├── .gitignore
├── README.md
├── notebooks
    ├── data_exploration.ipynb
    ├── input_mask_analysis.ipynb
    ├── model.py
    └── model_cnn.ipynb
└── requirements.txt


/.gitignore:
--------------------------------------------------------------------------------
  1 | # Byte-compiled / optimized / DLL files
  2 | __pycache__/
  3 | *.py[cod]
  4 | *$py.class
  5 | 
  6 | # C extensions
  7 | *.so
  8 | 
  9 | # Distribution / packaging
 10 | .Python
 11 | env/
 12 | build/
 13 | develop-eggs/
 14 | dist/
 15 | downloads/
 16 | eggs/
 17 | .eggs/
 18 | lib/
 19 | lib64/
 20 | parts/
 21 | sdist/
 22 | var/
 23 | wheels/
 24 | *.egg-info/
 25 | .installed.cfg
 26 | *.egg
 27 | 
 28 | # PyInstaller
 29 | #  Usually these files are written by a python script from a template
 30 | #  before PyInstaller builds the exe, so as to inject date/other infos into it.
 31 | *.manifest
 32 | *.spec
 33 | 
 34 | # Installer logs
 35 | pip-log.txt
 36 | pip-delete-this-directory.txt
 37 | 
 38 | # Unit test / coverage reports
 39 | htmlcov/
 40 | .tox/
 41 | .coverage
 42 | .coverage.*
 43 | .cache
 44 | nosetests.xml
 45 | coverage.xml
 46 | *.cover
 47 | .hypothesis/
 48 | 
 49 | # Translations
 50 | *.mo
 51 | *.pot
 52 | 
 53 | # Django stuff:
 54 | *.log
 55 | local_settings.py
 56 | 
 57 | # Flask stuff:
 58 | instance/
 59 | .webassets-cache
 60 | 
 61 | # Scrapy stuff:
 62 | .scrapy
 63 | 
 64 | # Sphinx documentation
 65 | docs/_build/
 66 | 
 67 | # PyBuilder
 68 | target/
 69 | 
 70 | # Jupyter Notebook
 71 | .ipynb_checkpoints
 72 | 
 73 | # pyenv
 74 | .python-version
 75 | 
 76 | # celery beat schedule file
 77 | celerybeat-schedule
 78 | 
 79 | # SageMath parsed files
 80 | *.sage.py
 81 | 
 82 | # dotenv
 83 | .env
 84 | 
 85 | # virtualenv
 86 | .venv
 87 | venv/
 88 | ENV/
 89 | venv-carvana/
 90 | 
 91 | # Spyder project settings
 92 | .spyderproject
 93 | .spyproject
 94 | 
 95 | # Rope project settings
 96 | .ropeproject
 97 | 
 98 | # mkdocs documentation
 99 | /site
100 | 
101 | # mypy
102 | .mypy_cache/
103 | 
104 | car/
105 | 


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
1 | # carvana-image-masking-challenge
2 | https://www.kaggle.com/c/carvana-image-masking-challenge
3 | 


--------------------------------------------------------------------------------
/notebooks/data_exploration.ipynb:
--------------------------------------------------------------------------------
  1 | {
  2 |  "cells": [
  3 |   {
  4 |    "cell_type": "code",
  5 |    "execution_count": 1,
  6 |    "metadata": {},
  7 |    "outputs": [
  8 |     {
  9 |      "name": "stderr",
 10 |      "output_type": "stream",
 11 |      "text": [
 12 |       "/kaggle/dev/ashish/carvana-image-masking-challenge/car/lib/python3.5/site-packages/sklearn/cross_validation.py:44: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. Also note that the interface of the new CV iterators are different from that of this module. This module will be removed in 0.20.\n",
 13 |       "  \"This module will be removed in 0.20.\", DeprecationWarning)\n",
 14 |       "Using TensorFlow backend.\n"
 15 |      ]
 16 |     }
 17 |    ],
 18 |    "source": [
 19 |     "import numpy as np # linear algebra\n",
 20 |     "import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n",
 21 |     "import matplotlib.pyplot as plt\n",
 22 |     "import matplotlib.image as mpimg\n",
 23 |     "import seaborn as sns\n",
 24 |     "from sklearn import model_selection, preprocessing, metrics\n",
 25 |     "from sklearn.utils import shuffle\n",
 26 |     "import xgboost as xgb\n",
 27 |     "import os\n",
 28 |     "import re\n",
 29 |     "import time\n",
 30 |     "#import h5py\n",
 31 |     "import pickle\n",
 32 |     "from tqdm import tqdm\n",
 33 |     "#import cv2\n",
 34 |     "from keras.models import Sequential, load_model\n",
 35 |     "from keras.layers.convolutional import Convolution2D\n",
 36 |     "from keras.layers.pooling import MaxPooling2D\n",
 37 |     "from keras.layers.core import Activation, Dropout, Flatten, Dense, Lambda\n",
 38 |     "from keras.layers import ELU\n",
 39 |     "from keras.optimizers import Adam\n",
 40 |     "import keras.backend.tensorflow_backend as KTF\n",
 41 |     "from keras.callbacks import EarlyStopping, ModelCheckpoint, TensorBoard\n",
 42 |     "color = sns.color_palette()\n",
 43 |     "\n",
 44 |     "%matplotlib inline"
 45 |    ]
 46 |   },
 47 |   {
 48 |    "cell_type": "code",
 49 |    "execution_count": 9,
 50 |    "metadata": {},
 51 |    "outputs": [
 52 |     {
 53 |      "name": "stdout",
 54 |      "output_type": "stream",
 55 |      "text": [
 56 |       "(2, 1) (1, 2)\n"
 57 |      ]
 58 |     },
 59 |     {
 60 |      "data": {
 61 |       "text/plain": [
 62 |        "array([[1, 2],\n",
 63 |        "       [3, 6]])"
 64 |       ]
 65 |      },
 66 |      "execution_count": 9,
 67 |      "metadata": {},
 68 |      "output_type": "execute_result"
 69 |     }
 70 |    ],
 71 |    "source": [
 72 |     "A = np.array([1, 3]).reshape((2, 1))\n",
 73 |     "B = np.array([1, 2]).reshape((1, 2))\n",
 74 |     "print(A.shape, B.shape)\n",
 75 |     "np.dot(A, B)"
 76 |    ]
 77 |   },
 78 |   {
 79 |    "cell_type": "code",
 80 |    "execution_count": 8,
 81 |    "metadata": {
 82 |     "collapsed": true
 83 |    },
 84 |    "outputs": [],
 85 |    "source": [
 86 |     "DATA_PATH = '/kaggle/dev/carvana-image-masking-challenge-data'\n",
 87 |     "RAW_DATA_PATH = os.path.join(DATA_PATH, 'raw_data')\n",
 88 |     "TRAIN_PATH = os.path.join(RAW_DATA_PATH, 'train')\n",
