "All rights reserved.<br \\>\n",
    "This material cannot be published, rewritten or redistributed in whole or part without the authors' written permission.\n",
    "\n",
    "**Convolutional Neural Net** will be abbreviated to **ConvNet** in the following sections.\n",
    "\n",
    "### Index\n",
    "1. [Pre-processing](#Pre-processing)\n",
    "2. [Recipes for ConvNet](#Recipes-Exclusively-for-Convolutional-Neural-Nets)\n",
    "    1. [Network Structure](#Network-Structure)\n",
    "        1. [Activity 1](#Tutorial/workshop-activity-1)\n",
    "    2. [Pooling](#Pooling)\n",
    "        1. [Activity 2](#Tutorial/workshop-activity-2)\n",
    "3. [Define Training Procedure](#Define-Training-Procedure)     \n",
    "4. [Start Training](#Start-Training) and [Observe Training Process](#Observe-Training-Process)\n",
    "5. [Testing](#Define-Testing-Procedure)\n",
    "\n",
    "During tutorial/workshop, attendees will be separated into three groups. Each group will be conducting different activities.<br />\n",
    "\n",
    "Activity: \"Cell\" $\\rightarrow$ \"Run All Below\" for testing the enviroment.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "################################################################\n",
    "#\n",
    "# All rights reserved.\n",
    "#\n",
    "# This is a codelab for Convolutional Neural Net.\n",
    "# Details include:\n",
    "#   - Pre-process dataset\n",
    "#   - Elaborate recipes\n",
    "#   - Define training procedures\n",
    "#   - Train and test models\n",
    "#   - Observe metrics\n",
    "#\n",
    "################################################################\n",
    "from __future__ import print_function\n",
    "\n",
    "import keras.callbacks as cb\n",
    "from keras.datasets import mnist\n",
    "from keras.layers import Activation, Dense, Dropout, Flatten\n",
    "from keras.layers import Convolution2D, MaxPooling2D\n",
    "from keras.models import Sequential\n",
    "from keras.optimizers import SGD\n",
    "from keras.regularizers import l1, l2\n",
    "from keras.utils import np_utils\n",
    "\n",
    "%matplotlib inline\n",
    "from matplotlib import pyplot as plt\n",
    "import numpy as np\n",
    "import time"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pre-processing\n",
    "\n",
    "Pre-processing for ConvNet is slightly different than that of FNNs. Here, you need to transform the data into a 3-D tensor with dimensions [width] $\\times$ [height] $\\times$  [channel].\n",
    "\n",
    "Width and height are from a image.\n",
    "Each channel represents a color.\n",
    "\n",
    "Common techniques for data augmentation:\n",
    "1. Cropping\n",
    "2. Random cropping\n",
    "3. Rotation\n",
    "4. Add random noises\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def PreprocessDataset():\n",
    "    # Load dataset\n",
    "    (x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "    # Set numeric type\n",
    "    x_train = x_train.astype('float32')\n",
    "    x_test = x_test.astype('float32')\n",
    "    # Normalize value to [0, 1]\n",
    "    x_train /= 255\n",
    "    x_test /= 255\n",
    "    # Transform lables to one-hot\n",
    "    y_train = np_utils.to_categorical(y_train, 10)\n",
    "    y_test = np_utils.to_categorical(y_test, 10)\n",
    "    # Reshape: here x_train is re-shaped to [channel] × [width] × [height]\n",
    "    # In other environment, the orders could be different; e.g., [height] × [width] × [channel].\n",
    "    x_train = x_train.reshape(x_train.shape[0], 28,28,1)\n",
    "    x_test = x_test.reshape(x_test.shape[0], 28,28,1)\n",
    "    return [x_train, x_test, y_train, y_test]\n",
    "\n",
    "x_train, x_test, y_train, y_test = PreprocessDataset()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Recipes Exclusively for Convolutional Neural Nets"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Convolutional Layer\n",
    "Network structure refers to components of a convnet.\n",
    "\n",
    "Ingredients:\n",
    "1. **Filters**: define the size of a filter, and the total number of filters in a convolutional layer.\n",
    "2. **Depth** represents the number of layers, usually 10-30.\n",
    "3. **Stride** defines the step at which the filter moves.\n",
    "4. **Padding** of the input data, either no padding or zero-padding.\n",
    "\n",
    "How to calculate the number of parameters\n",
    "\n",
    "Functionalityies of convolutional layers: \n",
    "- Capture patterns\n",
    "- Share parameters (otherwise too many parameters)\n",
    "\n",
    "### Pooling\n",
    "Pooling layer reduces the output of the previous layer.\n",
    "\n",
    "max-k pool (usually k = 1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "num_filters = 16\n",
    "filter_size = 3\n",
    "\n",
    "def DefineModel():\n",
    "    model = Sequential()\n",
    "    # First conv layer with filters of 3x3 pixels; also need to specify input shape.\n",
    "    # The number of filters is set to 16 for shorter runtime.\n",
    "    # Increase the number of filters for better accuracy.\n",
    "    # Strides are defined as in subsample=(1, 1)\n",
    "    model.add(Convolution2D(num_filters, filter_size, filter_size, border_mode='valid', input_shape=(28,28,1)))\n",
    "    # The activation for first layer is ReLU\n",
    "    model.add(Activation('relu'))\n",
    "    # Max pooling\n",
    "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "    # The last layer has the same dimension as the number of classes\n",
    "    model.add(Flatten())\n",
    "    model.add(Dense(10))\n",
    "    # For classification, the activation is softmax\n",
    "    model.add(Activation('softmax'))\n",
    "    # Define optimizer\n",
    "    optmzr = SGD(lr=0.1, clipnorm=5.)\n",
    "    # Define loss function = cross entropy\n",
    "    model.compile(loss='categorical_crossentropy', optimizer=optmzr, metrics=[\"accuracy\"])\n",
    "\n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Define Training Procedure"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def TrainModel(data=None, epochs=20, batch=256):\n",
    "    start_time = time.time()\n",
    "    model = DefineModel()\n",
    "    if data is None:\n",
    "        print(\"Must provide data.\")\n",
    "        return\n",
    "    x_train, x_test, y_train, y_test = data\n",
    "    print('Start training.')\n",
    "    history = model.fit(x_train, y_train, nb_epoch=epochs, batch_size=batch,\n",
    "              validation_data=(x_test, y_test), verbose=1)\n",
    "    print(\"Training took {0} seconds.\".format(time.time() - start_time))\n",
    "    return model, history"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Start Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Start training.\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/20\n",
      "60000/60000 [==============================] - 19s - loss: 0.5634 - acc: 0.8365 - val_loss: 0.3840 - val_acc: 0.8867\n",
      "Epoch 2/20\n",
      "60000/60000 [==============================] - 18s - loss: 0.3200 - acc: 0.9042 - val_loss: 0.3002 - val_acc: 0.9139\n",
      "Epoch 3/20\n",
      "60000/60000 [==============================] - 19s - loss: 0.2721 - acc: 0.9204 - val_loss: 0.2541 - val_acc: 0.9246\n",
      "Epoch 4/20\n",
      "60000/60000 [==============================] - 19s - loss: 0.2370 - acc: 0.9316 - val_loss: 0.2083 - val_acc: 0.9390\n",
      "Epoch 5/20\n",
      "60000/60000 [==============================] - 18s - loss: 0.2022 - acc: 0.9426 - val_loss: 0.1753 - val_acc: 0.9496\n",
      "Epoch 6/20\n",
      "60000/60000 [==============================] - 18s - loss: 0.1733 - acc: 0.9517 - val_loss: 0.1618 - val_acc: 0.9563\n",
      "Epoch 7/20\n",
      "60000/60000 [==============================] - 19s - loss: 0.1498 - acc: 0.9590 - val_loss: 0.1453 - val_acc: 0.9595\n",
      "Epoch 8/20\n",
      "60000/60000 [==============================] - 20s - loss: 0.1315 - acc: 0.9628 - val_loss: 0.1282 - val_acc: 0.9644\n",
      "Epoch 9/20\n",
      "60000/60000 [==============================] - 23s - loss: 0.1175 - acc: 0.9677 - val_loss: 0.1203 - val_acc: 0.9634\n",
      "Epoch 10/20\n",
      "60000/60000 [==============================] - 24s - loss: 0.1063 - acc: 0.9711 - val_loss: 0.1002 - val_acc: 0.9704\n",
      "Epoch 11/20\n",
      "60000/60000 [==============================] - 21s - loss: 0.0973 - acc: 0.9733 - val_loss: 0.0972 - val_acc: 0.9715\n",
      "Epoch 12/20\n",
      "60000/60000 [==============================] - 22s - loss: 0.0900 - acc: 0.9752 - val_loss: 0.0911 - val_acc: 0.9726\n",
      "Epoch 13/20\n",
      "60000/60000 [==============================] - 22s - loss: 0.0845 - acc: 0.9769 - val_loss: 0.0811 - val_acc: 0.9750\n",
      "Epoch 14/20\n",
      "60000/60000 [==============================] - 22s - loss: 0.0793 - acc: 0.9778 - val_loss: 0.0787 - val_acc: 0.9757\n",
      "Epoch 15/20\n",
      "60000/60000 [==============================] - 22s - loss: 0.0752 - acc: 0.9792 - val_loss: 0.0711 - val_acc: 0.9788\n",
      "Epoch 16/20\n",
      "60000/60000 [==============================] - 22s - loss: 0.0717 - acc: 0.9804 - val_loss: 0.0713 - val_acc: 0.9779\n",
      "Epoch 17/20\n",
      "60000/60000 [==============================] - 23s - loss: 0.0687 - acc: 0.9810 - val_loss: 0.0685 - val_acc: 0.9791\n",
      "Epoch 18/20\n",
      "60000/60000 [==============================] - 21s - loss: 0.0656 - acc: 0.9820 - val_loss: 0.0672 - val_acc: 0.9789\n",
      "Epoch 19/20\n",
      "60000/60000 [==============================] - 23s - loss: 0.0634 - acc: 0.9819 - val_loss: 0.0637 - val_acc: 0.9802\n",
      "Epoch 20/20\n",
      "60000/60000 [==============================] - 23s - loss: 0.0611 - acc: 0.9832 - val_loss: 0.0631 - val_acc: 0.9792\n",
      "Training took 431.017152786 seconds.\n"
     ]
    }
   ],
   "source": [
    "trained_model, training_history = TrainModel(data=[x_train, x_test, y_train, y_test])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Observe Training Process\n",
    "Plot the variation of loss during each batch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAh4AAAGHCAYAAAD/QltcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nfEbBQ0RERHxGwUNERER85v8DtCtZ8f677g0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10ebef090>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAicAAAGHCAYAAABrpPKuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n5f8DDh9nCRiYwFYAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10eba5050>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def PlotHistory(train_value, test_value, value_is_loss_or_acc):\n",
    "    f, ax = plt.subplots()\n",
    "    ax.plot([None] + train_value, 'o-')\n",
    "    ax.plot([None] + test_value, 'x-')\n",
    "    # Plot legend and use the best location automatically: loc = 0.\n",
    "    ax.legend(['Train ' + value_is_loss_or_acc, 'Validation ' + value_is_loss_or_acc], loc = 0) \n",
    "    ax.set_title('Training/Validation ' + value_is_loss_or_acc + ' per Epoch')\n",
    "    ax.set_xlabel('Epoch')\n",
    "    ax.set_ylabel(value_is_loss_or_acc)  \n",
    "    \n",
    "PlotHistory(training_history.history['loss'], training_history.history['val_loss'], 'Loss')\n",
    "PlotHistory(training_history.history['acc'], training_history.history['val_acc'], 'Accuracy')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def TestModel(model=None, data=None):\n",
    "    if model is None:\n",
    "        print(\"Must provide a trained model.\")\n",
    "        return\n",
    "    if data is None:\n",
    "        print(\"Must provide data.\")\n",
    "        return\n",
    "    x_test, y_test = data\n",
    "    scores = model.evaluate(x_test, y_test)\n",
    "    return scores"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test the trained model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 9984/10000 [============================>.] - ETA: 0sTest loss 0.0631, accuracy 97.92%\n"
     ]
    }
   ],
   "source": [
    "test_score = TestModel(model=trained_model, data=[x_test, y_test])\n",
    "print(\"Test loss {:.4f}, accuracy {:.2f}%\".format(test_score[0], test_score[1] * 100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Visualize Convolutional Layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def ShowConvolutionOutput(input_data, show_pic_idx):\n",
    "    from keras import backend as K\n",
    "    first_conv = K.function(trained_model.inputs, [trained_model.layers[1].output])\n",
    "    first_conv_out = first_conv([input_data])\n",
    "    fig, axes = plt.subplots(num_filters / 4, 4,figsize=(8, 8))\n",
    "    for filter_num in range(num_filters):\n",
    "        fig_x, fig_y = divmod(filter_num, 4)\n",
    "        if fig_y == 0: fig_x -= 1\n",
    "        axes[fig_x][fig_y].imshow(first_conv_out[0][show_pic_idx][filter_num])\n",
    "        plt.setp(axes[fig_x][fig_y].get_xticklabels(), visible=False)\n",
    "        plt.setp(axes[fig_x][fig_y].get_yticklabels(), visible=False)\n",
    "    fig.tight_layout(pad=0)\n",
    "    plt.show()\n",
    "\n",
    "def ShowInputImage(data):\n",
    "    plot = plt.figure()\n",
    "    plot.set_size_inches(2,2)\n",
    "    plt.imshow(np.reshape(-data, (28,28)), cmap='Greys_r')\n",
    "    plt.title(\"Input\")\n",
    "    plt.axis('off')\n",
    "    plt.show()\n",
    "\n",
    "def ShowFinalOutput(input_data, samp):\n",
    "    \"\"\"Calculate final prediction.\"\"\"\n",
    "    from keras import backend as K\n",
    "    # Calculate final prediction.\n",
    "    last_layer = K.function(trained_model.inputs, [trained_model.layers[-1].output])\n",
    "    last_layer_out = last_layer([input_data])\n",
    "    print(\"Final prediction: \" + str(np.argmax(last_layer_out[0][samp])) )\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Show Random Test Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAMYAAADXCAYAAABBES5jAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nofBAiaFQeKDEUCg8+P9PAfeNOlKX0gAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1121213d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsAAAAMzCAYAAACoXqTsAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCYgJaB2JSnINjJdaFVhde2fP1Zqulb
Download .txt
gitextract_g80rflq0/

