generate_patches_RGBA.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The autoreload extension is already loaded. To reload it, use:\n",
      "  %reload_ext autoreload\n"
     ]
    }
   ],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "import cPickle as pickle\n",
    "import mahotas as mh\n",
    "import numpy as np\n",
    "import os\n",
    "import time\n",
    "\n",
    "from matplotlib.pyplot import imshow\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "import mlproof as mlp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "PATCH_PATH = '/home/d/patches/cylinder2_rgba_small/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "def generate_patches(start_slice, end_slice):\n",
    "\n",
    "    patch_index = 0\n",
    "\n",
    "    all_patches_count = 0\n",
    "    patch_list = []\n",
    "    all_error_patches = []\n",
    "    all_correct_patches = []\n",
    "    \n",
    "    for z in range(start_slice, end_slice):\n",
    "\n",
    "        t0 = time.time()\n",
    "        print 'working on slice', z\n",
    "        input_image, input_prob, input_mask, input_gold, input_rhoana = mlp.Util.read_section('/home/d/data/cylinderNEW/',z)\n",
    "\n",
    "        error_patches, patches = mlp.Patch.patchify_maxoverlap(input_image, input_prob, input_mask, input_rhoana, input_gold, sample_rate=1)\n",
    "\n",
    "#         for e in patches:\n",
    "#             if e['image'].max() == 0.:\n",
    "#                 print 'ZERRRRO', z\n",
    "        \n",
    "        print 'Generated', len(error_patches), 'split error patches and', len(patches), ' correct patches in', time.time()-t0, 'seconds.'\n",
    "        \n",
    "        patch_list.append(patches)\n",
    "        all_error_patches += error_patches\n",
    "        all_correct_patches += patches\n",
    "        \n",
    "    \n",
    "    \n",
    "    NO_PATCHES = len(all_error_patches) + len(all_correct_patches)\n",
    "\n",
    "    print 'We have a total of',NO_PATCHES,'patches.'\n",
    "    print 'Errors:',len(all_error_patches)\n",
    "    print 'Correct:',len(all_correct_patches)    \n",
    "    \n",
    "#     with open(PATCH_PATH+'pickled_errors.p') as f:\n",
    "#         pickle.dump(all_error_patches)\n",
    "#     with open(PATCH_PATH+'pickled_correct.p') as f:\n",
    "#         pickle.dump(all_correct_patches)\n",
    "    \n",
    "#     print 'pickled.'\n",
    "    \n",
    "    PATCH_BYTES = 75*75\n",
    "    P_SIZE = (NO_PATCHES, 4, 75,75) # rather than raveled right now\n",
    "    \n",
    "    p_rgba = np.zeros(P_SIZE, dtype=np.float32)\n",
    "    p_rgba_large = np.zeros(P_SIZE, dtype=np.float32)    \n",
    "    \n",
    "#     p_image = np.zeros(P_SIZE, dtype=np.float32)\n",
    "#     p_prob = np.zeros(P_SIZE, dtype=np.float32)\n",
    "#     p_binary = np.zeros(P_SIZE, dtype=np.bool)\n",
    "#     p_merged_array = np.zeros(P_SIZE, dtype=np.bool)\n",
    "# #     p_dyn_obj = np.zeros((NO_PATCHES, PATCH_BYTES),dtype=np.bool)\n",
    "# #     p_dyn_bnd = np.zeros((NO_PATCHES, PATCH_BYTES),dtype=np.bool)\n",
    "#     p_border_overlap = np.zeros(P_SIZE, dtype=np.bool)\n",
    "#     p_larger_border_overlap = np.zeros(P_SIZE, dtype=np.bool)    \n",
    "    p_target = np.zeros(NO_PATCHES)\n",
    "\n",
    "\n",
    "    print 'starting loop'\n",
    "    \n",
    "    i = 0\n",
    "    for p in all_error_patches:\n",
    "#         print 'testing one y=1', p['image'].max()\n",
    "        p_rgba[i][0] = p['image']\n",
    "        p_rgba[i][1] = 1. - p['prob'] \n",
    "        p_rgba[i][2] = p['merged_array']\n",
    "        p_rgba[i][3] = p['border_overlap']\n",
    "        \n",
    "        p_rgba_large[i][0] = p['image']\n",
    "        p_rgba_large[i][1] = 1. - p['prob']    \n",
    "        p_rgba_large[i][2] = p['merged_array']\n",
    "        p_rgba_large[i][3] = p['larger_border_overlap']        \n",
    "        \n",
    "#         p_image[i] = (p['image'].astype(np.float32)/255.)#.ravel()\n",
    "#         p_prob[i] = (p['prob'].astype(np.float32)/255.)#.ravel()\n",
    "#         p_binary[i] = p['binary']#.ravel()\n",
    "#         p_merged_array[i] = p['merged_array']#.ravel()\n",
    "# #         p_dyn_obj[i] = p['dyn_obj'].ravel()\n",
    "# #         p_dyn_bnd[i] = p['dyn_bnd'].ravel()\n",
    "#         p_border_overlap[i] = p['border_overlap']#.ravel()\n",
    "#         p_larger_border_overlap[i] = p['larger_border_overlap']#.ravel()\n",
    "        p_target[i] = 1 # <--- important\n",
    "        i += 1\n",
    "\n",
    "        \n",
    "    for p in all_correct_patches:\n",
    "#         print 'testing one y=0', p['image'].max()\n",
    "        p_rgba[i][0] = p['image']\n",
    "        p_rgba[i][1] = 1. - p['prob']    \n",
    "        p_rgba[i][2] = p['merged_array']\n",
    "        p_rgba[i][3] = p['border_overlap']\n",
    "        \n",
    "        p_rgba_large[i][0] = p['image']\n",
    "        p_rgba_large[i][1] = 1. - p['prob']       \n",
    "        p_rgba_large[i][2] = p['merged_array']\n",
    "        p_rgba_large[i][3] = p['larger_border_overlap']        \n",
    "        \n",
    "#         p_image[i] = (p['image'].astype(np.float32)/255.)#.ravel()\n",
    "#         p_prob[i] = (p['prob'].astype(np.float32)/255.)#.ravel()\n",
    "#         p_binary[i] = p['binary']#.ravel()\n",
    "#         p_merged_array[i] = p['merged_array']#.ravel()\n",
    "# #         p_dyn_obj[i] = p['dyn_obj'].ravel()\n",
    "# #         p_dyn_bnd[i] = p['dyn_bnd'].ravel()\n",
    "#         p_border_overlap[i] = p['border_overlap']#.ravel()\n",
    "#         p_larger_border_overlap[i] = p['larger_border_overlap']#.ravel()\n",
    "        p_target[i] = 0 # <--- important\n",
    "        i+=1\n",
    "        \n",
    "    \n",
    "    return p_rgba, p_rgba_large, p_target\n",
    "\n",
    "\n",
    "\n",
    "def shuffle_in_unison_inplace(a, b, c):\n",
    "    assert len(a) == len(b)\n",
    "    p = np.random.permutation(len(a))\n",
    "    return a[p], b[p], c[p]\n",
    "\n",
    "\n",
    "\n",
    "def run(start_slice, end_slice, filename):\n",
    "    \n",
    "    if not os.path.exists(PATCH_PATH):\n",
    "        os.makedirs(PATCH_PATH)\n",
    "    \n",
    "    p = generate_patches(start_slice, end_slice)\n",
    "    \n",
    "    np.savez(PATCH_PATH+filename+'_unshuffled.npz', rgba=p[0], rgba_large=p[1])\n",
    "    np.savez(PATCH_PATH+filename+'_targets_unshuffled.npz', targets=p[2])\n",