 89 |     "TEST_PATH = os.path.join(RAW_DATA_PATH, 'test')\n",
 90 |     "TRAIN_MASKS_PATH = os.path.join(RAW_DATA_PATH, 'train_masks')\n",
 91 |     "TRAIN_MASKS_CSV_PATH = os.path.join(RAW_DATA_PATH, 'train_masks.csv')\n",
 92 |     "SAMPLE_SUBMISSION_PATH = os.path.join(RAW_DATA_PATH, 'sample_submission.csv')\n",
 93 |     "METADATA_PATH = os.path.join(RAW_DATA_PATH, 'metadata.csv')"
 94 |    ]
 95 |   },
 96 |   {
 97 |    "cell_type": "code",
 98 |    "execution_count": 104,
 99 |    "metadata": {},
100 |    "outputs": [
101 |     {
102 |      "name": "stdout",
103 |      "output_type": "stream",
104 |      "text": [
105 |       "train_masks_df.shape (5088, 2)\n"
106 |      ]
107 |     },
108 |     {
109 |      "data": {
110 |       "text/html": [
111 |        "<div>\n",
112 |        "<style>\n",
113 |        "    .dataframe thead tr:only-child th {\n",
114 |        "        text-align: right;\n",
115 |        "    }\n",
116 |        "\n",
117 |        "    .dataframe thead th {\n",
118 |        "        text-align: left;\n",
119 |        "    }\n",
120 |        "\n",
121 |        "    .dataframe tbody tr th {\n",
122 |        "        vertical-align: top;\n",
123 |        "    }\n",
124 |        "</style>\n",
125 |        "<table border=\"1\" class=\"dataframe\">\n",
126 |        "  <thead>\n",
127 |        "    <tr style=\"text-align: right;\">\n",
128 |        "      <th></th>\n",
129 |        "      <th>img</th>\n",
130 |        "      <th>rle_mask</th>\n",
131 |        "    </tr>\n",
132 |        "  </thead>\n",
133 |        "  <tbody>\n",
134 |        "    <tr>\n",
135 |        "      <th>0</th>\n",
136 |        "      <td>00087a6bd4dc_01.jpg</td>\n",
137 |        "      <td>879386 40 881253 141 883140 205 885009 17 8850...</td>\n",
138 |        "    </tr>\n",
139 |        "    <tr>\n",
140 |        "      <th>1</th>\n",
141 |        "      <td>00087a6bd4dc_02.jpg</td>\n",
142 |        "      <td>873779 4 875695 7 877612 9 879528 12 881267 15...</td>\n",
143 |        "    </tr>\n",
144 |        "    <tr>\n",
145 |        "      <th>2</th>\n",
146 |        "      <td>00087a6bd4dc_03.jpg</td>\n",
147 |        "      <td>864300 9 866217 13 868134 15 870051 16 871969 ...</td>\n",
148 |        "    </tr>\n",
149 |        "    <tr>\n",
150 |        "      <th>3</th>\n",
151 |        "      <td>00087a6bd4dc_04.jpg</td>\n",
152 |        "      <td>879735 20 881650 26 883315 92 883564 30 885208...</td>\n",
153 |        "    </tr>\n",
154 |        "    <tr>\n",
155 |        "      <th>4</th>\n",
156 |        "      <td>00087a6bd4dc_05.jpg</td>\n",
157 |        "      <td>883365 74 883638 28 885262 119 885550 34 88716...</td>\n",
158 |        "    </tr>\n",
159 |        "  </tbody>\n",
160 |        "</table>\n",
161 |        "</div>"
162 |       ],
163 |       "text/plain": [
164 |        "                   img                                           rle_mask\n",
165 |        "0  00087a6bd4dc_01.jpg  879386 40 881253 141 883140 205 885009 17 8850...\n",
166 |        "1  00087a6bd4dc_02.jpg  873779 4 875695 7 877612 9 879528 12 881267 15...\n",
167 |        "2  00087a6bd4dc_03.jpg  864300 9 866217 13 868134 15 870051 16 871969 ...\n",
168 |        "3  00087a6bd4dc_04.jpg  879735 20 881650 26 883315 92 883564 30 885208...\n",
169 |        "4  00087a6bd4dc_05.jpg  883365 74 883638 28 885262 119 885550 34 88716..."
170 |       ]
171 |      },
172 |      "execution_count": 104,
173 |      "metadata": {},
174 |      "output_type": "execute_result"
175 |     }
176 |    ],
177 |    "source": [
178 |     "train_masks_df = pd.read_csv(TRAIN_MASKS_CSV_PATH)\n",
179 |     "print('train_masks_df.shape', train_masks_df.shape)\n",
180 |     "train_masks_df.head()"
181 |    ]
182 |   },
183 |   {
184 |    "cell_type": "code",
185 |    "execution_count": 132,
186 |    "metadata": {},
187 |    "outputs": [
188 |     {
189 |      "name": "stdout",
190 |      "output_type": "stream",
191 |      "text": [
192 |       "metadata_df.shape (6572, 6)\n"
193 |      ]
194 |     },
195 |     {
196 |      "data": {
197 |       "text/html": [
198 |        "<div>\n",
199 |        "<style>\n",
200 |        "    .dataframe thead tr:only-child th {\n",
201 |        "        text-align: right;\n",
202 |        "    }\n",
203 |        "\n",
204 |        "    .dataframe thead th {\n",
205 |        "        text-align: left;\n",
206 |        "    }\n",
207 |        "\n",
208 |        "    .dataframe tbody tr th {\n",
209 |        "        vertical-align: top;\n",
210 |        "    }\n",
211 |        "</style>\n",
212 |        "<table border=\"1\" class=\"dataframe\">\n",
213 |        "  <thead>\n",
214 |        "    <tr style=\"text-align: right;\">\n",
215 |        "      <th></th>\n",
216 |        "      <th>id</th>\n",
217 |        "      <th>year</th>\n",
218 |        "      <th>make</th>\n",
219 |        "      <th>model</th>\n",
220 |        "      <th>trim1</th>\n",
221 |        "      <th>trim2</th>\n",
222 |        "    </tr>\n",
223 |        "  </thead>\n",
224 |        "  <tbody>\n",
225 |        "    <tr>\n",
226 |        "      <th>0</th>\n",
227 |        "      <td>0004d4463b50</td>\n",
228 |        "      <td>2014.0</td>\n",
229 |        "      <td>Acura</td>\n",
230 |        "      <td>TL</td>\n",
231 |        "      <td>TL</td>\n",
232 |        "      <td>w/SE</td>\n",
233 |        "    </tr>\n",
234 |        "    <tr>\n",
235 |        "      <th>1</th>\n",
236 |        "      <td>00087a6bd4dc</td>\n",
237 |        "      <td>2014.0</td>\n",
238 |        "      <td>Acura</td>\n",
239 |        "      <td>RLX</td>\n",
240 |        "      <td>RLX</td>\n",