├── .gitignore
├── Digit_Recognizer/
│   ├── random_forrest.txt
│   ├── test.csv
│   └── train.csv
├── Keras_CNN_BitTiger.ipynb
├── Keras_FNN_BitTiger.ipynb
├── Quora_questions_similar/
│   ├── .gitignore
│   ├── CNN_Question_pair.py
│   ├── README.md
│   ├── lstm_embedding.py
│   └── similar_questions.py
├── README.md
├── TensorFlow_Examples/
│   ├── .gitignore
│   ├── LICENSE
│   ├── _config.yml
│   ├── assets/
│   │   └── css/
│   │       └── style.scss
│   ├── labs/
│   │   ├── 02_recommendation/
│   │   │   ├── Explicit_Feedback_Neural_Recommender_System.ipynb
│   │   │   ├── Implicit_Feedback_Recsys_with_the_triplet_loss.ipynb
│   │   │   └── solutions/
│   │   │       ├── classification.py
│   │   │       ├── compute_errors.py
│   │   │       ├── deep_explicit_feedback_recsys.py
│   │   │       ├── deep_implicit_feedback_recsys.py
│   │   │       ├── deep_triplet_parameter_count.py
│   │   │       ├── ml-1m.py
│   │   │       ├── similarity.py
│   │   │       └── triplet_parameter_count.py
│   │   ├── 03_LSTM_NLP/
│   │   │   └── NLP_word_vectors_classification.ipynb
│   │   ├── 03_conv_nets/
│   │   │   ├── ConvNets_with_TensorFlow_rendered.ipynb
│   │   │   ├── Pretrained_ConvNets_with_Keras.ipynb
│   │   │   ├── cifar10_input.py
│   │   │   ├── img_emb.h5
│   │   │   ├── process_images.py
│   │   │   └── solutions/
│   │   │       ├── average_as_conv.py
│   │   │       ├── dogs_vs_cats_extract_features.py
│   │   │       ├── dogs_vs_cats_worst_predictions.py
│   │   │       ├── edge_detection.py
│   │   │       ├── mnist_conv.py
│   │   │       ├── mnist_conv_dropout.py
│   │   │       ├── pooling.py
│   │   │       ├── predict_image.py
│   │   │       ├── representations.py
│   │   │       └── strides_padding.py
│   │   └── 04_seq2seq/
│   │       ├── Translation_of_Numeric_Phrases_with_Seq2Seq_rendered.ipynb
│   │       ├── data_download.ipynb
│   │       ├── french_numbers.py
│   │       ├── solutions/
│   │       │   ├── beam_search.py
│   │       │   ├── interpret_output.py
│   │       │   ├── make_input_output.py
│   │       │   └── simple_seq2seq.py
│   │       └── translate.py
│   └── slides/
│       ├── 01_intro_to_deep_learning/
│       │   ├── index.html
│       │   ├── slides.css
│       │   ├── webfont-ubuntu-400-300-100.css
│       │   └── webfont-ubuntu-mono-400-700-400italic.css
│       ├── 02_recommender_systems/
│       │   ├── images/
│       │   │   └── make_dropout_curves.py
│       │   ├── index.html
│       │   ├── slides.css
│       │   ├── webfont-ubuntu-400-300-100.css
│       │   └── webfont-ubuntu-mono-400-700-400italic.css
│       ├── 03_conv_nets/
│       │   ├── index.html
│       │   ├── slides.css
│       │   ├── webfont-ubuntu-400-300-100.css
│       │   └── webfont-ubuntu-mono-400-700-400italic.css
│       ├── 04_conv_nets_2/
│       │   ├── index.html
│       │   ├── slides.css
│       │   ├── webfont-ubuntu-400-300-100.css
│       │   └── webfont-ubuntu-mono-400-700-400italic.css
│       ├── 05_deep_nlp/
│       │   ├── index.html
│       │   ├── slides.css
│       │   ├── webfont-ubuntu-400-300-100.css
│       │   └── webfont-ubuntu-mono-400-700-400italic.css
│       └── 06_expressivity_optimization_generalization/
│           ├── images/
│           │   ├── make_mini_mlp_loss_figures.py
│           │   ├── make_relu_approx_figures.py
│           │   ├── make_relu_composition_figures.py
│           │   └── random_relu_surface.py
│           ├── index.html
│           ├── slides.css
│           ├── webfont-ubuntu-400-300-100.css
│           └── webfont-ubuntu-mono-400-700-400italic.css
├── Workshop/
│   └── sessions/
│       ├── .ipynb_checkpoints/
│       │   ├── 2-3_classify-checkpoint.ipynb
│       │   └── 4_deep-checkpoint.ipynb
│       ├── 13_lstm.ipynb
│       ├── 14_embeddings.ipynb
│       ├── 15_news20.ipynb
│       ├── 16_QA.ipynb
│       ├── 2-3_classify.ipynb
│       ├── 20news_download.sh
│       ├── 4_deep.ipynb
│       ├── 5_convnets.ipynb
│       ├── 6_visualization.ipynb
│       ├── 7_gan.ipynb
│       ├── 8_dream.ipynb
│       ├── 8_style.ipynb
│       ├── detection/
│       │   ├── README.md
│       │   ├── data/
│       │   │   ├── checkpoints/
│       │   │   │   └── README.md
│       │   │   ├── files/
│       │   │   │   ├── VOC2007.pkl
│       │   │   │   ├── gt_pascal.pkl
│       │   │   │   └── prior_boxes_ssd300.pkl
│       │   │   └── out/
│       │   │       └── README.md
│       │   └── ssd_keras/
│       │       ├── PASCAL_VOC/
│       │       │   ├── __init__.py
│       │       │   └── get_data_from_XML.py
│       │       ├── ssd.py
│       │       ├── ssd_layers.py
│       │       ├── ssd_training.py
│       │       ├── ssd_utils.py
│       │       ├── testing.ipynb
│       │       ├── testing_video.ipynb
│       │       └── training.ipynb
│       ├── dream.py
│       ├── glove_download.sh
│       ├── style.py
│       ├── tracking/
│       │   ├── README.md
│       │   ├── data/
│       │   │   ├── ADL-Rundle-6/
│       │   │   │   └── det.txt
│       │   │   ├── ADL-Rundle-8/
│       │   │   │   └── det.txt
│       │   │   ├── ETH-Bahnhof/
│       │   │   │   └── det.txt
│       │   │   ├── ETH-Pedcross2/
│       │   │   │   └── det.txt
│       │   │   ├── ETH-Sunnyday/
│       │   │   │   └── det.txt
│       │   │   ├── KITTI-13/
│       │   │   │   └── det.txt
│       │   │   ├── KITTI-17/
│       │   │   │   └── det.txt
│       │   │   ├── PETS09-S2L1/
│       │   │   │   └── det.txt
│       │   │   ├── TUD-Campus/
│       │   │   │   └── det.txt
│       │   │   ├── TUD-Stadtmitte/
│       │   │   │   └── det.txt
│       │   │   └── Venice-2/
│       │   │       └── det.txt
│       │   ├── sort.py
│       │   ├── tracking_kalman.ipynb
│       │   └── tracking_siamese.ipynb
│       └── utils.py
├── doc/
│   ├── AI_chip.md
│   ├── Basics_Machine_learning.md
│   ├── CNN.md
│   ├── Kaggle.md
│   ├── NLP.md
│   ├── RNN.md
│   ├── Recommendation.md
│   ├── SVM.md
│   ├── Search.md
│   ├── Tensorflow_Keras.md
│   ├── decision_tree.md
│   ├── industrial_design.md
│   ├── pytorch.md
│   └── system.md
├── human-activity-recognition-with-smartphones/
│   ├── .gitignore
│   └── ADL_sensor_data.ipynb
├── keras_FNN_BitTiger.py
├── sklearn_test.py
├── tensor_flow_test.py
└── xgboost/
    ├── .gitignore
    ├── tree_xgboost.ipynb
    └── xgboost_feature.ipynb
Download .txt
SYMBOL INDEX (130 symbols across 27 files)