    "    print 'unshuffled saved'\n",
    "    shuffled = shuffle_in_unison_inplace(p[0],\n",
    "                                         p[1],\n",
    "                                         p[2]\n",
    "                                        )\n",
    "    \n",
    "    print 'saving..'\n",
    "    np.savez(PATCH_PATH+filename+'.npz', rgba=shuffled[0],\n",
    "                                         rgba_large=shuffled[1])\n",
    "    np.savez(PATCH_PATH+filename+'_targets.npz', targets=shuffled[2])\n",
    "    print 'Done!'\n",
    "    \n",
    "#     return shuffled\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "working on slice 0\n",
      "Loading z00000000.png\n",
      "Generated 0 split error patches and 0  correct patches in 1.41131496429 seconds.\n",
      "working on slice 1\n",
      "Loading z00000001.png\n",
      "Generated 0 split error patches and 0  correct patches in 1.12953090668 seconds.\n",
      "working on slice 2\n",
      "Loading z00000002.png\n",
      "Generated 0 split error patches and 0  correct patches in 1.12887477875 seconds.\n",
      "working on slice 3\n",
      "Loading z00000003.png\n",
      "Generated 0 split error patches and 0  correct patches in 1.13313388824 seconds.\n",
      "working on slice 4\n",
      "Loading z00000004.png\n",
      "Generated 0 split error patches and 0  correct patches in 1.10736107826 seconds.\n",
      "working on slice 5\n",
      "Loading z00000005.png\n",
      "Generated 0 split error patches and 0  correct patches in 1.10280394554 seconds.\n",
      "working on slice 6\n",
      "Loading z00000006.png\n",
      "Generated 0 split error patches and 0  correct patches in 1.12091302872 seconds.\n",
      "working on slice 7\n",
      "Loading z00000007.png\n",
      "Generated 0 split error patches and 0  correct patches in 1.22936701775 seconds.\n",
      "working on slice 8\n",
      "Loading z00000008.png\n",
      "Generated 0 split error patches and 0  correct patches in 2.21552610397 seconds.\n",
      "working on slice 9\n",
      "Loading z00000009.png\n",
      "Generated 4 split error patches and 4  correct patches in 5.76457595825 seconds.\n",
      "working on slice 10\n",
      "Loading z00000010.png\n",
      "Generated 0 split error patches and 0  correct patches in 4.69685697556 seconds.\n",
      "working on slice 11\n",
      "Loading z00000011.png\n",
      "Generated 4 split error patches and 4  correct patches in 5.57008695602 seconds.\n",
      "working on slice 12\n",
      "Loading z00000012.png\n",
      "Generated 0 split error patches and 0  correct patches in 7.85698199272 seconds.\n",
      "working on slice 13\n",
      "Loading z00000013.png\n",
      "Generated 4 split error patches and 4  correct patches in 14.6601099968 seconds.\n",
      "working on slice 14\n",
      "Loading z00000014.png\n",
      "Generated 12 split error patches and 12  correct patches in 23.6610169411 seconds.\n",
      "working on slice 15\n",
      "Loading z00000015.png\n",
      "Generated 20 split error patches and 20  correct patches in 31.5069689751 seconds.\n",
      "working on slice 16\n",
      "Loading z00000016.png\n",
      "Generated 24 split error patches and 24  correct patches in 35.6922240257 seconds.\n",
      "working on slice 17\n",
      "Loading z00000017.png\n",
      "Generated 28 split error patches and 28  correct patches in 37.468307972 seconds.\n",
      "working on slice 18\n",
      "Loading z00000018.png\n",
      "Generated 16 split error patches and 16  correct patches in 38.6474609375 seconds.\n",
      "working on slice 19\n",
      "Loading z00000019.png\n",
      "Generated 40 split error patches and 40  correct patches in 48.7613370419 seconds.\n",
      "working on slice 20\n",
      "Loading z00000020.png\n",
      "Generated 44 split error patches and 44  correct patches in 49.948693037 seconds.\n",
      "working on slice 21\n",
      "Loading z00000021.png\n",
      "Generated 52 split error patches and 52  correct patches in 58.687702179 seconds.\n",
      "working on slice 22\n",
      "Loading z00000022.png\n",
      "Generated 80 split error patches and 80  correct patches in 63.5206329823 seconds.\n",
      "working on slice 23\n",
      "Loading z00000023.png\n",
      "Generated 120 split error patches and 120  correct patches in 75.9672670364 seconds.\n",
      "working on slice 24\n",
      "Loading z00000024.png\n",
      "Generated 128 split error patches and 128  correct patches in 81.6129589081 seconds.\n",
      "working on slice 25\n",
      "Loading z00000025.png\n",
      "Generated 108 split error patches and 108  correct patches in 89.6895649433 seconds.\n",
      "working on slice 26\n",
      "Loading z00000026.png\n",
      "Generated 188 split error patches and 188  correct patches in 101.025290012 seconds.\n",
      "working on slice 27\n",
      "Loading z00000027.png\n",
      "Generated 116 split error patches and 116  correct patches in 95.698154211 seconds.\n",
      "working on slice 28\n",
      "Loading z00000028.png\n",
      "Generated 256 split error patches and 256  correct patches in 118.565156937 seconds.\n",
      "working on slice 29\n",
      "Loading z00000029.png\n",
      "Generated 228 split error patches and 228  correct patches in 132.280936003 seconds.\n",
      "working on slice 30\n",
      "Loading z00000030.png\n",
      "Generated 248 split error patches and 248  correct patches in 133.10925293 seconds.\n",
      "working on slice 31\n",
      "Loading z00000031.png\n",
      "Generated 228 split error patches and 228  correct patches in 147.133841038 seconds.\n",
      "working on slice 32\n",
      "Loading z00000032.png\n",
      "Generated 280 split error patches and 280  correct patches in 160.573714018 seconds.\n",
      "working on slice 33\n",
      "Loading z00000033.png\n",
      "Generated 488 split error patches and 488  correct patches in 198.714309931 seconds.\n",
      "working on slice 34\n",
      "Loading z00000034.png\n",
      "Generated 432 split error patches and 432  correct patches in 192.058409929 seconds.\n",
      "working on slice 35\n",
      "Loading z00000035.png\n",
      "Generated 496 split error patches and 496  correct patches in 220.318030119 seconds.\n",
      "working on slice 36\n",
      "Loading z00000036.png\n",
      "Generated 560 split error patches and 560  correct patches in 237.716424942 seconds.\n",