241 |        "      <td>w/Tech</td>\n",
242 |        "    </tr>\n",
243 |        "    <tr>\n",
244 |        "      <th>2</th>\n",
245 |        "      <td>000aa097d423</td>\n",
246 |        "      <td>2012.0</td>\n",
247 |        "      <td>Mazda</td>\n",
248 |        "      <td>MAZDA6</td>\n",
249 |        "      <td>MAZDA6</td>\n",
250 |        "      <td>i Sport</td>\n",
251 |        "    </tr>\n",
252 |        "    <tr>\n",
253 |        "      <th>3</th>\n",
254 |        "      <td>000f19f6e7d4</td>\n",
255 |        "      <td>2016.0</td>\n",
256 |        "      <td>Chevrolet</td>\n",
257 |        "      <td>Camaro</td>\n",
258 |        "      <td>Camaro</td>\n",
259 |        "      <td>SS</td>\n",
260 |        "    </tr>\n",
261 |        "    <tr>\n",
262 |        "      <th>4</th>\n",
263 |        "      <td>00144e887ae9</td>\n",
264 |        "      <td>2015.0</td>\n",
265 |        "      <td>Acura</td>\n",
266 |        "      <td>TLX</td>\n",
267 |        "      <td>TLX</td>\n",
268 |        "      <td>SH-AWD V6 w/Advance Pkg</td>\n",
269 |        "    </tr>\n",
270 |        "  </tbody>\n",
271 |        "</table>\n",
272 |        "</div>"
273 |       ],
274 |       "text/plain": [
275 |        "             id    year       make   model   trim1                    trim2\n",
276 |        "0  0004d4463b50  2014.0      Acura      TL      TL                     w/SE\n",
277 |        "1  00087a6bd4dc  2014.0      Acura     RLX     RLX                   w/Tech\n",
278 |        "2  000aa097d423  2012.0      Mazda  MAZDA6  MAZDA6                  i Sport\n",
279 |        "3  000f19f6e7d4  2016.0  Chevrolet  Camaro  Camaro                       SS\n",
280 |        "4  00144e887ae9  2015.0      Acura     TLX     TLX  SH-AWD V6 w/Advance Pkg"
281 |       ]
282 |      },
283 |      "execution_count": 132,
284 |      "metadata": {},
285 |      "output_type": "execute_result"
286 |     }
287 |    ],
288 |    "source": [
289 |     "metadata_df = pd.read_csv(METADATA_PATH)\n",
290 |     "print('metadata_df.shape', metadata_df.shape)\n",
291 |     "metadata_df.head()"
292 |    ]
293 |   },
294 |   {
295 |    "cell_type": "code",
296 |    "execution_count": 125,
297 |    "metadata": {},
298 |    "outputs": [
299 |     {
300 |      "name": "stdout",
301 |      "output_type": "stream",
302 |      "text": [
303 |       "Training set contains 318 cars with 16 images each\n"
304 |      ]
305 |     }
306 |    ],
307 |    "source": [
308 |     "# Verify there are 5088 train images\n",
309 |     "num_train_images = len(os.listdir(TRAIN_PATH))\n",
310 |     "print('Training set contains {} cars with {} images each'.format(int(num_train_images/16), 16))"
311 |    ]
312 |   },
313 |   {
314 |    "cell_type": "code",
315 |    "execution_count": 126,
316 |    "metadata": {},
317 |    "outputs": [
318 |     {
319 |      "name": "stdout",
320 |      "output_type": "stream",
321 |      "text": [
322 |       "Test set contains 6254 cars with 16 images each\n"
323 |      ]
324 |     }
325 |    ],
326 |    "source": [
327 |     "# Verify there are 100064 test images\n",
328 |     "num_train_images = len(os.listdir(TEST_PATH))\n",
329 |     "print('Test set contains {} cars with {} images each'.format(int(num_train_images/16), 16))"
330 |    ]
331 |   },
332 |   {
333 |    "cell_type": "code",
334 |    "execution_count": 127,
335 |    "metadata": {
336 |     "collapsed": true
337 |    },
338 |    "outputs": [],
339 |    "source": [
340 |     "# Helper functions to plot car, mask, masked_car\n",
341 |     "def plot_image(img_id):\n",
342 |     "    img = mpimg.imread(os.path.join(TRAIN_PATH, img_id + \".jpg\"))\n",
343 |     "    imgplot = plt.imshow(img)\n",
344 |     "    plt.axis('off')\n",
345 |     "    plt.show()\n",
346 |     "    \n",
347 |     "def plot_mask(img_id):\n",
348 |     "    mask = mpimg.imread(os.path.join(TRAIN_MASKS_PATH, img_id + \"_mask.gif\"))\n",
349 |     "    imgplot = plt.imshow(mask)\n",
350 |     "    plt.axis('off')\n",
351 |     "    plt.show()\n",
352 |     "    \n",
353 |     "def plot_masked_image(img_id):\n",
354 |     "    img = mpimg.imread(os.path.join(TRAIN_PATH, img_id + \".jpg\"))\n",
355 |     "    mask = mpimg.imread(os.path.join(TRAIN_MASKS_PATH, img_id + \"_mask.gif\"))\n",
356 |     "    mask = mask[:,:,0:3]\n",
357 |     "    mask[mask == 255] = 1 \n",
358 |     "    masked_img = img * mask\n",
359 |     "    imgplot = plt.imshow(masked_img)\n",
360 |     "    plt.axis('off')\n",
361 |     "    plt.show()"
362 |    ]
363 |   },
364 |   {
365 |    "cell_type": "code",
366 |    "execution_count": 133,
367 |    "metadata": {
368 |     "scrolled": false
369 |    },
370 |    "outputs": [
371 |     {
372 |      "name": "stdout",
373 |      "output_type": "stream",
374 |      "text": [
375 |       "Car id 6131a03dd028_04\n"
376 |      ]
377 |     },
378 |     {
379 |      "data": {
380 |       "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXEAAAD8CAYAAACB3pQWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTDPNNNMtppmJzzTTTDPdYvo/3/WeHlMK+0EAAAAASUVORK5CYII=\n",
381 |       "text/plain": [
382 |        "<matplotlib.figure.Figure at 0x7f4246a825f8>"
383 |       ]
384 |      },
385 |      "metadata": {},
386 |      "output_type": "display_data"
387 |     },
388 |     {
389 |      "data": {
390 |       "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXEAAAD8CAYAAACB3pQWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngGAiDhBMxAGCiThAMBEHCPY/A4WPy3/W7SkAAAAASUVORK5CYII=\n",
391 |       "text/plain": [
392 |        "<matplotlib.figure.Figure at 0x7f4248bd7278>"
393 |       ]