FILE: Quora_questions_similar/CNN_Question_pair.py
  function text_to_wordlist (line 84) | def text_to_wordlist(text, remove_stop_words=True, stem_words=False):
  function process_questions (line 168) | def process_questions(question_list, questions, question_list_name, data...

FILE: Quora_questions_similar/lstm_embedding.py
  function text_to_wordlist (line 82) | def text_to_wordlist(text, remove_stopwords=False, stem_words=False):

FILE: TensorFlow_Examples/labs/02_recommendation/solutions/deep_implicit_feedback_recsys.py
  function make_interaction_mlp (line 6) | def make_interaction_mlp(input_dim, n_hidden=1, hidden_size=64,
  function build_models (line 27) | def build_models(n_users, n_items, user_dim=32, item_dim=64,

FILE: TensorFlow_Examples/labs/02_recommendation/solutions/similarity.py
  function cosine (line 3) | def cosine(x, y):
  function cosine_similarities (line 9) | def cosine_similarities(x):
  function euclidean_distances (line 15) | def euclidean_distances(x):
  function most_similar (line 19) | def most_similar(idx, top_n=10, mode='euclidean'):

FILE: TensorFlow_Examples/labs/03_conv_nets/cifar10_input.py
  function read_cifar10 (line 38) | def read_cifar10(filename_queue):
  function _generate_image_and_label_batch (line 101) | def _generate_image_and_label_batch(image, label, min_queue_examples,
  function distorted_inputs (line 140) | def distorted_inputs(data_dir, batch_size):
  function inputs (line 203) | def inputs(eval_data, data_dir, batch_size, max_files=5, raw_images=False):

FILE: TensorFlow_Examples/labs/03_conv_nets/solutions/average_as_conv.py
  function conv (line 4) | def conv(x, k):

FILE: TensorFlow_Examples/labs/03_conv_nets/solutions/edge_detection.py
  function conv (line 4) | def conv(x, k):

FILE: TensorFlow_Examples/labs/03_conv_nets/solutions/strides_padding.py
  function conv (line 4) | def conv(x, k):
  function conv_valid (line 9) | def conv_valid(x, k):

FILE: TensorFlow_Examples/labs/04_seq2seq/french_numbers.py
  function to_french_phrase (line 74) | def to_french_phrase(number, is_units=True):
  function generate_translations (line 139) | def generate_translations(low=1, high=int(1e6) - 1, exhaustive=int(1e4),
  function prepare_datasets (line 171) | def prepare_datasets(data_dir, low=0, high=int(1e6) - 1,
  function tokenize (line 224) | def tokenize(sentence, word_level=True):
  function save_sentences (line 231) | def save_sentences(sentences, vocab, token_filename, text_filename,
  function save_vocabulary (line 247) | def save_vocabulary(rev_vocab, filename, encoding='utf-8'):
  function load_vocabulary (line 254) | def load_vocabulary(filename, encoding='utf-8'):
  function build_vocabulary (line 264) | def build_vocabulary(sentences, word_level=True):
  function sentence_to_token_ids (line 295) | def sentence_to_token_ids(sentence, vocabulary, word_level=True):
  function token_ids_to_sentence (line 315) | def token_ids_to_sentence(token_ids, rev_vocabulary, word_level=True):

