{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import tensorflow as tf\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import math\n",
    "import matplotlib.pyplot as plt\n",
    "import msgpack as msg\n",
    "import msgpack_numpy as m\n",
    "from sklearn.model_selection import train_test_split\n",
    "from skimage.transform import resize\n",
    "# from sklearn.color import rgb2gray\n",
    "m.patch()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "def load_face_files():\n",
    "    all_data = [msg.load(open('./face_images/face_images{}.bin'.format(i),'rb')) for i in range(1,6)]\n",
    "    images = np.vstack([i[b'images'] for i in all_data])\n",
    "    gray_images = np.dot(images,np.array([0.2125,0.7154,0.0721]))/255.0\n",
    "#     print(gray_images.shape)\n",
    "#     scaled_gray_images = resize(,(32,32))/255.0\n",
    "#     import pdb;pdb.set_trace()\n",
    "    coords = np.vstack([i[b'co-ords'] for i in all_data])\n",
    "    coords_norm = coords/255.0\n",
    "    return gray_images,coords_norm\n",
    "\n",
    "images,coords = load_face_files()\n",
    "\n",
    "def plot_image(idx,n=1):\n",
    "    im = images[idx]\n",
    "    part_coords = np.split(coords[idx],3)\n",
    "    plt.figure(figsize=(16,16),dpi=80)\n",
    "    plt.subplot(n,4,1)\n",
    "    plt.imshow(im,'gray')\n",
    "    for i in range(2,5):\n",
    "        [x,y,w,h] = part_coords[i-2]#[:4]\n",
    "        # print([x,y,w,h],all([i<0 for i in [x,y,w,h]]))\n",
    "        if not all([j<0 for j in [x,y,w,h]]):\n",
    "            plt.subplot(n,4,i)\n",
    "            plt.imshow(im[y:y+h,x:x+w],'gray')\n",
    "\n",
    "\n",
    "def get_head_images(c):\n",
    "    h_idx = []\n",
    "    for (idx,i) in enumerate(c):\n",
    "        head_coords = np.split(i,3)[0]\n",
    "        if not any([j<0 for j in head_coords]):\n",
    "            h_idx.append(idx)\n",
    "    return h_idx\n",
    "\n",
    "# plot_image(958)\n",
    "\n",
    "head_idxs = get_head_images(coords)\n",
    "head_images = images[head_idxs].reshape(-1,images.shape[1]*images.shape[2])\n",
    "head_coords = coords[head_idxs,:4].astype(np.float32)\n",
    "tr_head_imgs,te_head_imgs,tr_head_crds,te_head_crds = train_test_split(head_images,head_coords,test_size=0.33)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "def create_model(input_dim,output_dim):\n",
    "    g = tf.Graph()\n",
    "    with g.as_default():\n",
    "        learning_rate = tf.constant(0.1,name='learning_rate')\n",
    "        beta = tf.constant(0.00001,name='regularization_beta')\n",
    "        error_upper_bound = tf.constant(1.1,name='upper_bound')\n",
    "        error_lower_bound = tf.constant(0.9,name='lower_bound')\n",
    "        x = tf.placeholder(tf.float32, [None,input_dim])\n",
    "        y = tf.placeholder(tf.float32, [None,output_dim])\n",
    "        W1 = tf.Variable(tf.random_normal([input_dim, 512],stddev=2.0/math.sqrt(input_dim)),name='layer_1_weights')\n",
    "        b1 = tf.Variable(tf.random_normal([512]),name='bias_1_weights')\n",
    "        W2 = tf.Variable(tf.random_normal([512, 128],stddev=2.0/math.sqrt(512)),name='layer_2_weights')\n",
    "        b2 = tf.Variable(tf.random_normal([128]),name='bias_2_weights')\n",
    "        W_o = tf.Variable(tf.random_normal([128, output_dim],stddev=2.0/math.sqrt(128)),name='layer_output_weights')\n",
    "        b_o = tf.Variable(tf.random_normal([output_dim]),name='bias_output_weights')\n",
    "        layer_1 = tf.nn.relu(tf.add(tf.matmul(x,W1),b1))\n",
    "        layer_2 = tf.nn.relu(tf.add(tf.matmul(layer_1,W2),b2))\n",
    "        y_ = tf.nn.sigmoid(tf.add(tf.matmul(layer_2,W_o),b_o))\n",
    "        distance = tf.losses.mean_squared_error(labels = y,predictions = y_)\n",
    "        regularized_loss  = distance+beta*(tf.nn.l2_loss(W1)+tf.nn.l2_loss(W2)+tf.nn.l2_loss(W_o)+tf.nn.l2_loss(b1)+tf.nn.l2_loss(b2)+tf.nn.l2_loss(b_o))\n",
    "        ratio = tf.div(y,y_)\n",
    "        accuracy = distance#tf.reduce_mean(tf.cast((ratio < error_upper_bound) & (ratio > error_lower_bound), tf.float32))\n",
    "        tf.summary.scalar('distance', distance)\n",
    "        tf.summary.histogram('Weights1', W1)\n",
    "        tf.summary.histogram('Bias1', b1)\n",
    "        tf.summary.histogram('Weights2', W2)\n",
    "        tf.summary.histogram('Bias2', b2)\n",
    "        tf.summary.histogram('Weights_output', W_o)\n",
    "        tf.summary.histogram('Bias_output', b_o)\n",
    "        merged_summary = tf.summary.merge_all()\n",
    "        train_step = tf.train.GradientDescentOptimizer(learning_rate).minimize(regularized_loss)\n",
    "        return (g,x,y,y_,train_step,accuracy,merged_summary)\n",
    "\n",
    "(g,x,y,y_,train_step,accuracy,merged_summary) = create_model(tr_head_imgs.shape[1],tr_head_crds.shape[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "distance on validation set 0.0457288585603\n",
      "distance on validation set 0.0384670570493\n",
      "distance on validation set 0.0463402196765\n",
      "distance on validation set 0.0418722033501\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0mTraceback (most recent call last)",
      "\u001b[0;32m<ipython-input-56-ee62be87709e>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     21\u001b[0m                     \u001b[0;32mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"distance on validation set {}\"\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0macc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;31m#,'saved to ',save_path)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     22\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 23\u001b[0;31m \u001b[0mtrain_model\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mg\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0my\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0my_\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mtrain_step\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0maccuracy\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mmerged_summary\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m<ipython-input-56-ee62be87709e>\u001b[0m in \u001b[0;36mtrain_model\u001b[0;34m(g, x, y, y_, train_step, accuracy, merged_summary)\u001b[0m\n\u001b[1;32m     12\u001b[0m             \u001b[0;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m20000\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     13\u001b[0m                 \u001b[0mbatch_xs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mbatch_ys\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbatch_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtr_head_imgs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mtr_head_crds\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m128\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 14\u001b[0;31m                 \u001b[0ms\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mtrain_step\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mfeed_dict\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m{\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mbatch_xs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0my\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mbatch_ys\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     15\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0mi\u001b[0m\u001b[0;34m%\u001b[0m\u001b[0;36m100\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     16\u001b[0m                     \u001b[0mt_batch_xs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mt_batch_ys\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbatch_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtr_head_imgs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mtr_head_crds\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m32\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc\u001b[0m in \u001b[0;36mrun\u001b[0;34m(self, fetches, feed_dict, options, run_metadata)\u001b[0m\n\u001b[1;32m    765\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    766\u001b[0m       result = self._run(None, fetches, feed_dict, options_ptr,\n\u001b[0;32m--> 767\u001b[0;31m                          run_metadata_ptr)\n\u001b[0m\u001b[1;32m    768\u001b[0m       \u001b[0;32mif\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    769\u001b[0m         \u001b[0mproto_data\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtf_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTF_GetBuffer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrun_metadata_ptr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc\u001b[0m in \u001b[0;36m_run\u001b[0;34m(self, handle, fetches, feed_dict, options, run_metadata)\u001b[0m\n\u001b[1;32m    963\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mfinal_fetches\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mfinal_targets\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    964\u001b[0m       results = self._do_run(handle, final_targets, final_fetches,\n\u001b[0;32m--> 965\u001b[0;31m                              feed_dict_string, options, run_metadata)\n\u001b[0m\u001b[1;32m    966\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    967\u001b[0m       \u001b[0mresults\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc\u001b[0m in \u001b[0;36m_do_run\u001b[0;34m(self, handle, target_list, fetch_list, feed_dict, options, run_metadata)\u001b[0m\n\u001b[1;32m   1013\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mhandle\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1014\u001b[0m       return self._do_call(_run_fn, self._session, feed_dict, fetch_list,\n\u001b[0;32m-> 1015\u001b[0;31m                            target_list, options, run_metadata)\n\u001b[0m\u001b[1;32m   1016\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1017\u001b[0m       return self._do_call(_prun_fn, self._session, handle, feed_dict,\n",
      "\u001b[0;32m/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc\u001b[0m in \u001b[0;36m_do_call\u001b[0;34m(self, fn, *args)\u001b[0m\n\u001b[1;32m   1020\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0m_do_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1021\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1022\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1023\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0merrors\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mOpError\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1024\u001b[0m       \u001b[0mmessage\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcompat\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mas_text\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmessage\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc\u001b[0m in \u001b[0;36m_run_fn\u001b[0;34m(session, feed_dict, fetch_list, target_list, options, run_metadata)\u001b[0m\n\u001b[1;32m   1002\u001b[0m         return tf_session.TF_Run(session, options,\n\u001b[1;32m   1003\u001b[0m                                  \u001b[0mfeed_dict\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfetch_list\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtarget_list\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1004\u001b[0;31m                                  status, run_metadata)\n\u001b[0m\u001b[1;32m   1005\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1006\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_prun_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msession\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhandle\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeed_dict\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfetch_list\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "def batch_data(data_x,data_y,size=128):\n",
    "    batch_idxs = np.random.randint(0,data_x.shape[0],size=size)\n",
    "    return (data_x[batch_idxs],data_y[batch_idxs])\n",
    "\n",
    "def train_model(g,x,y,y_,train_step,accuracy,merged_summary):\n",
    "    with g.as_default():\n",
    "        with tf.Session() as s:\n",
    "            train_writer = tf.summary.FileWriter('./tensor_log',s.graph)\n",
    "            tf.global_variables_initializer().run()\n",
    "#             saver = tf.train.Saver()\n",
    "            # saver.restore(s, \"/tmp/model.ckpt\")\n",
    "            for i in range(20000):\n",
    "                batch_xs,batch_ys = batch_data(tr_head_imgs,tr_head_crds,128)\n",
    "                s.run([train_step],feed_dict={x:batch_xs,y:batch_ys})\n",
    "                if i%100 == 0:\n",
    "                    t_batch_xs,t_batch_ys = batch_data(tr_head_imgs,tr_head_crds,32)\n",
    "                    [acc]= s.run([accuracy], feed_dict={x: t_batch_xs,y: t_batch_ys})\n",
    "                   # train_writer.add_summary(summary,i)\n",
    "#                     save_path = saver.save(s, \"/tmp/model.ckpt\")\n",
    "#                     print(y_vals,t_batch_ys)\n",
    "                    print(\"distance on validation set {}\".format(acc))#,'saved to ',save_path)\n",
    "\n",
    "train_model(g,x,y,y_,train_step,accuracy,merged_summary)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "hide_input": false,
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}