      "working on slice 37\n",
      "Loading z00000037.png\n",
      "Generated 480 split error patches and 480  correct patches in 237.466375828 seconds.\n",
      "working on slice 38\n",
      "Loading z00000038.png\n",
      "Generated 636 split error patches and 636  correct patches in 261.414727926 seconds.\n",
      "working on slice 39\n",
      "Loading z00000039.png\n",
      "Generated 536 split error patches and 536  correct patches in 249.264129877 seconds.\n",
      "working on slice 40\n",
      "Loading z00000040.png\n",
      "Generated 744 split error patches and 744  correct patches in 295.664323092 seconds.\n",
      "working on slice 41\n",
      "Loading z00000041.png\n",
      "Generated 836 split error patches and 836  correct patches in 302.703818083 seconds.\n",
      "working on slice 42\n",
      "Loading z00000042.png\n",
      "Generated 868 split error patches and 868  correct patches in 308.151016951 seconds.\n",
      "working on slice 43\n",
      "Loading z00000043.png\n",
      "Generated 952 split error patches and 952  correct patches in 352.526884079 seconds.\n",
      "working on slice 44\n",
      "Loading z00000044.png\n",
      "Generated 996 split error patches and 996  correct patches in 368.925523996 seconds.\n",
      "working on slice 45\n",
      "Loading z00000045.png\n",
      "Generated 728 split error patches and 728  correct patches in 338.827229977 seconds.\n",
      "working on slice 46\n",
      "Loading z00000046.png\n",
      "Generated 620 split error patches and 620  correct patches in 331.799098969 seconds.\n",
      "working on slice 47\n",
      "Loading z00000047.png\n",
      "Generated 612 split error patches and 612  correct patches in 345.889595032 seconds.\n",
      "working on slice 48\n",
      "Loading z00000048.png\n",
      "Generated 652 split error patches and 652  correct patches in 356.863452911 seconds.\n",
      "working on slice 49\n",
      "Loading z00000049.png\n",
      "Generated 840 split error patches and 840  correct patches in 383.008934975 seconds.\n",
      "working on slice 50\n",
      "Loading z00000050.png\n",
      "Generated 960 split error patches and 960  correct patches in 405.958745003 seconds.\n",
      "working on slice 51\n",
      "Loading z00000051.png\n",
      "Generated 884 split error patches and 884  correct patches in 383.994342089 seconds.\n",
      "working on slice 52\n",
      "Loading z00000052.png\n",
      "Generated 744 split error patches and 744  correct patches in 373.592616081 seconds.\n",
      "working on slice 53\n",
      "Loading z00000053.png\n",
      "Generated 800 split error patches and 800  correct patches in 385.871507883 seconds.\n",
      "working on slice 54\n",
      "Loading z00000054.png\n",
      "Generated 848 split error patches and 848  correct patches in 391.951431036 seconds.\n",
      "working on slice 55\n",
      "Loading z00000055.png\n",
      "Generated 608 split error patches and 608  correct patches in 386.085554838 seconds.\n",
      "working on slice 56\n",
      "Loading z00000056.png\n",
      "Generated 576 split error patches and 576  correct patches in 393.193368912 seconds.\n",
      "working on slice 57\n",
      "Loading z00000057.png\n",
      "Generated 592 split error patches and 592  correct patches in 402.884635925 seconds.\n",
      "working on slice 58\n",
      "Loading z00000058.png\n",
      "Generated 756 split error patches and 756  correct patches in 409.775440931 seconds.\n",
      "working on slice 59\n",
      "Loading z00000059.png\n",
      "Generated 1016 split error patches and 1016  correct patches in 446.880795002 seconds.\n",
      "working on slice 60\n",
      "Loading z00000060.png\n",
      "Generated 724 split error patches and 724  correct patches in 424.864104033 seconds.\n",
      "working on slice 61\n",
      "Loading z00000061.png\n",
      "Generated 760 split error patches and 760  correct patches in 422.126460075 seconds.\n",
      "working on slice 62\n",
      "Loading z00000062.png\n",
      "Generated 1048 split error patches and 1048  correct patches in 446.953583002 seconds.\n",
      "working on slice 63\n",
      "Loading z00000063.png\n",
      "Generated 896 split error patches and 896  correct patches in 434.666806936 seconds.\n",
      "working on slice 64\n",
      "Loading z00000064.png\n",
      "Generated 788 split error patches and 788  correct patches in 451.01591301 seconds.\n",
      "working on slice 65\n",
      "Loading z00000065.png\n",
      "Generated 708 split error patches and 708  correct patches in 435.70527792 seconds.\n",
      "working on slice 66\n",
      "Loading z00000066.png\n",
      "Generated 564 split error patches and 564  correct patches in 437.955293894 seconds.\n",
      "working on slice 67\n",
      "Loading z00000067.png\n",
      "Generated 592 split error patches and 592  correct patches in 447.203082085 seconds.\n",
      "working on slice 68\n",
      "Loading z00000068.png\n",
      "Generated 544 split error patches and 544  correct patches in 438.513465881 seconds.\n",
      "working on slice 69\n",
      "Loading z00000069.png\n",
      "Generated 428 split error patches and 428  correct patches in 430.268140078 seconds.\n",
      "working on slice 70\n",
      "Loading z00000070.png\n",
      "Generated 248 split error patches and 248  correct patches in 399.889695168 seconds.\n",
      "working on slice 71\n",
      "Loading z00000071.png\n",
      "Generated 364 split error patches and 364  correct patches in 435.256233931 seconds.\n",
      "working on slice 72\n",
      "Loading z00000072.png\n",
      "Generated 468 split error patches and 468  correct patches in 473.61661005 seconds.\n",
      "working on slice 73\n",
      "Loading z00000073.png\n",
      "Generated 452 split error patches and 452  correct patches in 474.385110855 seconds.\n",
      "working on slice 74\n",
      "Loading z00000074.png\n",
      "Generated 384 split error patches and 384  correct patches in 454.740281105 seconds.\n",
      "working on slice 75\n",
      "Loading z00000075.png\n",
      "Generated 260 split error patches and 260  correct patches in 440.889096022 seconds.\n",
      "working on slice 76\n",
      "Loading z00000076.png\n",
      "Generated 212 split error patches and 212  correct patches in 445.458086967 seconds.\n",
      "working on slice 77\n",
      "Loading z00000077.png\n",