FILE: TensorFlow_Examples/labs/04_seq2seq/solutions/beam_search.py
  function beam_translate (line 1) | def beam_translate(model, source_sequence, shared_vocab, rev_shared_vocab,

FILE: TensorFlow_Examples/labs/04_seq2seq/solutions/make_input_output.py
  function make_input_output (line 1) | def make_input_output(source_tokens, target_tokens, reverse_source=True):

FILE: TensorFlow_Examples/labs/04_seq2seq/translate.py
  function read_data (line 85) | def read_data(source_path, target_path, max_size=None):
  function create_model (line 126) | def create_model(session, source_vocab_size, target_vocab_size,
  function train (line 152) | def train():
  function decode (line 247) | def decode():
  function self_test (line 299) | def self_test():
  function main (line 321) | def main(_):

FILE: TensorFlow_Examples/slides/02_recommender_systems/images/make_dropout_curves.py
  function make_curves (line 18) | def make_curves(random_state=42):

FILE: TensorFlow_Examples/slides/06_expressivity_optimization_generalization/images/make_mini_mlp_loss_figures.py
  function relu (line 18) | def relu(x):
  function mini_mlp (line 22) | def mini_mlp(x, w1, w2):
  function dloss_mini_mlp (line 26) | def dloss_mini_mlp(x, y, w1, w2):

FILE: TensorFlow_Examples/slides/06_expressivity_optimization_generalization/images/make_relu_approx_figures.py
  function relu (line 9) | def relu(x):
  function rect (line 13) | def rect(x, a=1, b=2, h=2, eps=1e-7):

FILE: TensorFlow_Examples/slides/06_expressivity_optimization_generalization/images/make_relu_composition_figures.py
  function relu (line 9) | def relu(x):
  function tri (line 13) | def tri(x):

FILE: TensorFlow_Examples/slides/06_expressivity_optimization_generalization/images/random_relu_surface.py
  function weight_init (line 11) | def weight_init(n_in, n_out):
  function bias_init (line 15) | def bias_init(n):

FILE: Workshop/sessions/detection/ssd_keras/PASCAL_VOC/get_data_from_XML.py
  class XML_preprocessor (line 5) | class XML_preprocessor(object):
    method __init__ (line 7) | def __init__(self, data_path):
    method _preprocess_XML (line 13) | def _preprocess_XML(self):
    method _to_one_hot (line 40) | def _to_one_hot(self,name):

FILE: Workshop/sessions/detection/ssd_keras/ssd.py
  function SSD300 (line 23) | def SSD300(input_shape, num_classes=21,weights=None):

FILE: Workshop/sessions/detection/ssd_keras/ssd_layers.py
  class Normalize (line 10) | class Normalize(Layer):
    method __init__ (line 31) | def __init__(self, scale, **kwargs):
    method build (line 39) | def build(self, input_shape):
    method call (line 46) | def call(self, x, mask=None):
  class PriorBox (line 52) | class PriorBox(Layer):
    method __init__ (line 82) | def __init__(self, img_size, min_size, max_size=None, aspect_ratios=None,
    method get_output_shape_for (line 111) | def get_output_shape_for(self, input_shape):
    method call (line 118) | def call(self, x, mask=None):

FILE: Workshop/sessions/detection/ssd_keras/ssd_training.py
  class MultiboxLoss (line 6) | class MultiboxLoss(object):
    method __init__ (line 23) | def __init__(self, num_classes, alpha=1.0, neg_pos_ratio=3.0,
    method _l1_smooth_loss (line 33) | def _l1_smooth_loss(self, y_true, y_pred):
    method _softmax_loss (line 53) | def _softmax_loss(self, y_true, y_pred):
    method compute_loss (line 70) | def compute_loss(self, y_true, y_pred):

FILE: Workshop/sessions/detection/ssd_keras/ssd_utils.py
  class BBoxUtility (line 13) | class BBoxUtility(object):
    method __init__ (line 28) | def __init__(self, num_classes, priors=None, overlap_threshold=0.5,
    method nms_thresh (line 44) | def nms_thresh(self):
    method nms_thresh (line 48) | def nms_thresh(self, value):
    method top_k (line 55) | def top_k(self):
    method top_k (line 59) | def top_k(self, value):
    method iou (line 65) | def iou(self, box):
    method encode_box (line 90) | def encode_box(self, box, return_iou=True):
    method assign_boxes (line 124) | def assign_boxes(self, boxes):
    method decode_boxes (line 159) | def decode_boxes(self, mbox_loc, mbox_priorbox, variances):
    method detection_out (line 193) | def detection_out(self, predictions, background_label_id=0, keep_top_k...
  class Generator (line 243) | class Generator(object):
    method __init__ (line 244) | def __init__(self, gt, bbox_util,
    method grayscale (line 282) | def grayscale(self, rgb):
    method saturation (line 285) | def saturation(self, rgb):
    method brightness (line 292) | def brightness(self, rgb):
    method contrast (line 298) | def contrast(self, rgb):
    method lighting (line 305) | def lighting(self, img):
    method horizontal_flip (line 313) | def horizontal_flip(self, img, y):
    method vertical_flip (line 319) | def vertical_flip(self, img, y):
    method random_sized_crop (line 325) | def random_sized_crop(self, img, targets):
    method generate (line 375) | def generate(self, train=True):

FILE: Workshop/sessions/dream.py
  function preprocess_image (line 9) | def preprocess_image(image_path,img_height,img_width):
  function deprocess_image (line 17) | def deprocess_image(x,img_height,img_width):
  function continuity_loss (line 32) | def continuity_loss(x,img_height,img_width):
  function eval_loss_and_grads (line 41) | def eval_loss_and_grads(x,img_size,f_outputs):
  class Evaluator (line 59) | class Evaluator(object):
    method __init__ (line 61) | def __init__(self,img_size,f_outputs):
    method loss (line 67) | def loss(self, x):
    method grads (line 75) | def grads(self, x):

FILE: Workshop/sessions/style.py
  function preprocess_image (line 14) | def preprocess_image(image_path,img_nrows,img_ncols):
  function deprocess_image (line 22) | def deprocess_image(x,img_nrows,img_ncols):
  function gram_matrix (line 34) | def gram_matrix(x):
  function style_loss (line 50) | def style_loss(style, combination,img_nrows,img_ncols):
  function content_loss (line 62) | def content_loss(base, combination):
  function total_variation_loss (line 68) | def total_variation_loss(x,img_nrows,img_ncols):
  function eval_loss_and_grads (line 76) | def eval_loss_and_grads(x,img_nrows,img_ncols,f_outputs):
  class Evaluator (line 96) | class Evaluator(object):
    method __init__ (line 98) | def __init__(self,img_nrows,img_ncols,f_outputs):
    method loss (line 105) | def loss(self, x):
    method grads (line 114) | def grads(self, x):

FILE: Workshop/sessions/tracking/sort.py
  function iou (line 31) | def iou(bb_test,bb_gt):
  function convert_bbox_to_z (line 46) | def convert_bbox_to_z(bbox):
  function convert_x_to_bbox (line 60) | def convert_x_to_bbox(x,score=None):
  class KalmanBoxTracker (line 73) | class KalmanBoxTracker(object):
    method __init__ (line 79) | def __init__(self,bbox):
    method update (line 104) | def update(self,bbox):
    method predict (line 114) | def predict(self):
    method get_state (line 128) | def get_state(self):
  function associate_detections_to_trackers (line 134) | def associate_detections_to_trackers(detections,trackers,iou_threshold =...
  class Sort (line 175) | class Sort(object):
    method __init__ (line 176) | def __init__(self,max_age=1,min_hits=3):
    method update (line 185) | def update(self,dets):

FILE: Workshop/sessions/utils.py
  function plot_confusion_matrix (line 6) | def plot_confusion_matrix(cm, classes,
  function plot_samples (line 40) | def plot_samples(X_train,N=5):
  function plot_curves (line 62) | def plot_curves(history,nb_epoch):
  function deprocess_image (line 86) | def deprocess_image(x):
  function normalize (line 103) | def normalize(x):