394 |      },
395 |      "metadata": {},
396 |      "output_type": "display_data"
397 |     },
398 |     {
399 |      "data": {
400 |       "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXEAAAD8CAYAAACB3pQWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nxfw/3l0Ar8powpIAAAAASUVORK5CYII=\n",
401 |       "text/plain": [
402 |        "<matplotlib.figure.Figure at 0x7f4248c1d7b8>"
403 |       ]
404 |      },
405 |      "metadata": {},
406 |      "output_type": "display_data"
407 |     }
408 |    ],
409 |    "source": [
410 |     "# Plot random car, mask, masked_car\n",
411 |     "random_car_idx = int(np.random.random() * len(train_masks_df))\n",
412 |     "random_img_path = train_masks_df.iloc[random_car_idx].values[0]\n",
413 |     "m = re.match(r'([a-f0-9].*).jpg', random_img_path)\n",
414 |     "random_img_id = m.group(1)\n",
415 |     "print('Car id {}'.format(random_img_id))\n",
416 |     "plot_image(random_img_id)\n",
417 |     "plot_mask(random_img_id)\n",
418 |     "plot_masked_image(random_img_id)"
419 |    ]
420 |   },
421 |   {
422 |    "cell_type": "code",
423 |    "execution_count": 121,
424 |    "metadata": {},
425 |    "outputs": [
426 |     {
427 |      "name": "stdout",
428 |      "output_type": "stream",
429 |      "text": [
430 |       "sample_submission_df.shape (100064, 2)\n"
431 |      ]
432 |     },
433 |     {
434 |      "data": {
435 |       "text/html": [
436 |        "<div>\n",
437 |        "<style>\n",
438 |        "    .dataframe thead tr:only-child th {\n",
439 |        "        text-align: right;\n",
440 |        "    }\n",
441 |        "\n",
442 |        "    .dataframe thead th {\n",
443 |        "        text-align: left;\n",
444 |        "    }\n",
445 |        "\n",
446 |        "    .dataframe tbody tr th {\n",
447 |        "        vertical-align: top;\n",
448 |        "    }\n",
449 |        "</style>\n",
450 |        "<table border=\"1\" class=\"dataframe\">\n",
451 |        "  <thead>\n",
452 |        "    <tr style=\"text-align: right;\">\n",
453 |        "      <th></th>\n",
454 |        "      <th>img</th>\n",
455 |        "      <th>rle_mask</th>\n",
456 |        "    </tr>\n",
457 |        "  </thead>\n",
458 |        "  <tbody>\n",
459 |        "    <tr>\n",
460 |        "      <th>0</th>\n",
461 |        "      <td>0004d4463b50_01.jpg</td>\n",
462 |        "      <td>1 1</td>\n",
463 |        "    </tr>\n",
464 |        "    <tr>\n",
465 |        "      <th>1</th>\n",
466 |        "      <td>0004d4463b50_02.jpg</td>\n",
467 |        "      <td>1 1</td>\n",
468 |        "    </tr>\n",
469 |        "    <tr>\n",
470 |        "      <th>2</th>\n",
471 |        "      <td>0004d4463b50_03.jpg</td>\n",
472 |        "      <td>1 1</td>\n",
473 |        "    </tr>\n",
474 |        "    <tr>\n",
475 |        "      <th>3</th>\n",
476 |        "      <td>0004d4463b50_04.jpg</td>\n",
477 |        "      <td>1 1</td>\n",
478 |        "    </tr>\n",
479 |        "    <tr>\n",
480 |        "      <th>4</th>\n",
481 |        "      <td>0004d4463b50_05.jpg</td>\n",
482 |        "      <td>1 1</td>\n",
483 |        "    </tr>\n",
484 |        "  </tbody>\n",
485 |        "</table>\n",
486 |        "</div>"
487 |       ],
488 |       "text/plain": [
489 |        "                   img rle_mask\n",
490 |        "0  0004d4463b50_01.jpg      1 1\n",
491 |        "1  0004d4463b50_02.jpg      1 1\n",
492 |        "2  0004d4463b50_03.jpg      1 1\n",
493 |        "3  0004d4463b50_04.jpg      1 1\n",
494 |        "4  0004d4463b50_05.jpg      1 1"
495 |       ]
496 |      },
497 |      "execution_count": 121,
498 |      "metadata": {},
499 |      "output_type": "execute_result"
500 |     }
501 |    ],
502 |    "source": [
503 |     "# Sample submission\n",
504 |     "sample_submission_df = pd.read_csv(SAMPLE_SUBMISSION_PATH)\n",
505 |     "print('sample_submission_df.shape', sample_submission_df.shape)\n",
506 |     "sample_submission_df.head()"
507 |    ]
508 |   },
509 |   {
510 |    "cell_type": "code",
511 |    "execution_count": 16,
512 |    "metadata": {},
513 |    "outputs": [
514 |     {
515 |      "name": "stdout",
516 |      "output_type": "stream",
517 |      "text": [
518 |       "_________________________________________________________________\n",
519 |       "Layer (type)                 Output Shape              Param #   \n",
520 |       "=================================================================\n",
521 |       "input_7 (InputLayer)         (None, 2, 2, 2)           0         \n",
522 |       "_________________________________________________________________\n",
523 |       "up_sampling2d_6 (UpSampling2 (None, 2, 4, 2)           0         \n",
524 |       "=================================================================\n",
525 |       "Total params: 0\n",
526 |       "Trainable params: 0\n",
527 |       "Non-trainable params: 0\n",
528 |       "_________________________________________________________________\n"
529 |      ]
530 |     }
531 |    ],
532 |    "source": [
533 |     "from keras.models import Model, load_model\n",
534 |     "from keras.layers import *\n",
535 |     "from keras.optimizers import Adam, RMSprop\n",
536 |     "from keras.losses import binary_crossentropy\n",
537 |     "import keras.backend.tensorflow_backend as KTF\n",
538 |     "def simple_model():\n",
539 |     "    inputs = Input(shape=(2,2,2))\n",
540 |     "    x = UpSampling2D((1,2))(inputs)\n",
541 |     "    #x = MaxPooling2D((2, 2), strides=(2, 2))(x)\n",
542 |     "    model = Model(inputs=inputs, outputs=x)\n",
543 |     "    return model\n",
544 |     "\n",