      "Generated 368 split error patches and 368  correct patches in 481.603005886 seconds.\n",
      "working on slice 78\n",
      "Loading z00000078.png\n",
      "Generated 404 split error patches and 404  correct patches in 474.614570856 seconds.\n",
      "working on slice 79\n",
      "Loading z00000079.png\n",
      "Generated 548 split error patches and 548  correct patches in 503.264626026 seconds.\n",
      "working on slice 80\n",
      "Loading z00000080.png\n",
      "Generated 364 split error patches and 364  correct patches in 476.154119968 seconds.\n",
      "working on slice 81\n",
      "Loading z00000081.png\n",
      "Generated 456 split error patches and 456  correct patches in 500.688076973 seconds.\n",
      "working on slice 82\n",
      "Loading z00000082.png\n",
      "Generated 452 split error patches and 452  correct patches in 507.854372025 seconds.\n",
      "working on slice 83\n",
      "Loading z00000083.png\n",
      "Generated 316 split error patches and 316  correct patches in 488.750401974 seconds.\n",
      "working on slice 84\n",
      "Loading z00000084.png\n",
      "Generated 476 split error patches and 476  correct patches in 533.764137983 seconds.\n",
      "working on slice 85\n",
      "Loading z00000085.png\n",
      "Generated 480 split error patches and 480  correct patches in 519.206175089 seconds.\n",
      "working on slice 86\n",
      "Loading z00000086.png\n",
      "Generated 456 split error patches and 456  correct patches in 521.445481062 seconds.\n",
      "working on slice 87\n",
      "Loading z00000087.png\n",
      "Generated 440 split error patches and 440  correct patches in 510.196522951 seconds.\n",
      "working on slice 88\n",
      "Loading z00000088.png\n",
      "Generated 396 split error patches and 396  correct patches in 516.533308983 seconds.\n",
      "working on slice 89\n",
      "Loading z00000089.png\n",
      "Generated 308 split error patches and 308  correct patches in 516.468945026 seconds.\n",
      "working on slice 90\n",
      "Loading z00000090.png\n",
      "Generated 344 split error patches and 344  correct patches in 519.171138048 seconds.\n",
      "working on slice 91\n",
      "Loading z00000091.png\n",
      "Generated 372 split error patches and 372  correct patches in 509.992311001 seconds.\n",
      "working on slice 92\n",
      "Loading z00000092.png\n",
      "Generated 416 split error patches and 416  correct patches in 512.261558056 seconds.\n",
      "working on slice 93\n",
      "Loading z00000093.png\n",
      "Generated 460 split error patches and 460  correct patches in 519.235606909 seconds.\n",
      "working on slice 94\n",
      "Loading z00000094.png\n",
      "Generated 304 split error patches and 304  correct patches in 505.255247116 seconds.\n",
      "working on slice 95\n",
      "Loading z00000095.png\n",
      "Generated 420 split error patches and 420  correct patches in 517.834512949 seconds.\n",
      "working on slice 96\n",
      "Loading z00000096.png\n",
      "Generated 484 split error patches and 484  correct patches in 538.116242886 seconds.\n",
      "working on slice 97\n",
      "Loading z00000097.png\n",
      "Generated 560 split error patches and 560  correct patches in 515.507833004 seconds.\n",
      "working on slice 98\n",
      "Loading z00000098.png\n",
      "Generated 668 split error patches and 668  correct patches in 540.652815104 seconds.\n",
      "working on slice 99\n",
      "Loading z00000099.png\n",
      "Generated 688 split error patches and 688  correct patches in 561.37883687 seconds.\n",
      "working on slice 100\n",
      "Loading z00000100.png\n",
      "Generated 728 split error patches and 728  correct patches in 563.775197983 seconds.\n",
      "working on slice 101\n",
      "Loading z00000101.png\n",
      "Generated 576 split error patches and 576  correct patches in 549.546740055 seconds.\n",
      "working on slice 102\n",
      "Loading z00000102.png\n",
      "Generated 688 split error patches and 688  correct patches in 549.451019049 seconds.\n",
      "working on slice 103\n",
      "Loading z00000103.png\n",
      "Generated 608 split error patches and 608  correct patches in 549.778550863 seconds.\n",
      "working on slice 104\n",
      "Loading z00000104.png\n",
      "Generated 576 split error patches and 576  correct patches in 534.565827131 seconds.\n",
      "working on slice 105\n",
      "Loading z00000105.png\n",
      "Generated 536 split error patches and 536  correct patches in 523.061214924 seconds.\n",
      "working on slice 106\n",
      "Loading z00000106.png\n",
      "Generated 412 split error patches and 412  correct patches in 509.109124899 seconds.\n",
      "working on slice 107\n",
      "Loading z00000107.png\n",
      "Generated 324 split error patches and 324  correct patches in 507.970405102 seconds.\n",
      "working on slice 108\n",
      "Loading z00000108.png\n",
      "Generated 576 split error patches and 576  correct patches in 566.801594973 seconds.\n",
      "working on slice 109\n",
      "Loading z00000109.png\n",
      "Generated 592 split error patches and 592  correct patches in 560.885504961 seconds.\n",
      "working on slice 110\n",
      "Loading z00000110.png\n",
      "Generated 356 split error patches and 356  correct patches in 517.742475033 seconds.\n",
      "working on slice 111\n",
      "Loading z00000111.png\n",
      "Generated 436 split error patches and 436  correct patches in 529.580899954 seconds.\n",
      "working on slice 112\n",
      "Loading z00000112.png\n",
      "Generated 400 split error patches and 400  correct patches in 523.406587124 seconds.\n",
      "working on slice 113\n",
      "Loading z00000113.png\n",
      "Generated 364 split error patches and 364  correct patches in 530.343599081 seconds.\n",
      "working on slice 114\n",
      "Loading z00000114.png\n",
      "Generated 364 split error patches and 364  correct patches in 522.524276018 seconds.\n",
      "working on slice 115\n",
      "Loading z00000115.png\n",
      "Generated 400 split error patches and 400  correct patches in 507.424499035 seconds.\n",
      "working on slice 116\n",
      "Loading z00000116.png\n",
      "Generated 392 split error patches and 392  correct patches in 514.885039091 seconds.\n",
      "working on slice 117\n",
      "Loading z00000117.png\n",
      "Generated 488 split error patches and 488  correct patches in 495.725244999 seconds.\n",