FILE: keras_FNN_BitTiger.py
  function PreprocessDataset (line 56) | def PreprocessDataset():
  function DefineModel (line 157) | def DefineModel():
  function TrainModel (line 289) | def TrainModel(data=None, epochs=20):
  function PlotHistory (line 327) | def PlotHistory(train_value, test_value, value_is_loss_or_acc):
  function drawWeightHistogram (line 352) | def drawWeightHistogram(x):
  function TestModel (line 374) | def TestModel(model=None, data=None):
  function ShowInputImage (line 397) | def ShowInputImage(data):
  function ShowHiddenLayerOutput (line 406) | def ShowHiddenLayerOutput(input_data, target_layer_num):
  function ShowFinalOutput (line 420) | def ShowFinalOutput(input_data):
Copy disabled (too large) Download .json
Condensed preview — 149 files, each showing path, character count, and a content snippet. Download the .json file for the full structured content (17,217K chars).
[
  {
    "path": ".gitignore",
    "chars": 133,
    "preview": "*.docx\n*.pdf\n/book\n.idea/*\n.ipynb_checkpoints/*\nxgboost/.ipynb_checkpoints/*\n\n/TensorFlow_Examples/labs/03_conv_nets/ima"
  },
  {
    "path": "Digit_Recognizer/random_forrest.txt",
    "chars": 758,
    "preview": "from sklearn.ensemble import RandomForestClassifier\r\nfrom numpy import genfromtxt, savetxt\r\n\r\ndef main():\r\n    #create t"
  },
  {
    "path": "Keras_CNN_BitTiger.ipynb",
    "chars": 1423678,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {},\n   \"source\": [\n    \"# Codelab for Convolutional Neural "
  },
  {
    "path": "Keras_FNN_BitTiger.ipynb",
    "chars": 28247,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {},\n   \"source\": [\n    \"# Codelab for Feedforward Neural Ne"
  },
  {
    "path": "Quora_questions_similar/.gitignore",
    "chars": 70,
    "preview": "# Created by .ignore support plugin (hsz.mobi)\n*.npy\n*.json\n*.h5\n*.csv"
  },
  {
    "path": "Quora_questions_similar/CNN_Question_pair.py",
    "chars": 15880,
    "preview": "# coding: utf-8\n\n# # Predicting Duplicate Questions\n\n# In[5]:\n\nimport pandas as pd\nimport numpy as np\nimport nltk\nfrom n"
  },
  {
    "path": "Quora_questions_similar/README.md",
    "chars": 558,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "Quora_questions_similar/lstm_embedding.py",
    "chars": 10140,
    "preview": "'''\nSingle model may achieve LB scores at around 0.29+ ~ 0.30+\nAverage ensembles can easily get 0.28+ or less\nDon't need"
  },
  {
    "path": "Quora_questions_similar/similar_questions.py",
    "chars": 8149,
    "preview": "from __future__ import print_function\nimport numpy as np\nimport csv, datetime, time, json\nfrom zipfile import ZipFile\nfr"
  },
  {
    "path": "README.md",
    "chars": 7065,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "TensorFlow_Examples/.gitignore",
    "chars": 635,
    "preview": "# random crap\n*~\n*.swp\n\\#*\\#\n.\\#*\n\n# python crap\n*.pyc\n__pycache__/\n\nevaluations/*.pdf\n\n# Dataset files and folders\nml-1"
  },
  {
    "path": "TensorFlow_Examples/LICENSE",
    "chars": 1057,
    "preview": "MIT License\n\nCopyright (c) 2017 \n\nPermission is hereby granted, free of charge, to any person obtaining a copy\nof this s"
  },
  {
    "path": "TensorFlow_Examples/_config.yml",
    "chars": 25,
    "preview": "theme: jekyll-theme-slate"
  },
  {
    "path": "TensorFlow_Examples/assets/css/style.scss",
    "chars": 72,
    "preview": "---\n---\n\n@import \"{{ site.theme }}\";\n\n#header_wrap {\n  display: none;\n}\n"
  },
  {
    "path": "TensorFlow_Examples/labs/02_recommendation/Explicit_Feedback_Neural_Recommender_System.ipynb",
    "chars": 172908,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "TensorFlow_Examples/labs/02_recommendation/Implicit_Feedback_Recsys_with_the_triplet_loss.ipynb",
    "chars": 57073,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "TensorFlow_Examples/labs/02_recommendation/solutions/classification.py",
    "chars": 2103,
    "preview": "# For each sample we input the integer identifiers\n# of a single user and a single item\nuser_id_input = Input(shape=[1],"
  },
  {
    "path": "TensorFlow_Examples/labs/02_recommendation/solutions/compute_errors.py",
    "chars": 600,
    "preview": "squared_differences = np.square(initial_train_preds - rating_train.values)\nabsolute_differences = np.abs(initial_train_p"
  },
  {
    "path": "TensorFlow_Examples/labs/02_recommendation/solutions/deep_explicit_feedback_recsys.py",
    "chars": 1485,
    "preview": "# For each sample we input the integer identifiers\n# of a single user and a single item\nuser_id_input = Input(shape=[1],"
  },
  {
    "path": "TensorFlow_Examples/labs/02_recommendation/solutions/deep_implicit_feedback_recsys.py",
    "chars": 4730,
    "preview": "from keras.models import Model, Sequential\nfrom keras.layers import Embedding, Flatten, Input, Dense, Dropout, merge\nfro"
  },
  {
    "path": "TensorFlow_Examples/labs/02_recommendation/solutions/deep_triplet_parameter_count.py",
    "chars": 1080,
    "preview": "# Analysis:\n#\n# Both models have again exactly the same number of parameters,\n# namely the parameters of the 2 embedding"
  },
  {
    "path": "TensorFlow_Examples/labs/02_recommendation/solutions/ml-1m.py",
    "chars": 2942,
    "preview": "import matplotlib.pyplot as plt\n\nfrom pathlib import Path\nfrom zipfile import ZipFile\nfrom urllib.request import urlretr"
  },
  {
    "path": "TensorFlow_Examples/labs/02_recommendation/solutions/similarity.py",
    "chars": 2171,
    "preview": "EPSILON = 1e-07\n\ndef cosine(x, y):\n    dot_pdt = np.dot(x, y.T)\n    norms = np.linalg.norm(x) * np.linalg.norm(y)\n    re"
  },
  {
    "path": "TensorFlow_Examples/labs/02_recommendation/solutions/triplet_parameter_count.py",
    "chars": 695,
    "preview": "# Analysis:\n#\n# Both models have exactly the same number of parameters,\n# namely the parameters of the 2 embeddings:\n#\n#"
  },
  {
    "path": "TensorFlow_Examples/labs/03_LSTM_NLP/NLP_word_vectors_classification.ipynb",
    "chars": 29446,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "TensorFlow_Examples/labs/03_conv_nets/ConvNets_with_TensorFlow_rendered.ipynb",
    "chars": 791971,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": 28,\n   \"metadata\": {\n    \"collapsed\": false,\n    \"deleta"
  },
  {
    "path": "TensorFlow_Examples/labs/03_conv_nets/cifar10_input.py",
    "chars": 9849,
    "preview": "# Copyright 2015 The TensorFlow Authors. All Rights Reserved.\n#\n# Licensed under the Apache License, Version 2.0 (the \"L"
  },
  {
    "path": "TensorFlow_Examples/labs/03_conv_nets/process_images.py",
    "chars": 1393,
    "preview": "from keras.applications.resnet50 import ResNet50\nfrom keras.models import Model\nfrom keras.preprocessing import image\nfr"
  },
  {
    "path": "TensorFlow_Examples/labs/03_conv_nets/solutions/average_as_conv.py",
    "chars": 1057,
    "preview": "image = tf.placeholder(tf.float32, [None, None, None, 3])\nkernel = tf.placeholder(tf.float32, [3, 3, 3, 3])\n\ndef conv(x,"
  },
  {
    "path": "TensorFlow_Examples/labs/03_conv_nets/solutions/dogs_vs_cats_extract_features.py",
    "chars": 503,
    "preview": "from time import time\n\nfeatures = []\nlabels = []\n\nt0 = time()\ncount = 0\nfor X, y in train_flow:\n    labels.append(y)\n   "
  },
  {
    "path": "TensorFlow_Examples/labs/03_conv_nets/solutions/dogs_vs_cats_worst_predictions.py",
    "chars": 1606,
    "preview": "from IPython.display import Image, display\n\npredicted_batches = []\nlabel_batches = []\nn_batches = val_flow.nb_sample // "