545 |     "model = simple_model()\n",
546 |     "model.compile(loss='binary_crossentropy', optimizer=RMSprop(lr=0.1))\n",
547 |     "model.summary()"
548 |    ]
549 |   },
550 |   {
551 |    "cell_type": "code",
552 |    "execution_count": 17,
553 |    "metadata": {},
554 |    "outputs": [
555 |     {
556 |      "name": "stdout",
557 |      "output_type": "stream",
558 |      "text": [
559 |       "[[[1 2]\n",
560 |       "  [3 4]]\n",
561 |       "\n",
562 |       " [[5 6]\n",
563 |       "  [7 8]]]\n"
564 |      ]
565 |     },
566 |     {
567 |      "data": {
568 |       "text/plain": [
569 |        "array([[[ 1.,  2.],\n",
570 |        "        [ 1.,  2.],\n",
571 |        "        [ 3.,  4.],\n",
572 |        "        [ 3.,  4.]],\n",
573 |        "\n",
574 |        "       [[ 5.,  6.],\n",
575 |        "        [ 5.,  6.],\n",
576 |        "        [ 7.,  8.],\n",
577 |        "        [ 7.,  8.]]], dtype=float32)"
578 |       ]
579 |      },
580 |      "execution_count": 17,
581 |      "metadata": {},
582 |      "output_type": "execute_result"
583 |     }
584 |    ],
585 |    "source": [
586 |     "random_input = np.asarray([[[[1, 2], [3, 4]], [[5, 6], [7, 8]]]])\n",
587 |     "print(random_input[0])\n",
588 |     "#print(random_input.shape)\n",
589 |     "model.predict(random_input)[0]"
590 |    ]
591 |   }
592 |  ],
593 |  "metadata": {
594 |   "kernelspec": {
595 |    "display_name": "Python 3",
596 |    "language": "python",
597 |    "name": "python3"
598 |   },
599 |   "language_info": {
600 |    "codemirror_mode": {
601 |     "name": "ipython",
602 |     "version": 3
603 |    },
604 |    "file_extension": ".py",
605 |    "mimetype": "text/x-python",
606 |    "name": "python",
607 |    "nbconvert_exporter": "python",
608 |    "pygments_lexer": "ipython3",
609 |    "version": "3.5.2"
610 |   }
611 |  },
612 |  "nbformat": 4,
613 |  "nbformat_minor": 2
614 | }
615 | 


--------------------------------------------------------------------------------
/notebooks/model.py:
--------------------------------------------------------------------------------
  1 | import os
  2 | import time
  3 | import h5py
  4 | import math
  5 | import pickle
  6 | import numpy as np
  7 | import pandas as pd
  8 | import cv2
  9 | import threading
 10 | import queue
 11 | import matplotlib.pyplot as plt
 12 | import seaborn as sns
 13 | from scipy import misc, ndimage
 14 | from sklearn import model_selection, preprocessing, metrics
 15 | from sklearn.utils import shuffle
 16 | from skimage import transform
 17 | from tqdm import tqdm
 18 | from keras.regularizers import l2
 19 | from keras.models import Model, load_model
 20 | from keras.layers import *
 21 | from keras.optimizers import *
 22 | from keras.callbacks import EarlyStopping, ModelCheckpoint, TensorBoard
 23 | from keras import backend as K
 24 | from keras.losses import binary_crossentropy
 25 | import keras.backend.tensorflow_backend as KTF
 26 | import tensorflow as tf
 27 | from tensorflow.python.client import device_lib
 28 | 
 29 | DATA_PATH = '/kaggle/dev/carvana-image-masking-challenge-data'
 30 | RAW_DATA_PATH = os.path.join(DATA_PATH, 'raw_data')
 31 | TRAIN_PATH = os.path.join(RAW_DATA_PATH, 'train')
 32 | TEST_PATH = os.path.join(RAW_DATA_PATH, 'test')
 33 | TRAIN_MASKS_PATH = os.path.join(RAW_DATA_PATH, 'train_masks')
 34 | #TRAIN_MASKS_FIXED_PATH = os.path.join(DATA_PATH, 'fixed_masks/fix-HCK')
 35 | TRAIN_MASKS_CSV_PATH = os.path.join(RAW_DATA_PATH, 'train_masks.csv')
 36 | SAMPLE_SUBMISSION_PATH = os.path.join(RAW_DATA_PATH, 'sample_submission.csv')
 37 | METADATA_PATH = os.path.join(RAW_DATA_PATH, 'metadata.csv')
 38 | SUBMISSION_PATH = os.path.join(DATA_PATH, 'submissions')
 39 | ASSETS_PATH = os.path.join(DATA_PATH, 'assets')
 40 | MODELS_PATH = os.path.join(ASSETS_PATH, 'models')
 41 | TENSORBOARD_PATH = os.path.join(ASSETS_PATH, 'tensorboard')
 42 | 
 43 | train_masks_df = pd.read_csv(TRAIN_MASKS_CSV_PATH)
 44 | print('train_masks_df.shape', train_masks_df.shape)
 45 | 
 46 | # Constants
 47 | HEIGHT_ORIG = 1280
 48 | WIDTH_ORIG = 1918
 49 | CHANNELS_ORIG = 3
 50 | 
 51 | HEIGHT = 1024
 52 | WIDTH = 1024
 53 | CHANNELS = 3
 54 | new_shape = (HEIGHT, WIDTH, CHANNELS)
 55 | mask_shape = (new_shape[0], new_shape[1], 1)
 56 | 
 57 | def get_img_id(img_path):
 58 |     return img_path[:15]
 59 | 
 60 | img_ids = list(map(get_img_id, list(train_masks_df.img.values)))
 61 | 
 62 | def load_image_disk(img_id, folder=TRAIN_PATH):
 63 |     img = misc.imread(os.path.join(folder, img_id + ".jpg"))
 64 |     return img
 65 | 
 66 | def get_image(img_id):
 67 |     return train_imgs[img_id]
 68 | 
 69 | # Return mask as 1/0 binary img with single channel
 70 | def load_mask_disk(img_id, folder=TRAIN_MASKS_PATH, filetype='gif'):
 71 |     mask = misc.imread(os.path.join(folder,  "{}_mask.{}".format(img_id, filetype)), flatten=True)
 72 |     mask[mask > 128] = 1
 73 |     if len(mask.shape) == 2:
 74 |         mask = mask.reshape(mask.shape[0], mask.shape[1], 1)
 75 |     return mask
 76 | 
 77 | def get_mask(img_id):
 78 |     return train_masks[img_id]
 79 | 
 80 | # Helper functions to plot car, mask, masked_car
 81 | def plot_image(img_id):
 82 |     img = misc.imread(os.path.join(TRAIN_PATH, img_id + ".jpg"))
 83 |     imgplot = plt.imshow(img)
 84 |     plt.axis('off')
 85 |     plt.show()
 86 |     
 87 | def plot_mask(img_id, folder=TRAIN_MASKS_PATH, filetype='gif', ax=None):
 88 |     mask = misc.imread(os.path.join(folder, "{}_mask.{}".format(img_id, filetype)))
 89 |     if ax == None:
 90 |         imgplot = plt.imshow(mask)