      "working on slice 118\n",
      "Loading z00000118.png\n",
      "Generated 632 split error patches and 632  correct patches in 528.642037868 seconds.\n",
      "working on slice 119\n",
      "Loading z00000119.png\n",
      "Generated 376 split error patches and 376  correct patches in 492.045344114 seconds.\n",
      "working on slice 120\n",
      "Loading z00000120.png\n",
      "Generated 472 split error patches and 472  correct patches in 505.118974924 seconds.\n",
      "working on slice 121\n",
      "Loading z00000121.png\n",
      "Generated 296 split error patches and 296  correct patches in 496.502156973 seconds.\n",
      "working on slice 122\n",
      "Loading z00000122.png\n",
      "Generated 332 split error patches and 332  correct patches in 486.991145134 seconds.\n",
      "working on slice 123\n",
      "Loading z00000123.png\n",
      "Generated 372 split error patches and 372  correct patches in 487.497493029 seconds.\n",
      "working on slice 124\n",
      "Loading z00000124.png\n",
      "Generated 460 split error patches and 460  correct patches in 501.836264133 seconds.\n",
      "working on slice 125\n",
      "Loading z00000125.png\n",
      "Generated 460 split error patches and 460  correct patches in 500.570104122 seconds.\n",
      "working on slice 126\n",
      "Loading z00000126.png\n",
      "Generated 412 split error patches and 412  correct patches in 503.297901869 seconds.\n",
      "working on slice 127\n",
      "Loading z00000127.png\n",
      "Generated 480 split error patches and 480  correct patches in 504.847102165 seconds.\n",
      "working on slice 128\n",
      "Loading z00000128.png\n",
      "Generated 432 split error patches and 432  correct patches in 495.266516209 seconds.\n",
      "working on slice 129\n",
      "Loading z00000129.png\n",
      "Generated 380 split error patches and 380  correct patches in 500.419028997 seconds.\n",
      "working on slice 130\n",
      "Loading z00000130.png\n",
      "Generated 272 split error patches and 272  correct patches in 487.957016945 seconds.\n",
      "working on slice 131\n",
      "Loading z00000131.png\n",
      "Generated 288 split error patches and 288  correct patches in 491.962775946 seconds.\n",
      "working on slice 132\n",
      "Loading z00000132.png\n",
      "Generated 240 split error patches and 240  correct patches in 486.22044301 seconds.\n",
      "working on slice 133\n",
      "Loading z00000133.png\n",
      "Generated 232 split error patches and 232  correct patches in 476.466511965 seconds.\n",
      "working on slice 134\n",
      "Loading z00000134.png\n",
      "Generated 216 split error patches and 216  correct patches in 494.261609077 seconds.\n",
      "working on slice 135\n",
      "Loading z00000135.png\n",
      "Generated 300 split error patches and 300  correct patches in 491.95644784 seconds.\n",
      "working on slice 136\n",
      "Loading z00000136.png\n",
      "Generated 332 split error patches and 332  correct patches in 485.217125177 seconds.\n",
      "working on slice 137\n",
      "Loading z00000137.png\n",
      "Generated 408 split error patches and 408  correct patches in 525.324605942 seconds.\n",
      "working on slice 138\n",
      "Loading z00000138.png\n",
      "Generated 424 split error patches and 424  correct patches in 510.656581879 seconds.\n",
      "working on slice 139\n",
      "Loading z00000139.png\n",
      "Generated 328 split error patches and 328  correct patches in 493.992324114 seconds.\n",
      "working on slice 140\n",
      "Loading z00000140.png\n",
      "Generated 296 split error patches and 296  correct patches in 493.216361046 seconds.\n",
      "working on slice 141\n",
      "Loading z00000141.png\n",
      "Generated 332 split error patches and 332  correct patches in 512.92335701 seconds.\n",
      "working on slice 142\n",
      "Loading z00000142.png\n",
      "Generated 260 split error patches and 260  correct patches in 469.272783995 seconds.\n",
      "working on slice 143\n",
      "Loading z00000143.png\n",
      "Generated 428 split error patches and 428  correct patches in 517.341932058 seconds.\n",
      "working on slice 144\n",
      "Loading z00000144.png\n",
      "Generated 456 split error patches and 456  correct patches in 509.360266924 seconds.\n",
      "working on slice 145\n",
      "Loading z00000145.png\n",
      "Generated 472 split error patches and 472  correct patches in 513.181310892 seconds.\n",
      "working on slice 146\n",
      "Loading z00000146.png\n",
      "Generated 312 split error patches and 312  correct patches in 502.545536041 seconds.\n",
      "working on slice 147\n",
      "Loading z00000147.png\n",
      "Generated 276 split error patches and 276  correct patches in 478.890866041 seconds.\n",
      "working on slice 148\n",
      "Loading z00000148.png\n",
      "Generated 312 split error patches and 312  correct patches in 511.757537127 seconds.\n",
      "working on slice 149\n",
      "Loading z00000149.png\n",
      "Generated 384 split error patches and 384  correct patches in 535.00584507 seconds.\n",
      "working on slice 150\n",
      "Loading z00000150.png\n",
      "Generated 348 split error patches and 348  correct patches in 510.416869164 seconds.\n",
      "working on slice 151\n",
      "Loading z00000151.png\n",
      "Generated 356 split error patches and 356  correct patches in 511.164813042 seconds.\n",
      "working on slice 152\n",
      "Loading z00000152.png\n",
      "Generated 340 split error patches and 340  correct patches in 525.044653177 seconds.\n",
      "working on slice 153\n",
      "Loading z00000153.png\n",
      "Generated 360 split error patches and 360  correct patches in 506.439213037 seconds.\n",
      "working on slice 154\n",
      "Loading z00000154.png\n",
      "Generated 352 split error patches and 352  correct patches in 508.421803951 seconds.\n",
      "working on slice 155\n",
      "Loading z00000155.png\n",
      "Generated 424 split error patches and 424  correct patches in 513.134411097 seconds.\n",
      "working on slice 156\n",
      "Loading z00000156.png\n",
      "Generated 388 split error patches and 388  correct patches in 507.169883966 seconds.\n",
      "working on slice 157\n",
      "Loading z00000157.png\n",
      "Generated 352 split error patches and 352  correct patches in 524.854075909 seconds.\n",
      "working on slice 158\n",