  },
  {
    "path": "TensorFlow_Examples/labs/03_conv_nets/solutions/edge_detection.py",
    "chars": 961,
    "preview": "image = tf.placeholder(tf.float32, [None, None, None, 1])\nkernel = tf.placeholder(tf.float32, [3, 3])\n\ndef conv(x, k):\n "
  },
  {
    "path": "TensorFlow_Examples/labs/03_conv_nets/solutions/mnist_conv.py",
    "chars": 616,
    "preview": "W_conv1 = weight_variable([5, 5, 1, 32])\nb_conv1 = bias_variable([32])\n\nh_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + "
  },
  {
    "path": "TensorFlow_Examples/labs/03_conv_nets/solutions/mnist_conv_dropout.py",
    "chars": 1867,
    "preview": "# Model definition\nW_conv1 = weight_variable([5, 5, 1, 32])\nb_conv1 = bias_variable([32])\n\nh_conv1 = tf.nn.relu(conv2d(x"
  },
  {
    "path": "TensorFlow_Examples/labs/03_conv_nets/solutions/pooling.py",
    "chars": 563,
    "preview": "image = tf.placeholder(tf.float32, [None, None, None, 3])\noutput_image = tf.nn.max_pool(image, ksize=[1, 2, 2, 1],\n     "
  },
  {
    "path": "TensorFlow_Examples/labs/03_conv_nets/solutions/predict_image.py",
    "chars": 542,
    "preview": "\nfrom scipy.misc import imread, imresize\nfrom keras.applications.imagenet_utils import preprocess_input\nfrom keras.appli"
  },
  {
    "path": "TensorFlow_Examples/labs/03_conv_nets/solutions/representations.py",
    "chars": 368,
    "preview": "# Proportion of zeros in a representation\nprint(\"proportion of zeros\", np.mean(out_tensors[0]==0.0))\n\n# For all represen"
  },
  {
    "path": "TensorFlow_Examples/labs/03_conv_nets/solutions/strides_padding.py",
    "chars": 1339,
    "preview": "image = tf.placeholder(tf.float32, [None, None, None, 3])\nkernel = tf.placeholder(tf.float32, [3, 3, 3])\n\ndef conv(x, k)"
  },
  {
    "path": "TensorFlow_Examples/labs/04_seq2seq/Translation_of_Numeric_Phrases_with_Seq2Seq_rendered.ipynb",
    "chars": 91071,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "TensorFlow_Examples/labs/04_seq2seq/data_download.ipynb",
    "chars": 1380,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": null,\n   \"metadata\": {\n    \"collapsed\": false,\n    \"dele"
  },
  {
    "path": "TensorFlow_Examples/labs/04_seq2seq/french_numbers.py",
    "chars": 11193,
    "preview": "from __future__ import print_function\nimport os.path as op\nfrom random import Random\nfrom math import log10\nfrom sklearn"
  },
  {
    "path": "TensorFlow_Examples/labs/04_seq2seq/solutions/beam_search.py",
    "chars": 3127,
    "preview": "def beam_translate(model, source_sequence, shared_vocab, rev_shared_vocab,\n                   word_level_source=True, wo"
  },
  {
    "path": "TensorFlow_Examples/labs/04_seq2seq/solutions/interpret_output.py",
    "chars": 970,
    "preview": "prediction = simple_seq2seq.predict(first_test_sequence)\nprint(\"prediction shape:\", prediction.shape)\n\n# Let's use `argm"
  },
  {
    "path": "TensorFlow_Examples/labs/04_seq2seq/solutions/make_input_output.py",
    "chars": 278,
    "preview": "def make_input_output(source_tokens, target_tokens, reverse_source=True):\n    if reverse_source:\n        source_tokens ="
  },
  {
    "path": "TensorFlow_Examples/labs/04_seq2seq/solutions/simple_seq2seq.py",
    "chars": 612,
    "preview": "from keras.models import Sequential\nfrom keras.layers import Embedding, Dropout, GRU, Dense\n\nvocab_size = len(shared_voc"
  },
  {
    "path": "TensorFlow_Examples/labs/04_seq2seq/translate.py",
    "chars": 14957,
    "preview": "# copyright 2017 Olivier Grisel\n# Copyright 2015 The TensorFlow Authors. All Rights Reserved.\n#\n# Original code from:\n#\n"
  },
  {
    "path": "TensorFlow_Examples/slides/01_intro_to_deep_learning/index.html",
    "chars": 18357,
    "preview": "<!DOCTYPE html>\n<html>\n  <head>\n    <title>Deep Learning Lectures</title>\n    <meta charset=\"utf-8\">\n    <style>\n     .l"
  },
  {
    "path": "TensorFlow_Examples/slides/01_intro_to_deep_learning/slides.css",
    "chars": 1674,
    "preview": "@import url(webfont-ubuntu-400-300-100.css);\n@import url(webfont-ubuntu-mono-400-700-400italic.css);\n\nbody {\n  font-fami"
  },
  {
    "path": "TensorFlow_Examples/slides/01_intro_to_deep_learning/webfont-ubuntu-400-300-100.css",
    "chars": 3779,
    "preview": "/* cyrillic-ext */\n@font-face {\n  font-family: 'Ubuntu';\n  font-style: normal;\n  font-weight: 300;\n  src: local('Ubuntu "
  },
  {
    "path": "TensorFlow_Examples/slides/01_intro_to_deep_learning/webfont-ubuntu-mono-400-700-400italic.css",
    "chars": 6261,
    "preview": "/* cyrillic-ext */\n@font-face {\n  font-family: 'Ubuntu Mono';\n  font-style: normal;\n  font-weight: 400;\n  src: local('Ub"
  },
  {
    "path": "TensorFlow_Examples/slides/02_recommender_systems/images/make_dropout_curves.py",
    "chars": 3803,
    "preview": "# display figures in the notebook\nimport tensorflow as tf\nimport matplotlib.pyplot as plt\nimport numpy as np\nfrom sklear"
  },
  {
    "path": "TensorFlow_Examples/slides/02_recommender_systems/index.html",
    "chars": 13155,
    "preview": "<!DOCTYPE html>\n<html>\n  <head>\n    <title>Deep Learning Lectures</title>\n    <meta charset=\"utf-8\">\n    <style>\n     .l"
  },
  {
    "path": "TensorFlow_Examples/slides/02_recommender_systems/slides.css",
    "chars": 1674,
    "preview": "@import url(webfont-ubuntu-400-300-100.css);\n@import url(webfont-ubuntu-mono-400-700-400italic.css);\n\nbody {\n  font-fami"
  },
  {
    "path": "TensorFlow_Examples/slides/02_recommender_systems/webfont-ubuntu-400-300-100.css",
    "chars": 3779,
    "preview": "/* cyrillic-ext */\n@font-face {\n  font-family: 'Ubuntu';\n  font-style: normal;\n  font-weight: 300;\n  src: local('Ubuntu "
  },
  {
    "path": "TensorFlow_Examples/slides/02_recommender_systems/webfont-ubuntu-mono-400-700-400italic.css",
    "chars": 6261,
    "preview": "/* cyrillic-ext */\n@font-face {\n  font-family: 'Ubuntu Mono';\n  font-style: normal;\n  font-weight: 400;\n  src: local('Ub"
  },
  {
    "path": "TensorFlow_Examples/slides/03_conv_nets/index.html",
    "chars": 16337,
    "preview": "<!DOCTYPE html>\n<html>\n  <head>\n    <title>Deep Learning Lectures</title>\n    <meta charset=\"utf-8\">\n    <style>\n     .l"
  },
  {
    "path": "TensorFlow_Examples/slides/03_conv_nets/slides.css",
    "chars": 1674,
    "preview": "@import url(webfont-ubuntu-400-300-100.css);\n@import url(webfont-ubuntu-mono-400-700-400italic.css);\n\nbody {\n  font-fami"
  },
  {
    "path": "TensorFlow_Examples/slides/03_conv_nets/webfont-ubuntu-400-300-100.css",
    "chars": 3779,
    "preview": "/* cyrillic-ext */\n@font-face {\n  font-family: 'Ubuntu';\n  font-style: normal;\n  font-weight: 300;\n  src: local('Ubuntu "
  },
  {
    "path": "TensorFlow_Examples/slides/03_conv_nets/webfont-ubuntu-mono-400-700-400italic.css",
    "chars": 6261,
    "preview": "/* cyrillic-ext */\n@font-face {\n  font-family: 'Ubuntu Mono';\n  font-style: normal;\n  font-weight: 400;\n  src: local('Ub"
  },
  {
    "path": "TensorFlow_Examples/slides/04_conv_nets_2/index.html",
    "chars": 13628,
    "preview": "<!DOCTYPE html>\n<html>\n  <head>\n    <title>Deep Learning Lectures</title>\n    <meta charset=\"utf-8\">\n    <style>\n     .l"
  },
  {
    "path": "TensorFlow_Examples/slides/04_conv_nets_2/slides.css",
    "chars": 1674,
    "preview": "@import url(webfont-ubuntu-400-300-100.css);\n@import url(webfont-ubuntu-mono-400-700-400italic.css);\n\nbody {\n  font-fami"
  },
  {