 91 |         plt.axis('off')
 92 |         plt.show()
 93 |     else:
 94 |         ax.imshow(mask)
 95 |         ax.axis('off')
 96 |     
 97 | def plot_masked_image(img_id, ax=None):
 98 |     img = misc.imread(os.path.join(TRAIN_PATH, img_id + ".jpg"))
 99 |     mask = misc.imread(os.path.join(TRAIN_MASKS_PATH, img_id + "_mask.gif"))
100 |     mask = mask[:,:,0:3]
101 |     mask[mask == 255] = 1 
102 |     masked_img = img * mask
103 |     if ax == None:
104 |         imgplot = plt.imshow(masked_img)
105 |         plt.axis('off')
106 |         plt.show()
107 |     else:
108 |         ax.imshow(masked_img)
109 |         ax.axis('off')
110 | 
111 | def gray2rgb(img):
112 |     img = np.squeeze(img)
113 |     w, h = img.shape
114 |     ret = np.empty((w, h, 3), dtype=np.uint8)
115 |     ret[:, :, 0] = img
116 |     ret[:, :, 1] = img
117 |     ret[:, :, 2] = img
118 |     return ret
119 | 
120 | def resize_img(img, new_s = new_shape):
121 |     return transform.resize(img, new_s)
122 | 
123 | train_imgs = {}
124 | for img_path in tqdm(os.listdir(TRAIN_PATH)):
125 |     img_id = get_img_id(img_path)
126 |     train_imgs[img_id] = cv2.resize(load_image_disk(img_id), (new_shape[0], new_shape[1]))
127 |     
128 | train_masks = {}
129 | for img_path in tqdm(os.listdir(TRAIN_MASKS_PATH)):
130 |     img_id = get_img_id(img_path)
131 |     train_masks[img_id] = np.expand_dims(cv2.resize(load_mask_disk(img_id), (new_shape[0], new_shape[1])), axis=2)
132 | 
133 | def randomHueSaturationValue(image, hue_shift_limit=(-180, 180),
134 |                              sat_shift_limit=(-255, 255),
135 |                              val_shift_limit=(-255, 255), u=0.5):
136 |     if np.random.random() < u:
137 |         image = cv2.cvtColor(image, cv2.COLOR_RGB2HSV)
138 |         h, s, v = cv2.split(image)
139 |         hue_shift = np.random.uniform(hue_shift_limit[0], hue_shift_limit[1])
140 |         h = cv2.add(h, hue_shift)
141 |         sat_shift = np.random.uniform(sat_shift_limit[0], sat_shift_limit[1])
142 |         s = cv2.add(s, sat_shift)
143 |         val_shift = np.random.uniform(val_shift_limit[0], val_shift_limit[1])
144 |         v = cv2.add(v, val_shift)
145 |         image = cv2.merge((h, s, v))
146 |         image = cv2.cvtColor(image, cv2.COLOR_HSV2RGB)
147 |     return image
148 | 
149 | def randomShiftScaleRotate(image, mask,
150 |                            shift_limit=(-0.0625, 0.0625),
151 |                            scale_limit=(-0.1, 0.1),
152 |                            rotate_limit=(-45, 45), aspect_limit=(0, 0),
153 |                            borderMode=cv2.BORDER_REFLECT_101, u=0.5):
154 |     if np.random.random() < u:
155 |         height, width, channel = image.shape
156 | 
157 |         angle = np.random.uniform(rotate_limit[0], rotate_limit[1])  # degree
158 |         scale = np.random.uniform(1 + scale_limit[0], 1 + scale_limit[1])
159 |         aspect = np.random.uniform(1 + aspect_limit[0], 1 + aspect_limit[1])
160 |         sx = scale * aspect / (aspect ** 0.5)
161 |         sy = scale / (aspect ** 0.5)
162 |         dx = round(np.random.uniform(shift_limit[0], shift_limit[1]) * width)
163 |         dy = round(np.random.uniform(shift_limit[0], shift_limit[1]) * height)
164 | 
165 |         cc = np.math.cos(angle / 180 * np.math.pi) * sx
166 |         ss = np.math.sin(angle / 180 * np.math.pi) * sy
167 |         rotate_matrix = np.array([[cc, -ss], [ss, cc]])
168 | 
169 |         box0 = np.array([[0, 0], [width, 0], [width, height], [0, height], ])
170 |         box1 = box0 - np.array([width / 2, height / 2])
171 |         box1 = np.dot(box1, rotate_matrix.T) + np.array([width / 2 + dx, height / 2 + dy])
172 | 
173 |         box0 = box0.astype(np.float32)
174 |         box1 = box1.astype(np.float32)
175 |         mat = cv2.getPerspectiveTransform(box0, box1)
176 |         image = cv2.warpPerspective(image, mat, (width, height), flags=cv2.INTER_LINEAR, borderMode=borderMode,
177 |                                     borderValue=(0, 0, 0,))
178 |         mask = cv2.warpPerspective(mask, mat, (width, height), flags=cv2.INTER_LINEAR, borderMode=borderMode,
179 |                                    borderValue=(0, 0, 0,))
180 |         if len(mask.shape) == 2:
181 |             mask = np.expand_dims(mask, axis=2)
182 | 
183 |     return image, mask
184 | 
185 | def randomHorizontalFlip(image, mask, u=0.5):
186 |     if np.random.random() < u:
187 |         image = cv2.flip(image, 1)
188 |         mask = cv2.flip(mask, 1)
189 | 
190 |     return image, mask
191 | 
192 | def generate_training_batch(data, batch_size):
193 |     while True:
194 |         X_batch = []
195 |         Y_batch = []
196 |         batch_ids = np.random.choice(data,
197 |                                      size=batch_size,
198 |                                      replace=False)
199 |         for idx, img_id in enumerate(batch_ids):
200 |             x = get_image(img_id)
201 |             y = get_mask(img_id)
202 |             x, y = randomShiftScaleRotate(x, y,
203 |                                           shift_limit=(-0.0625, 0.0625),
204 |                                           scale_limit=(-0.1, 0.1),
205 |                                           rotate_limit=(-0, 0))
206 | #             x = randomHueSaturationValue(x,
207 | #                                hue_shift_limit=(-50, 50),
208 | #                                sat_shift_limit=(-5, 5),
209 | #                                val_shift_limit=(-15, 15))
210 |             X_batch.append(x)
211 |             Y_batch.append(y)