      "Loading z00000158.png\n",
      "Generated 392 split error patches and 392  correct patches in 523.156852961 seconds.\n",
      "working on slice 159\n",
      "Loading z00000159.png\n",
      "Generated 484 split error patches and 484  correct patches in 502.590487957 seconds.\n",
      "working on slice 160\n",
      "Loading z00000160.png\n",
      "Generated 356 split error patches and 356  correct patches in 518.868691921 seconds.\n",
      "working on slice 161\n",
      "Loading z00000161.png\n",
      "Generated 368 split error patches and 368  correct patches in 509.3326509 seconds.\n",
      "working on slice 162\n",
      "Loading z00000162.png\n",
      "Generated 408 split error patches and 408  correct patches in 509.870839834 seconds.\n",
      "working on slice 163\n",
      "Loading z00000163.png\n",
      "Generated 312 split error patches and 312  correct patches in 531.736149073 seconds.\n",
      "working on slice 164\n",
      "Loading z00000164.png\n",
      "Generated 348 split error patches and 348  correct patches in 513.256715059 seconds.\n",
      "working on slice 165\n",
      "Loading z00000165.png\n",
      "Generated 432 split error patches and 432  correct patches in 531.631705999 seconds.\n",
      "working on slice 166\n",
      "Loading z00000166.png\n",
      "Generated 452 split error patches and 452  correct patches in 525.471704006 seconds.\n",
      "working on slice 167\n",
      "Loading z00000167.png\n",
      "Generated 412 split error patches and 412  correct patches in 543.708986998 seconds.\n",
      "working on slice 168\n",
      "Loading z00000168.png\n",
      "Generated 540 split error patches and 540  correct patches in 542.516698837 seconds.\n",
      "working on slice 169\n",
      "Loading z00000169.png\n",
      "Generated 528 split error patches and 528  correct patches in 542.870447159 seconds.\n",
      "working on slice 170\n",
      "Loading z00000170.png\n",
      "Generated 496 split error patches and 496  correct patches in 542.668106079 seconds.\n",
      "working on slice 171\n",
      "Loading z00000171.png\n",
      "Generated 564 split error patches and 564  correct patches in 531.310811043 seconds.\n",
      "working on slice 172\n",
      "Loading z00000172.png\n",
      "Generated 588 split error patches and 588  correct patches in 516.371713161 seconds.\n",
      "working on slice 173\n",
      "Loading z00000173.png\n",
      "Generated 728 split error patches and 728  correct patches in 578.415820122 seconds.\n",
      "working on slice 174\n",
      "Loading z00000174.png\n",
      "Generated 584 split error patches and 584  correct patches in 551.354780197 seconds.\n",
      "working on slice 175\n",
      "Loading z00000175.png\n",
      "Generated 568 split error patches and 568  correct patches in 572.390740871 seconds.\n",
      "working on slice 176\n",
      "Loading z00000176.png\n",
      "Generated 560 split error patches and 560  correct patches in 559.132805824 seconds.\n",
      "working on slice 177\n",
      "Loading z00000177.png\n",
      "Generated 676 split error patches and 676  correct patches in 583.40300107 seconds.\n",
      "working on slice 178\n",
      "Loading z00000178.png\n",
      "Generated 644 split error patches and 644  correct patches in 551.443649054 seconds.\n",
      "working on slice 179\n",
      "Loading z00000179.png\n",
      "Generated 692 split error patches and 692  correct patches in 561.365392923 seconds.\n",
      "working on slice 180\n",
      "Loading z00000180.png\n",
      "Generated 676 split error patches and 676  correct patches in 575.097265959 seconds.\n",
      "working on slice 181\n",
      "Loading z00000181.png\n",
      "Generated 524 split error patches and 524  correct patches in 552.039777994 seconds.\n",
      "working on slice 182\n",
      "Loading z00000182.png\n",
      "Generated 524 split error patches and 524  correct patches in 553.965954065 seconds.\n",
      "working on slice 183\n",
      "Loading z00000183.png\n",
      "Generated 456 split error patches and 456  correct patches in 542.458884954 seconds.\n",
      "working on slice 184\n",
      "Loading z00000184.png\n",
      "Generated 492 split error patches and 492  correct patches in 551.351524115 seconds.\n",
      "working on slice 185\n",
      "Loading z00000185.png\n",
      "Generated 616 split error patches and 616  correct patches in 581.156431913 seconds.\n",
      "working on slice 186\n",
      "Loading z00000186.png\n",
      "Generated 600 split error patches and 600  correct patches in 552.196703911 seconds.\n",
      "working on slice 187\n",
      "Loading z00000187.png\n",
      "Generated 844 split error patches and 844  correct patches in 625.39528513 seconds.\n",
      "working on slice 188\n",
      "Loading z00000188.png\n",
      "Generated 1020 split error patches and 1020  correct patches in 767.94753623 seconds.\n",
      "working on slice 189\n",
      "Loading z00000189.png\n",
      "Generated 972 split error patches and 972  correct patches in 1080.00082994 seconds.\n",
      "working on slice 190\n",
      "Loading z00000190.png\n",
      "Generated 788 split error patches and 788  correct patches in 973.957851171 seconds.\n",
      "working on slice 191\n",
      "Loading z00000191.png\n",
      "Generated 728 split error patches and 728  correct patches in 940.250479937 seconds.\n",
      "working on slice 192\n",
      "Loading z00000192.png\n",
      "Generated 660 split error patches and 660  correct patches in 891.353976965 seconds.\n",
      "working on slice 193\n",
      "Loading z00000193.png\n",
      "Generated 932 split error patches and 932  correct patches in 1109.11925197 seconds.\n",
      "working on slice 194\n",
      "Loading z00000194.png\n",
      "Generated 900 split error patches and 900  correct patches in 1171.76196218 seconds.\n",
      "working on slice 195\n",
      "Loading z00000195.png\n",
      "Generated 988 split error patches and 988  correct patches in 901.326671839 seconds.\n",
      "working on slice 196\n",
      "Loading z00000196.png\n",
      "Generated 808 split error patches and 808  correct patches in 1083.76884699 seconds.\n",
      "working on slice 197\n",
      "Loading z00000197.png\n",
      "Generated 780 split error patches and 780  correct patches in 1037.59643316 seconds.\n",
      "working on slice 198\n",
      "Loading z00000198.png\n",