    "path": "TensorFlow_Examples/slides/04_conv_nets_2/webfont-ubuntu-400-300-100.css",
    "chars": 3779,
    "preview": "/* cyrillic-ext */\n@font-face {\n  font-family: 'Ubuntu';\n  font-style: normal;\n  font-weight: 300;\n  src: local('Ubuntu "
  },
  {
    "path": "TensorFlow_Examples/slides/04_conv_nets_2/webfont-ubuntu-mono-400-700-400italic.css",
    "chars": 6261,
    "preview": "/* cyrillic-ext */\n@font-face {\n  font-family: 'Ubuntu Mono';\n  font-style: normal;\n  font-weight: 400;\n  src: local('Ub"
  },
  {
    "path": "TensorFlow_Examples/slides/05_deep_nlp/index.html",
    "chars": 21123,
    "preview": "<!DOCTYPE html>\n<html>\n  <head>\n    <title>Deep Learning Lectures</title>\n    <meta charset=\"utf-8\">\n    <style>\n     .l"
  },
  {
    "path": "TensorFlow_Examples/slides/05_deep_nlp/slides.css",
    "chars": 1725,
    "preview": "@import url(webfont-ubuntu-400-300-100.css);\n@import url(webfont-ubuntu-mono-400-700-400italic.css);\n\nbody {\n  font-fami"
  },
  {
    "path": "TensorFlow_Examples/slides/05_deep_nlp/webfont-ubuntu-400-300-100.css",
    "chars": 3779,
    "preview": "/* cyrillic-ext */\n@font-face {\n  font-family: 'Ubuntu';\n  font-style: normal;\n  font-weight: 300;\n  src: local('Ubuntu "
  },
  {
    "path": "TensorFlow_Examples/slides/05_deep_nlp/webfont-ubuntu-mono-400-700-400italic.css",
    "chars": 6261,
    "preview": "/* cyrillic-ext */\n@font-face {\n  font-family: 'Ubuntu Mono';\n  font-style: normal;\n  font-weight: 400;\n  src: local('Ub"
  },
  {
    "path": "TensorFlow_Examples/slides/06_expressivity_optimization_generalization/images/make_mini_mlp_loss_figures.py",
    "chars": 3601,
    "preview": "import numpy as np\nimport matplotlib.pyplot as plt\nimport os\nimport shutil\n\n# Generate the figures in the same folder\nos"
  },
  {
    "path": "TensorFlow_Examples/slides/06_expressivity_optimization_generalization/images/make_relu_approx_figures.py",
    "chars": 655,
    "preview": "import matplotlib.pyplot as plt\nimport numpy as np\nimport os\n\n# Generate the figures in the same folder\nos.chdir(os.path"
  },
  {
    "path": "TensorFlow_Examples/slides/06_expressivity_optimization_generalization/images/make_relu_composition_figures.py",
    "chars": 698,
    "preview": "import matplotlib.pyplot as plt\nimport numpy as np\nimport os\n\n# Generate the figures in the same folder\nos.chdir(os.path"
  },
  {
    "path": "TensorFlow_Examples/slides/06_expressivity_optimization_generalization/images/random_relu_surface.py",
    "chars": 1277,
    "preview": "import matplotlib.pyplot as plt\nimport numpy as np\nfrom keras.layers import Dense\nfrom keras.models import Sequential\n\nn"
  },
  {
    "path": "TensorFlow_Examples/slides/06_expressivity_optimization_generalization/index.html",
    "chars": 21431,
    "preview": "<!DOCTYPE html>\n<html>\n  <head>\n    <title>Deep Learning Lectures</title>\n    <meta charset=\"utf-8\">\n    <style>\n     .l"
  },
  {
    "path": "TensorFlow_Examples/slides/06_expressivity_optimization_generalization/slides.css",
    "chars": 1674,
    "preview": "@import url(webfont-ubuntu-400-300-100.css);\n@import url(webfont-ubuntu-mono-400-700-400italic.css);\n\nbody {\n  font-fami"
  },
  {
    "path": "TensorFlow_Examples/slides/06_expressivity_optimization_generalization/webfont-ubuntu-400-300-100.css",
    "chars": 3779,
    "preview": "/* cyrillic-ext */\n@font-face {\n  font-family: 'Ubuntu';\n  font-style: normal;\n  font-weight: 300;\n  src: local('Ubuntu "
  },
  {
    "path": "TensorFlow_Examples/slides/06_expressivity_optimization_generalization/webfont-ubuntu-mono-400-700-400italic.css",
    "chars": 6261,
    "preview": "/* cyrillic-ext */\n@font-face {\n  font-family: 'Ubuntu Mono';\n  font-style: normal;\n  font-weight: 400;\n  src: local('Ub"
  },
  {
    "path": "Workshop/sessions/.ipynb_checkpoints/2-3_classify-checkpoint.ipynb",
    "chars": 136896,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/.ipynb_checkpoints/4_deep-checkpoint.ipynb",
    "chars": 71902,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/13_lstm.ipynb",
    "chars": 31752,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/14_embeddings.ipynb",
    "chars": 6864,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/15_news20.ipynb",
    "chars": 9213,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/16_QA.ipynb",
    "chars": 13228,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/2-3_classify.ipynb",
    "chars": 134542,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/20news_download.sh",
    "chars": 161,
    "preview": "#!/bin/bash\nmkdir -p ../../data/20news\npushd ../../data/20news\nwget http://qwone.com/~jason/20Newsgroups/20news-bydate.t"
  },
  {
    "path": "Workshop/sessions/4_deep.ipynb",
    "chars": 156702,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/5_convnets.ipynb",
    "chars": 16169,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/6_visualization.ipynb",
    "chars": 1083297,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/7_gan.ipynb",
    "chars": 679438,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/8_dream.ipynb",
    "chars": 2067541,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/8_style.ipynb",
    "chars": 2894468,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/detection/README.md",
    "chars": 1837,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "Workshop/sessions/detection/data/checkpoints/README.md",
    "chars": 0,
    "preview": ""
  },
  {
    "path": "Workshop/sessions/detection/data/out/README.md",
    "chars": 0,
    "preview": ""
  },
  {
    "path": "Workshop/sessions/detection/ssd_keras/PASCAL_VOC/__init__.py",
    "chars": 0,
    "preview": ""
  },
  {
    "path": "Workshop/sessions/detection/ssd_keras/PASCAL_VOC/get_data_from_XML.py",
    "chars": 3279,
    "preview": "import numpy as np\nimport os\nfrom xml.etree import ElementTree\n\nclass XML_preprocessor(object):\n\n    def __init__(self, "
  },
  {
    "path": "Workshop/sessions/detection/ssd_keras/ssd.py",
    "chars": 12891,
    "preview": "\"\"\"Keras implementation of SSD.\"\"\"\n\nimport keras.backend as K\nfrom keras.layers import Activation\nfrom keras.layers impo"
  },
  {
    "path": "Workshop/sessions/detection/ssd_keras/ssd_layers.py",
    "chars": 6719,
    "preview": "\"\"\"Some special pupropse layers for SSD.\"\"\"\n\nimport keras.backend as K\nfrom keras.engine.topology import InputSpec\nfrom "
  },
  {
    "path": "Workshop/sessions/detection/ssd_keras/ssd_training.py",
    "chars": 5819,
    "preview": "\"\"\"SSD training utils.\"\"\"\n\nimport tensorflow as tf\n\n\nclass MultiboxLoss(object):\n    \"\"\"Multibox loss with some helper f"
  },
  {
    "path": "Workshop/sessions/detection/ssd_keras/ssd_utils.py",
    "chars": 17290,
    "preview": "\"\"\"Some utils for SSD.\"\"\"\n\nimport numpy as np\nimport tensorflow as tf\nfrom random import shuffle\nfrom scipy.misc import "
  },
  {
    "path": "Workshop/sessions/detection/ssd_keras/testing.ipynb",
    "chars": 1511143,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/detection/ssd_keras/testing_video.ipynb",
    "chars": 1678064,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/detection/ssd_keras/training.ipynb",
    "chars": 16791,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/dream.py",
    "chars": 2694,
    "preview": "from keras.preprocessing.image import load_img, img_to_array\r\nimport numpy as np\r\nfrom scipy.optimize import fmin_l_bfgs"
  },
  {
    "path": "Workshop/sessions/glove_download.sh",