212 |         X = np.asarray(X_batch, dtype=np.float32)
213 |         Y = np.asarray(Y_batch, dtype=np.float32)
214 |         yield X, Y
215 | 
216 | def generate_validation_batch(data, batch_size):
217 |     while True:
218 |         X_batch = []
219 |         Y_batch = []
220 |         batch_ids = np.random.choice(data,
221 |                                      size=batch_size,
222 |                                      replace=False)
223 |         for idx, img_id in enumerate(batch_ids):
224 |             x = get_image(img_id)
225 |             y = get_mask(img_id)
226 |             X_batch.append(x)
227 |             Y_batch.append(y)
228 |         X = np.asarray(X_batch, dtype=np.float32)
229 |         Y = np.asarray(Y_batch, dtype=np.float32)
230 |         yield X, Y
231 | 
232 | def generate_validation_data_seq(data):
233 |     idx = 0
234 |     while True:
235 |         img_id = data[idx]
236 |         X = get_image(img_id)
237 |         Y = get_mask(img_id)
238 |         yield img_id, X, Y
239 |         idx  += 1
240 |         if idx >= len(data):
241 |             break
242 | 
243 | def get_model_memory_usage(batch_size, model):
244 |     from keras import backend as K
245 | 
246 |     shapes_mem_count = 0
247 |     for l in model.layers:
248 |         single_layer_mem = 1
249 |         for s in l.output_shape:
250 |             if s is None:
251 |                 continue
252 |             single_layer_mem *= s
253 |         shapes_mem_count += single_layer_mem
254 | 
255 |     trainable_count = int(np.sum([K.count_params(p) for p in set(model.trainable_weights)]))
256 |     non_trainable_count = int(np.sum([K.count_params(p) for p in set(model.non_trainable_weights)]))
257 | 
258 |     total_memory = 4*batch_size*(shapes_mem_count + trainable_count + non_trainable_count)
259 |     gbytes = round(total_memory / (1024 ** 3), 3)
260 |     mbytes = round(total_memory / (1024 ** 2), 3)
261 |     
262 |     print('trainable_count', trainable_count, 'non_trainable_count', non_trainable_count, 'gbytes', gbytes, 'mbytes', mbytes)
263 | 
264 | def down(filters, input_):
265 |     down_ = Conv2D(filters, (3, 3), padding='same')(input_)
266 |     down_ = BatchNormalization(epsilon=1e-4)(down_)
267 |     down_ = Activation('relu')(down_)
268 |     down_ = Conv2D(filters, (3, 3), padding='same')(down_)
269 |     down_ = BatchNormalization(epsilon=1e-4)(down_)
270 |     down_res = Activation('relu')(down_)
271 |     down_pool = MaxPooling2D((2, 2), strides=(2, 2))(down_)
272 |     return down_pool, down_res
273 | 
274 | def up(filters, input_, down_):
275 |     up_ = UpSampling2D((2, 2))(input_)
276 |     up_ = concatenate([down_, up_], axis=3)
277 |     up_ = Conv2D(filters, (3, 3), padding='same')(up_)
278 |     up_ = BatchNormalization(epsilon=1e-4)(up_)
279 |     up_ = Activation('relu')(up_)
280 |     up_ = Conv2D(filters, (3, 3), padding='same')(up_)
281 |     up_ = BatchNormalization(epsilon=1e-4)(up_)
282 |     up_ = Activation('relu')(up_)
283 |     up_ = Conv2D(filters, (3, 3), padding='same')(up_)
284 |     up_ = BatchNormalization(epsilon=1e-4)(up_)
285 |     up_ = Activation('relu')(up_)
286 |     return up_
287 | 
288 | def get_unet_1024(input_shape=(HEIGHT, WIDTH, CHANNELS), num_classes=1):
289 |     inputs = Input(shape=input_shape)
290 | 
291 |     #down0b, down0b_res = down(8, inputs)
292 |     down0a, down0a_res = down(24, inputs)
293 |     down0, down0_res = down(64, down0a)
294 |     down1, down1_res = down(128, down0)
295 |     down2, down2_res = down(256, down1)
296 |     down3, down3_res = down(512, down2)
297 |     down4, down4_res = down(768, down3)
298 |     
299 |     center = Conv2D(768, (3, 3), padding='same')(down4)
300 |     center = BatchNormalization(epsilon=1e-4)(center)
301 |     center = Activation('relu')(center)
302 |     center = Conv2D(768, (3, 3), padding='same')(center)
303 |     center = BatchNormalization(epsilon=1e-4)(center)
304 |     center = Activation('relu')(center)
305 | 
306 |     up4 = up(768, center, down4_res)
307 |     up3 = up(512, up4, down3_res)
308 |     up2 = up(256, up3, down2_res)
309 |     up1 = up(128, up2, down1_res)
310 |     up0 = up(64, up1, down0_res)
311 |     up0a = up(24, up0, down0a_res)
312 |     #up0b = up(8, up0a, down0b_res)
313 | 
314 |     classify = Conv2D(num_classes, (1, 1), activation='sigmoid', name='final_layer')(up0a)
315 | 
316 |     model = Model(inputs=inputs, outputs=classify)
317 | 
318 |     return model
319 | 
320 | def dice_coef(y_true, y_pred, smooth=1):
321 |     y_true_f = K.flatten(y_true)
322 |     y_pred_f = K.flatten(y_pred)
323 |     
324 |     intersection = K.sum(y_true_f * y_pred_f)
325 |     return (2. * intersection + smooth) / (K.sum(y_true_f) + K.sum(y_pred_f) + smooth)
326 | 
327 | def dice_coef_loss(y_true, y_pred):
328 |     return 1-dice_coef(y_true, y_pred)
329 | 
330 | def bce_dice_loss(y_true, y_pred):
331 |     return binary_crossentropy(y_true, y_pred) + dice_coef_loss(y_true, y_pred)
332 | 
333 | BATCH_SIZE = 2
334 | 
335 | # Training new model
336 | ts = str(int(time.time()))
337 | model_name = 'malhot'
338 | num_epochs = 30
339 | steps_per_epoch = int(len(img_ids) * 0.8/BATCH_SIZE)
340 | run_name = 'model={}-batch_size={}-num_epoch={}-steps_per_epoch={}-ts={}'.format(model_name,
341 |                                                                           BATCH_SIZE,
342 |                                                                           num_epochs,
343 |                                                                           steps_per_epoch,