      "Generated 708 split error patches and 708  correct patches in 998.431375027 seconds.\n",
      "working on slice 199\n",
      "Loading z00000199.png\n",
      "Generated 884 split error patches and 884  correct patches in 983.651901007 seconds.\n",
      "working on slice 200\n",
      "Loading z00000200.png\n",
      "Generated 928 split error patches and 928  correct patches in 960.943857908 seconds.\n",
      "working on slice 201\n",
      "Loading z00000201.png\n",
      "Generated 704 split error patches and 704  correct patches in 745.688416004 seconds.\n",
      "working on slice 202\n",
      "Loading z00000202.png\n",
      "Generated 584 split error patches and 584  correct patches in 568.871387959 seconds.\n",
      "working on slice 203\n",
      "Loading z00000203.png\n",
      "Generated 748 split error patches and 748  correct patches in 544.396362066 seconds.\n",
      "working on slice 204\n",
      "Loading z00000204.png\n",
      "Generated 640 split error patches and 640  correct patches in 534.640718937 seconds.\n",
      "working on slice 205\n",
      "Loading z00000205.png\n",
      "Generated 604 split error patches and 604  correct patches in 508.629528999 seconds.\n",
      "working on slice 206\n",
      "Loading z00000206.png\n",
      "Generated 792 split error patches and 792  correct patches in 511.279141903 seconds.\n",
      "working on slice 207\n",
      "Loading z00000207.png\n",
      "Generated 668 split error patches and 668  correct patches in 503.120079994 seconds.\n",
      "working on slice 208\n",
      "Loading z00000208.png\n",
      "Generated 848 split error patches and 848  correct patches in 519.827909946 seconds.\n",
      "working on slice 209\n",
      "Loading z00000209.png\n",
      "Generated 744 split error patches and 744  correct patches in 516.249357939 seconds.\n",
      "working on slice 210\n",
      "Loading z00000210.png\n",
      "Generated 668 split error patches and 668  correct patches in 508.327284813 seconds.\n",
      "working on slice 211\n",
      "Loading z00000211.png\n",
      "Generated 568 split error patches and 568  correct patches in 513.312182188 seconds.\n",
      "working on slice 212\n",
      "Loading z00000212.png\n",
      "Generated 728 split error patches and 728  correct patches in 510.850212097 seconds.\n",
      "working on slice 213\n",
      "Loading z00000213.png\n",
      "Generated 604 split error patches and 604  correct patches in 518.571683884 seconds.\n",
      "working on slice 214\n",
      "Loading z00000214.png\n",
      "Generated 720 split error patches and 720  correct patches in 506.796018839 seconds.\n",
      "working on slice 215\n",
      "Loading z00000215.png\n",
      "Generated 500 split error patches and 500  correct patches in 494.14139986 seconds.\n",
      "working on slice 216\n",
      "Loading z00000216.png\n",
      "Generated 784 split error patches and 784  correct patches in 548.977982044 seconds.\n",
      "working on slice 217\n",
      "Loading z00000217.png\n",
      "Generated 732 split error patches and 732  correct patches in 499.404901981 seconds.\n",
      "working on slice 218\n",
      "Loading z00000218.png\n",
      "Generated 636 split error patches and 636  correct patches in 464.317550898 seconds.\n",
      "working on slice 219\n",
      "Loading z00000219.png\n",
      "Generated 348 split error patches and 348  correct patches in 451.430663109 seconds.\n",
      "working on slice 220\n",
      "Loading z00000220.png\n",
      "Generated 364 split error patches and 364  correct patches in 471.711945057 seconds.\n",
      "working on slice 221\n",
      "Loading z00000221.png\n",
      "Generated 344 split error patches and 344  correct patches in 466.412420988 seconds.\n",
      "working on slice 222\n",
      "Loading z00000222.png\n",
      "Generated 460 split error patches and 460  correct patches in 463.806288958 seconds.\n",
      "working on slice 223\n",
      "Loading z00000223.png\n",
      "Generated 320 split error patches and 320  correct patches in 446.018151045 seconds.\n",
      "working on slice 224\n",
      "Loading z00000224.png\n",
      "Generated 320 split error patches and 320  correct patches in 453.76038003 seconds.\n",
      "working on slice 225\n",
      "Loading z00000225.png\n",
      "Generated 520 split error patches and 520  correct patches in 465.909071922 seconds.\n",
      "working on slice 226\n",
      "Loading z00000226.png\n",
      "Generated 384 split error patches and 384  correct patches in 450.557800055 seconds.\n",
      "working on slice 227\n",
      "Loading z00000227.png\n",
      "Generated 416 split error patches and 416  correct patches in 450.905888081 seconds.\n",
      "working on slice 228\n",
      "Loading z00000228.png\n",
      "Generated 300 split error patches and 300  correct patches in 439.465064049 seconds.\n",
      "working on slice 229\n",
      "Loading z00000229.png\n",
      "Generated 252 split error patches and 252  correct patches in 396.969323874 seconds.\n",
      "working on slice 230\n",
      "Loading z00000230.png\n",
      "Generated 268 split error patches and 268  correct patches in 410.194236994 seconds.\n",
      "working on slice 231\n",
      "Loading z00000231.png\n",
      "Generated 400 split error patches and 400  correct patches in 405.687110186 seconds.\n",
      "working on slice 232\n",
      "Loading z00000232.png\n",
      "Generated 296 split error patches and 296  correct patches in 373.247498035 seconds.\n",
      "working on slice 233\n",
      "Loading z00000233.png\n",
      "Generated 316 split error patches and 316  correct patches in 376.821807861 seconds.\n",
      "working on slice 234\n",
      "Loading z00000234.png\n",
      "Generated 296 split error patches and 296  correct patches in 396.151175022 seconds.\n",
      "working on slice 235\n",
      "Loading z00000235.png\n",
      "Generated 208 split error patches and 208  correct patches in 350.326941967 seconds.\n",
      "working on slice 236\n",
      "Loading z00000236.png\n",
      "Generated 220 split error patches and 220  correct patches in 384.972400188 seconds.\n",
      "working on slice 237\n",
      "Loading z00000237.png\n",
      "Generated 240 split error patches and 240  correct patches in 375.272922039 seconds.\n",
      "working on slice 238\n",
      "Loading z00000238.png\n",
      "Generated 244 split error patches and 244  correct patches in 367.718622923 seconds.\n",