    "chars": 156,
    "preview": "#!/bin/bash\nmkdir -p ../../data/glove\npushd ../../data/glove\n#wget http://nlp.stanford.edu/data/glove.6B.zip\n#tar -xvf g"
  },
  {
    "path": "Workshop/sessions/style.py",
    "chars": 4130,
    "preview": "'''Neural style transfer with Keras.\r\n# References\r\n    - [A Neural Algorithm of Artistic Style](http://arxiv.org/abs/15"
  },
  {
    "path": "Workshop/sessions/tracking/README.md",
    "chars": 3184,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "Workshop/sessions/tracking/data/ADL-Rundle-6/det.txt",
    "chars": 238414,
    "preview": "1,-1,1691.97,381.048,152.23,352.617,0.995616,-1,-1,-1\n1,-1,1233.55,467.507,133.65,218.985,0.980069,-1,-1,-1\n1,-1,108.484"
  },
  {
    "path": "Workshop/sessions/tracking/data/ADL-Rundle-8/det.txt",
    "chars": 286808,
    "preview": "1,-1,472.128,437.845,78.717,161.327,0.988879,-1,-1,-1\n1,-1,1469.45,415.254,91.93,187.818,0.987547,-1,-1,-1\n1,-1,1352.14,"
  },
  {
    "path": "Workshop/sessions/tracking/data/ETH-Bahnhof/det.txt",
    "chars": 340657,
    "preview": "1,-1,286.552,154.138,71.337,167.328,0.998331,-1,-1,-1\n1,-1,222.571,179.989,33.767,108.927,0.995026,-1,-1,-1\n1,-1,456.542"
  },
  {
    "path": "Workshop/sessions/tracking/data/ETH-Pedcross2/det.txt",
    "chars": 253192,
    "preview": "1,-1,11.2975,148.802,79.1665,197.675,0.996934,-1,-1,-1\n1,-1,127.539,149.108,61.742,185.324,0.996496,-1,-1,-1\n2,-1,18.788"
  },
  {
    "path": "Workshop/sessions/tracking/data/ETH-Sunnyday/det.txt",
    "chars": 119602,
    "preview": "1,-1,291.557,192.468,39.494,119.227,0.995128,-1,-1,-1\n1,-1,53.8152,193.418,48.4288,130.772,0.992938,-1,-1,-1\n1,-1,586.98"
  },
  {
    "path": "Workshop/sessions/tracking/data/KITTI-13/det.txt",
    "chars": 51457,
    "preview": "4,-1,748.744,152.562,32.441,55.121,0.672558,-1,-1,-1\n5,-1,747.246,152.965,34.561,53.954,0.778875,-1,-1,-1\n6,-1,748.641,1"
  },
  {
    "path": "Workshop/sessions/tracking/data/KITTI-17/det.txt",
    "chars": 32194,
    "preview": "1,-1,477.417,133.575,60.972,206.672,0.985314,-1,-1,-1\n1,-1,387.83,158.467,100.509,182.139,0.97094,-1,-1,-1\n2,-1,417.554,"
  },
  {
    "path": "Workshop/sessions/tracking/data/PETS09-S2L1/det.txt",
    "chars": 238216,
    "preview": "1,-1,649.441,231.502,44.417,86.13,0.995474,-1,-1,-1\n1,-1,252.783,207.732,35.813,96.641,0.991175,-1,-1,-1\n1,-1,499.296,15"
  },
  {
    "path": "Workshop/sessions/tracking/data/TUD-Campus/det.txt",
    "chars": 17492,
    "preview": "1,-1,281.931,187.466,79.93,209.537,0.997784,-1,-1,-1\n1,-1,56.6878,144.225,93.5572,295.907,0.997601,-1,-1,-1\n1,-1,378.618"
  },
  {
    "path": "Workshop/sessions/tracking/data/TUD-Stadtmitte/det.txt",
    "chars": 52208,
    "preview": "1,-1,340.829,79.4999,87.662,244.25,0.998128,-1,-1,-1\n1,-1,570.789,80.8348,68.211,203.59,0.998087,-1,-1,-1\n1,-1,181.388,8"
  },
  {
    "path": "Workshop/sessions/tracking/data/Venice-2/det.txt",
    "chars": 301407,
    "preview": "1,-1,1312.61,434.465,203.63,350.962,0.995395,-1,-1,-1\n1,-1,1459.3,441.642,119.05,340.046,0.985937,-1,-1,-1\n1,-1,442.754,"
  },
  {
    "path": "Workshop/sessions/tracking/sort.py",
    "chars": 7533,
    "preview": "\"\"\"\n    SORT: A Simple, Online and Realtime Tracker\n    Copyright (C) 2016 Alex Bewley alex@dynamicdetection.com\n\n    Th"
  },
  {
    "path": "Workshop/sessions/tracking/tracking_kalman.ipynb",
    "chars": 423750,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/tracking/tracking_siamese.ipynb",
    "chars": 35082,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {\n    \"deletable\": true,\n    \"editable\": true\n   },\n   \"sou"
  },
  {
    "path": "Workshop/sessions/utils.py",
    "chars": 2978,
    "preview": "import matplotlib.pyplot as plt\nimport numpy as np\nimport itertools\nimport keras.backend as K\n\ndef plot_confusion_matrix"
  },
  {
    "path": "doc/AI_chip.md",
    "chars": 6575,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "doc/Basics_Machine_learning.md",
    "chars": 23865,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "doc/CNN.md",
    "chars": 25956,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "doc/Kaggle.md",
    "chars": 10746,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "doc/NLP.md",
    "chars": 7224,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "doc/RNN.md",
    "chars": 23349,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "doc/Recommendation.md",
    "chars": 44569,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "doc/SVM.md",
    "chars": 8407,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "doc/Search.md",
    "chars": 8747,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "doc/Tensorflow_Keras.md",
    "chars": 20500,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "doc/decision_tree.md",
    "chars": 17208,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "doc/industrial_design.md",
    "chars": 64431,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "doc/pytorch.md",
    "chars": 12712,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "doc/system.md",
    "chars": 59491,
    "preview": "<!-- START doctoc generated TOC please keep comment here to allow auto update -->\n<!-- DON'T EDIT THIS SECTION, INSTEAD "
  },
  {
    "path": "human-activity-recognition-with-smartphones/.gitignore",
    "chars": 39,
    "preview": ".ipynb_checkpoints/*\ntest.csv\ntrain.csv"
  },
  {
    "path": "human-activity-recognition-with-smartphones/ADL_sensor_data.ipynb",
    "chars": 36932,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"code\",\n   \"execution_count\": null,\n   \"metadata\": {\n    \"collapsed\": true\n   },\n   \"o"
  },
  {
    "path": "keras_FNN_BitTiger.py",
    "chars": 13369,
    "preview": "\n# coding: utf-8\n\n# # Codelab for Feedforward Neural Net\n# \n# All rights reserved.\n# \n# This material cannot be publishe"
  },
  {
    "path": "sklearn_test.py",
    "chars": 242,
    "preview": "from sklearn import datasets\niris = datasets.load_iris()\ndigits = datasets.load_digits()\nfrom sklearn import svm\nclf = s"
  },
  {
    "path": "tensor_flow_test.py",
    "chars": 99,
    "preview": "import os;\nimport inspect;\nimport tensorflow;\n\nprint(os.path.dirname(inspect.getfile(tensorflow)))\n"
  },
  {
    "path": "xgboost/.gitignore",
    "chars": 26,
    "preview": ".ipynb_checkpoints/*\n*.csv"
  },
  {
    "path": "xgboost/tree_xgboost.ipynb",
    "chars": 404998,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {},\n   \"source\": [\n    \"# Comparison of Tree and [xgboost]("
  },
  {
    "path": "xgboost/xgboost_feature.ipynb",
    "chars": 261108,
    "preview": "{\n \"cells\": [\n  {\n   \"cell_type\": \"markdown\",\n   \"metadata\": {},\n   \"source\": [\n    \"# Xgboost Feature\"\n   ]\n  },\n  {\n  "
  }
]

// ... and 7 more files (download for full content)

About this extraction

This page contains the full source code of the zhangruiskyline/DeepLearning GitHub repository, extracted and formatted as plain text for AI agents and large language models (LLMs). The extraction includes 149 files (171.7 MB), approximately 4.2M tokens, and a symbol index with 130 extracted functions, classes, methods, constants, and types. Use this with OpenClaw, Claude, ChatGPT, Cursor, Windsurf, or any other AI tool that accepts text input. You can copy the full output to your clipboard or download it as a .txt file.

Extracted by GitExtract — free GitHub repo to text converter for AI. Built by Nikandr Surkov.

Copied to clipboard!