DSOD_train.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import keras\n",
    "import os\n",
    "import time\n",
    "import glob\n",
    "from tqdm import tqdm\n",
    "\n",
    "from ssd_model import DSOD300, DSOD512\n",
    "from ssd_utils import PriorUtil\n",
    "from ssd_data import InputGenerator\n",
    "from ssd_training import SSDLoss, SSDFocalLoss\n",
    "from utils.training import Logger, LearningRateDecay, AdamAccumulate\n",
    "from utils.model import load_weights\n",
    "\n",
    "\n",
    "np.set_printoptions(linewidth=120, suppress=True, precision=8)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Pascal VOC 2012 + MS COCO\n",
    "\n",
    "from data_voc import GTUtility\n",
    "gt_util_voc = GTUtility('data/VOC2012/')\n",
    "\n",
    "from data_coco import GTUtility\n",
    "gt_util_coco = GTUtility('data/COCO/')\n",
    "gt_util_coco_val = GTUtility('data/COCO/', validation=True)\n",
    "\n",
    "gt_util_train = GTUtility.merge(gt_util_voc, gt_util_coco.convert_to_voc())\n",
    "gt_util_val = gt_util_coco_val.convert_to_voc()\n",
    "\n",
    "class_weights = np.array([0.00007205, 1.3919328 , 1.43665262, 1.30902077, 1.36668928, 1.2391509 , 1.21337629, \n",
    "                          0.41527107, 1.1458096 , 0.29150119, 1.25713104, 0.61941517, 1.03175604, 1.21542005, \n",
    "                          1.01947561, 0.0542007 , 1.12664538, 1.14966073, 1.12464889, 1.49998021, 1.09218961])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# DSOD 300\n",
    "model = DSOD300(num_classes=gt_util_train.num_classes, softmax=False)\n",
    "prior_util = PriorUtil(model)\n",
    "\n",
    "experiment = 'dsod300fl_voccoco'\n",
    "batch_size = 16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# DSOD 512\n",
    "model = DSOD512(num_classes=gt_util_train.num_classes, softmax=False)\n",
    "prior_util = PriorUtil(model)\n",
    "\n",
    "experiment = 'dsod512fl_voccoco'\n",
    "batch_size = 6"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "prior_util.compute_class_weights(gt_util_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# DSOD paper\n",
    "# batch size 128\n",
    "# 320k iterations\n",
    "# initial learning rate 0.1\n",
    "\n",
    "epochs = 1000\n",
    "initial_epoch = 0\n",
    "\n",
    "gen_train = InputGenerator(gt_util_train, prior_util, batch_size, model.image_size, augmentation=True)\n",
    "gen_val = InputGenerator(gt_util_val, prior_util, batch_size, model.image_size, augmentation=True)\n",
    "\n",
    "checkdir = './checkpoints/' + time.strftime('%Y%m%d%H%M') + '_' + experiment\n",
    "if not os.path.exists(checkdir):\n",
    "    os.makedirs(checkdir)\n",
    "\n",
    "with open(checkdir+'/source.py','wb') as f:\n",
    "    source = ''.join(['# In[%i]\\n%s\\n\\n' % (i, In[i]) for i in range(len(In))])\n",
    "    f.write(source.encode())\n",
    "\n",
    "#optim = keras.optimizers.SGD(lr=1e-3, momentum=0.9, decay=0, nesterov=True)\n",
    "#optim = keras.optimizers.Adam(lr=0.1, beta_1=0.9, beta_2=0.999, epsilon=1e-08, decay=0.0)\n",
    "optim = AdamAccumulate(lr=0.1, beta_1=0.9, beta_2=0.999, epsilon=1e-08, accum_iters=128//batch_size) # virtual batch size 128\n",
    "\n",
    "# weight decay\n",
    "regularizer = keras.regularizers.l2(5e-4) # None if disabled\n",
    "for l in model.layers:\n",
    "    if l.__class__.__name__.startswith('Conv'):\n",
    "        l.kernel_regularizer = regularizer\n",
    "\n",
    "#loss = SSDLoss(alpha=1.0, neg_pos_ratio=3.0)\n",
    "loss = SSDFocalLoss(lambda_conf=1.0, class_weights=class_weights)\n",
    "\n",
    "model.compile(optimizer=optim, loss=loss.compute, metrics=loss.metrics)\n",
    "        \n",
    "history = model.fit_generator(\n",
    "        gen_train.generate(), \n",
    "        steps_per_epoch=gen_train.num_batches,\n",
    "        epochs=epochs, \n",
    "        verbose=1, \n",
    "        callbacks=[\n",
    "            keras.callbacks.ModelCheckpoint(checkdir+'/weights.{epoch:03d}.h5', verbose=1, save_weights_only=True),\n",
    "            Logger(checkdir),\n",
    "            #LearningRateDecay()\n",
    "        ], \n",
    "        validation_data=gen_val.generate(), \n",
    "        validation_steps=gen_val.num_batches, \n",
    "        class_weight=None, \n",
    "        #max_queue_size=10, \n",
    "        workers=1, \n",
    "        #use_multiprocessing=False, \n",
    "        initial_epoch=initial_epoch)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model.load_weights('./checkpoints/201808261035_dsod300fl_voccoco/weights.110.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "_, inputs, images, data = gt_util_val.sample_random_batch(batch_size=16, input_size=model.image_size, seed=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "preds = model.predict(inputs, batch_size=1, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "for i in range(len(preds)):\n",
    "    plt.figure(figsize=[8]*2)\n",
    "    plt.imshow(images[i])\n",
    "    res = prior_util.decode(preds[i], confidence_threshold=0.40)\n",
    "    prior_util.plot_results(res, classes=gt_util_val.classes, show_labels=True, gt_data=data[i])\n",
    "    plt.axis('off')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}