shape-0.0001.csv

shape.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.6/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "import tensorflow\n",
    "\n",
    "import keras_rcnn.datasets.shape\n",
    "import keras_rcnn.models\n",
    "import keras_rcnn.preprocessing\n",
    "import keras_rcnn.utils"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "configuration = tensorflow.ConfigProto()\n",
    "\n",
    "configuration.gpu_options.allow_growth = True\n",
    "\n",
    "configuration.gpu_options.visible_device_list = \"1\"\n",
    "\n",
    "session = tensorflow.Session(config=configuration)\n",
    "\n",
    "keras.backend.set_session(session)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "categories = {\n",
    "    \"circle\": 1,\n",
    "    \"rectangle\": 2,\n",
    "    \"triangle\": 3\n",
    "}\n",
    "\n",
    "training_dictionaries, test_dictionaries = keras_rcnn.datasets.shape.load_data()\n",
    "\n",
    "generator = keras_rcnn.preprocessing.ObjectDetectionGenerator()\n",
    "\n",
    "generator = generator.flow_from_dictionary(\n",
    "    dictionary=training_dictionaries,\n",
    "    categories=categories,\n",
    "    target_size=(256, 256)\n",
    ")\n",
    "\n",
    "validation_data = keras_rcnn.preprocessing.ObjectDetectionGenerator()\n",
    "\n",
    "validation_data = validation_data.flow_from_dictionary(\n",
    "    dictionary=test_dictionaries,\n",
    "    categories=categories,\n",
    "    target_size=(256, 256)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/storage/com/github/broadinstitute/keras-rcnn/keras_rcnn/models/_rcnn.py:373: UserWarning: Output \"object_detection_2\" missing from loss dictionary. We assume this was done on purpose, and we will not be expecting any data to be passed to \"object_detection_2\" during training.\n",
      "  super(RCNN, self).compile(optimizer, None)\n",
      "/usr/local/lib/python3.6/site-packages/tensorflow/python/ops/gradients_impl.py:100: UserWarning: Converting sparse IndexedSlices to a dense Tensor of unknown shape. This may consume a large amount of memory.\n",
      "  \"Converting sparse IndexedSlices to a dense Tensor of unknown shape. \"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/0x00b1/com/github/scikit-image/scikit-image/skimage/transform/_warps.py:24: UserWarning: The default multichannel argument (None) is deprecated.  Please specify either True or False explicitly.  multichannel will default to False starting with release 0.16.\n",
      "  warn('The default multichannel argument (None) is deprecated.  Please '\n",
      "/home/0x00b1/com/github/scikit-image/scikit-image/skimage/transform/_warps.py:105: UserWarning: The default mode, 'constant', will be changed to 'reflect' in skimage 0.15.\n",
      "  warn(\"The default mode, 'constant', will be changed to 'reflect' in \"\n",
      "/home/0x00b1/com/github/scikit-image/scikit-image/skimage/transform/_warps.py:110: UserWarning: Anti-aliasing will be enabled by default in skimage 0.15 to avoid aliasing artifacts when down-sampling images.\n",
      "  warn(\"Anti-aliasing will be enabled by default in skimage 0.15 to \"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "256/256 [==============================] - 45s 175ms/step - loss: 2.0908\n",
      "Epoch 2/100\n",
      "256/256 [==============================] - 41s 161ms/step - loss: 1.8718\n",
      "Epoch 3/100\n",
      "256/256 [==============================] - 41s 159ms/step - loss: 1.7826\n",
      "Epoch 4/100\n",
      "256/256 [==============================] - 40s 158ms/step - loss: 1.7202\n",
      "Epoch 5/100\n",
      "256/256 [==============================] - 40s 157ms/step - loss: 1.6324\n",
      "Epoch 6/100\n",
      "256/256 [==============================] - 40s 157ms/step - loss: 1.5461\n",
      "Epoch 7/100\n",
      "256/256 [==============================] - 40s 156ms/step - loss: 1.4596\n",
      "Epoch 8/100\n",
      "256/256 [==============================] - 40s 155ms/step - loss: 1.3670\n",
      "Epoch 9/100\n",
      "256/256 [==============================] - 43s 166ms/step - loss: 1.2746\n",
      "Epoch 10/100\n",
      "256/256 [==============================] - 43s 166ms/step - loss: 1.2227\n",
      "Epoch 11/100\n",
      "256/256 [==============================] - 41s 161ms/step - loss: 1.1378\n",
      "Epoch 12/100\n",
      "256/256 [==============================] - 41s 158ms/step - loss: 1.0527\n",
      "Epoch 13/100\n",
      "256/256 [==============================] - 40s 157ms/step - loss: 0.9792\n",
      "Epoch 14/100\n",
      "256/256 [==============================] - 40s 155ms/step - loss: 0.8874\n",
      "Epoch 15/100\n",
      "256/256 [==============================] - 39s 154ms/step - loss: 0.8340\n",
      "Epoch 16/100\n",
      "256/256 [==============================] - 40s 155ms/step - loss: 0.8049\n",
      "Epoch 17/100\n",
      "256/256 [==============================] - 39s 153ms/step - loss: 0.7366\n",
      "Epoch 18/100\n",
      "256/256 [==============================] - 39s 153ms/step - loss: 0.7036\n",
      "Epoch 19/100\n",
      "256/256 [==============================] - 40s 154ms/step - loss: 0.6563\n",
      "Epoch 20/100\n",
      "256/256 [==============================] - 41s 161ms/step - loss: 0.6076\n",
      "Epoch 21/100\n",
      "256/256 [==============================] - 41s 160ms/step - loss: 0.5938\n",
      "Epoch 22/100\n",
      "256/256 [==============================] - 39s 154ms/step - loss: 0.5471\n",
      "Epoch 23/100\n",
      "256/256 [==============================] - 39s 154ms/step - loss: 0.5107\n",
      "Epoch 24/100\n",
      "256/256 [==============================] - 39s 152ms/step - loss: 0.4975\n",
      "Epoch 25/100\n",
      "256/256 [==============================] - 39s 154ms/step - loss: 0.4807\n",
      "Epoch 26/100\n",
      "165/256 [==================>...........] - ETA: 13s - loss: 0.4598"
     ]
    }
   ],
   "source": [
    "learning_rate = 0.0001\n",
    "\n",
    "keras.backend.set_learning_phase(1)\n",
    "\n",
    "model = keras_rcnn.models.RCNN(\n",
    "    backbone=keras_rcnn.models.backbone.VGG19,\n",
    "    categories=categories.keys(),\n",
    "    dense_units=512,\n",
    "    input_shape=(256, 256, 3)\n",
    ")\n",
    "\n",
    "optimizer = keras.optimizers.Adam(learning_rate)\n",
    "\n",
    "model.compile(optimizer)\n",
    "\n",
    "filename = \"shape-weights-{}.hdf5\".format(learning_rate)\n",
    "\n",
    "callbacks = [\n",
    "    keras.callbacks.ModelCheckpoint(filename, \"loss\", save_best_only=True, save_weights_only=True),\n",
    "    keras.callbacks.CSVLogger(\"shape-{}.csv\".format(learning_rate)),\n",
    "    keras.callbacks.TerminateOnNaN()\n",
    "]\n",
    "\n",
    "model.fit_generator(\n",
    "    callbacks=callbacks,\n",
    "    epochs=100,\n",
    "    generator=generator\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}