344 |                                                                           ts)
345 | tensorboard_loc = os.path.join(TENSORBOARD_PATH, run_name)
346 | checkpoint_loc = os.path.join(MODELS_PATH, 'model-{}-weights.h5'.format(ts))
347 | 
348 | earlyStopping = EarlyStopping(monitor='val_loss', 
349 |                               patience=2, 
350 |                               verbose=1, 
351 |                               min_delta = 0.0001,
352 |                               mode='min',)
353 | 
354 | modelCheckpoint = ModelCheckpoint(checkpoint_loc,
355 |                                   monitor = 'val_loss', 
356 |                                   save_best_only = True, 
357 |                                   mode = 'min', 
358 |                                   verbose = 1,
359 |                                   save_weights_only = True)
360 | 
361 | tensorboard = TensorBoard(log_dir=tensorboard_loc, histogram_freq=0, write_graph=True, write_images=True)
362 | 
363 | callbacks_list = [modelCheckpoint, earlyStopping, tensorboard]
364 | 
365 | model = get_unet_1024()
366 | model.compile(loss=bce_dice_loss, optimizer=Adam(lr=1e-4), metrics=[dice_coef])
367 | print(model.summary())
368 | get_model_memory_usage(BATCH_SIZE, model)
369 | 
370 | train_ids, validation_ids = model_selection.train_test_split(img_ids, random_state=42, test_size=0.20)
371 | train_generator = generate_training_batch(train_ids, BATCH_SIZE)
372 | valid_generator = generate_validation_batch(validation_ids, BATCH_SIZE)
373 | VALIDATION_STEPS = int(len(validation_ids) / BATCH_SIZE)
374 | 
375 | print('Starting run {}'.format(run_name))
376 | history = model.fit_generator(
377 |         train_generator, 
378 |         steps_per_epoch = steps_per_epoch, 
379 |         epochs = num_epochs,
380 |         callbacks = callbacks_list,
381 |         verbose = 1,
382 |         validation_data = valid_generator,
383 |         validation_steps = VALIDATION_STEPS)
384 | 
385 | model_path = os.path.join(MODELS_PATH, 'model-{}.h5'.format(ts))
386 | history_path = os.path.join(MODELS_PATH, 'model-{}.history'.format(ts))
387 | model.save(model_path)
388 | pickle.dump(history.history, open(history_path, "wb"))
389 | print('Saved model at {}'.format(model_path))
390 | print('Saved model history at {}'.format(history_path))
391 | 


--------------------------------------------------------------------------------
/requirements.txt:
--------------------------------------------------------------------------------
  1 | appdirs==1.4.3
  2 | bleach==2.0.0
  3 | cycler==0.10.0
  4 | decorator==4.0.11
  5 | entrypoints==0.2.2
  6 | h5py==2.7.0
  7 | html5lib==0.999999999
  8 | ipykernel==4.6.1
  9 | ipython==6.0.0
 10 | ipython-genutils==0.2.0
 11 | ipywidgets==6.0.0
 12 | jedi==0.10.2
 13 | Jinja2==2.9.6
 14 | jsonschema==2.6.0
 15 | jupyter==1.0.0
 16 | jupyter-client==5.0.1
 17 | jupyter-console==5.1.0
 18 | jupyter-core==4.3.0
 19 | Keras==2.0.4
 20 | MarkupSafe==1.0
 21 | matplotlib==2.0.2
 22 | mistune==0.7.4
 23 | nbconvert==5.1.1
 24 | nbformat==4.3.0
 25 | networkx==1.11
 26 | notebook==5.0.0
 27 | numpy==1.13.0
 28 | olefile==0.44
 29 | packaging==16.8
 30 | pandas==0.20.1
 31 | pandocfilters==1.4.1
 32 | pexpect==4.2.1
 33 | pickleshare==0.7.4
 34 | Pillow==4.2.1
 35 | prompt-toolkit==1.0.14
 36 | protobuf==3.3.0
 37 | ptyprocess==0.5.1
 38 | Pygments==2.2.0
 39 | pyparsing==2.2.0
 40 | pyreadline==2.1
 41 | python-dateutil==2.6.0
 42 | pytz==2017.2
 43 | PyWavelets==0.5.2
 44 | PyYAML==3.12
 45 | pyzmq==16.0.2
 46 | qtconsole==4.3.0
 47 | scikit-image==0.13.0
 48 | scikit-learn==0.18.1
 49 | scipy==0.19.0
 50 | seaborn==0.7.1
 51 | simplegeneric==0.8.1
 52 | six==1.10.0
 53 | sklearn==0.0
 54 | tensorflow==1.0.0
 55 | tensorflow-gpu==1.1.0
 56 | terminado==0.6
 57 | testpath==0.3
 58 | Theano==0.9.0
 59 | tornado==4.5.1
 60 | tqdm==4.15.0
 61 | traitlets==4.3.2
 62 | wcwidth==0.1.7
 63 | webencodings==0.5.1
 64 | Werkzeug==0.12.2
 65 | widgetsnbextension==2.0.0
 66 | xgboost==0.6a2
 67 | appdirs==1.4.3
 68 | bleach==2.0.0
 69 | cycler==0.10.0
 70 | decorator==4.0.11
 71 | entrypoints==0.2.2
 72 | h5py==2.7.0
 73 | html5lib==0.999999999
 74 | ipykernel==4.6.1
 75 | ipython==6.0.0
 76 | ipython-genutils==0.2.0
 77 | ipywidgets==6.0.0
 78 | jedi==0.10.2
 79 | Jinja2==2.9.6
 80 | jsonschema==2.6.0
 81 | jupyter==1.0.0
 82 | jupyter-client==5.0.1
 83 | jupyter-console==5.1.0
 84 | jupyter-core==4.3.0
 85 | Keras==2.0.4
 86 | MarkupSafe==1.0
 87 | matplotlib==2.0.2
 88 | mistune==0.7.4
 89 | nbconvert==5.1.1
 90 | nbformat==4.3.0
 91 | networkx==1.11
 92 | notebook==5.0.0
 93 | numpy==1.13.0
 94 | olefile==0.44
 95 | opencv-python==3.3.0.9
 96 | packaging==16.8
 97 | pandas==0.20.1
 98 | pandocfilters==1.4.1
 99 | pexpect==4.2.1
100 | pickleshare==0.7.4
101 | Pillow==4.2.1
102 | prompt-toolkit==1.0.14
103 | protobuf==3.3.0
104 | ptyprocess==0.5.1
105 | Pygments==2.2.0
106 | pyparsing==2.2.0
107 | pyreadline==2.1
108 | python-dateutil==2.6.0
109 | pytz==2017.2
110 | PyWavelets==0.5.2
111 | PyYAML==3.12
112 | pyzmq==16.0.2
113 | qtconsole==4.3.0
114 | scikit-image==0.13.0
115 | scikit-learn==0.18.1
116 | scipy==0.19.0
117 | seaborn==0.7.1
118 | simplegeneric==0.8.1
119 | six==1.10.0
120 | sklearn==0.0
121 | tensorflow==1.0.0
122 | tensorflow-gpu==1.1.0
123 | terminado==0.6
124 | testpath==0.3
125 | Theano==0.9.0
126 | tornado==4.5.1
127 | tqdm==4.15.0
128 | traitlets==4.3.2
129 | wcwidth==0.1.7
130 | webencodings==0.5.1
131 | Werkzeug==0.12.2
132 | widgetsnbextension==2.0.0
133 | xgboost==0.6a2
134 | 


--------------------------------------------------------------------------------