      "working on slice 239\n",
      "Loading z00000239.png\n",
      "Generated 248 split error patches and 248  correct patches in 355.273553848 seconds.\n",
      "working on slice 240\n",
      "Loading z00000240.png\n",
      "Generated 168 split error patches and 168  correct patches in 341.090656996 seconds.\n",
      "working on slice 241\n",
      "Loading z00000241.png\n",
      "Generated 164 split error patches and 164  correct patches in 327.014652967 seconds.\n",
      "working on slice 242\n",
      "Loading z00000242.png\n",
      "Generated 184 split error patches and 184  correct patches in 318.411050797 seconds.\n",
      "working on slice 243\n",
      "Loading z00000243.png\n",
      "Generated 116 split error patches and 116  correct patches in 326.400463104 seconds.\n",
      "working on slice 244\n",
      "Loading z00000244.png\n",
      "Generated 156 split error patches and 156  correct patches in 324.41843009 seconds.\n",
      "working on slice 245\n",
      "Loading z00000245.png\n",
      "Generated 208 split error patches and 208  correct patches in 299.270974874 seconds.\n",
      "working on slice 246\n",
      "Loading z00000246.png\n",
      "Generated 404 split error patches and 404  correct patches in 340.226310015 seconds.\n",
      "working on slice 247\n",
      "Loading z00000247.png\n",
      "Generated 312 split error patches and 312  correct patches in 323.971551895 seconds.\n",
      "working on slice 248\n",
      "Loading z00000248.png\n",
      "Generated 260 split error patches and 260  correct patches in 298.456943989 seconds.\n",
      "working on slice 249\n",
      "Loading z00000249.png\n",
      "Generated 276 split error patches and 276  correct patches in 312.570869923 seconds.\n",
      "We have a total of 225520 patches.\n",
      "Errors: 112760\n",
      "Correct: 112760\n"
     ]
    }
   ],
   "source": [
    "run(0,250, 'train')\n",
    "run(250,300, 'test')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "working on slice 200\n",
      "Loading z00000200.png\n",
      "Generated 928 split error patches and 928  correct patches in 582.849436998 seconds.\n",
      "working on slice 201\n",
      "Loading z00000201.png\n",
      "Generated 704 split error patches and 704  correct patches in 575.221096039 seconds.\n",
      "working on slice 202\n",
      "Loading z00000202.png\n",
      "Generated 584 split error patches and 584  correct patches in 531.806138992 seconds.\n",
      "working on slice 203\n",
      "Loading z00000203.png\n",
      "Generated 748 split error patches and 748  correct patches in 560.802320957 seconds.\n",
      "working on slice 204\n",
      "Loading z00000204.png\n",
      "Generated 640 split error patches and 640  correct patches in 550.972352982 seconds.\n",
      "working on slice 205\n",
      "Loading z00000205.png\n",
      "Generated 604 split error patches and 604  correct patches in 529.127249002 seconds.\n",
      "working on slice 206\n",
      "Loading z00000206.png\n",
      "Generated 792 split error patches and 792  correct patches in 549.462425947 seconds.\n",
      "working on slice 207\n",
      "Loading z00000207.png\n",
      "Generated 668 split error patches and 668  correct patches in 530.575075865 seconds.\n",
      "working on slice 208\n",
      "Loading z00000208.png\n",
      "Generated 848 split error patches and 848  correct patches in 545.316864014 seconds.\n",
      "working on slice 209\n",
      "Loading z00000209.png\n",
      "Generated 744 split error patches and 744  correct patches in 527.447477102 seconds.\n",
      "working on slice 210\n",
      "Loading z00000210.png\n",
      "Generated 668 split error patches and 668  correct patches in 527.445876122 seconds.\n",
      "working on slice 211\n",
      "Loading z00000211.png\n",
      "Generated 568 split error patches and 568  correct patches in 517.545186043 seconds.\n",
      "working on slice 212\n",
      "Loading z00000212.png\n",
      "Generated 728 split error patches and 728  correct patches in 549.048825026 seconds.\n",
      "working on slice 213\n",
      "Loading z00000213.png\n",
      "Generated 604 split error patches and 604  correct patches in 536.001394987 seconds.\n",
      "working on slice 214\n",
      "Loading z00000214.png\n",
      "Generated 720 split error patches and 720  correct patches in 531.028899908 seconds.\n",
      "working on slice 215\n",
      "Loading z00000215.png\n",
      "Generated 500 split error patches and 500  correct patches in 526.350455046 seconds.\n",
      "working on slice 216\n",
      "Loading z00000216.png\n",
      "Generated 784 split error patches and 784  correct patches in 544.110040903 seconds.\n",
      "working on slice 217\n",
      "Loading z00000217.png\n",
      "Generated 732 split error patches and 732  correct patches in 523.687410116 seconds.\n",
      "working on slice 218\n",
      "Loading z00000218.png\n",
      "Generated 636 split error patches and 636  correct patches in 498.820012093 seconds.\n",
      "working on slice 219\n",
      "Loading z00000219.png\n",
      "Generated 348 split error patches and 348  correct patches in 470.433403015 seconds.\n",
      "working on slice 220\n",
      "Loading z00000220.png\n",
      "Generated 364 split error patches and 364  correct patches in 476.567451954 seconds.\n",
      "working on slice 221\n",
      "Loading z00000221.png\n",
      "Generated 344 split error patches and 344  correct patches in 463.371654034 seconds.\n",
      "working on slice 222\n",
      "Loading z00000222.png\n",
      "Generated 460 split error patches and 460  correct patches in 476.857083082 seconds.\n",
      "working on slice 223\n",
      "Loading z00000223.png\n",
      "Generated 320 split error patches and 320  correct patches in 461.848649979 seconds.\n",
      "working on slice 224\n",
      "Loading z00000224.png\n",
      "Generated 320 split error patches and 320  correct patches in 450.647064924 seconds.\n",
      "working on slice 225\n",
      "Loading z00000225.png\n",
      "Generated 520 split error patches and 520  correct patches in 471.250283003 seconds.\n",
      "working on slice 226\n",
      "Loading z00000226.png\n",
      "Generated 384 split error patches and 384  correct patches in 448.770314932 seconds.\n",
      "working on slice 227\n",
      "Loading z00000227.png\n",
      "Generated 416 split error patches and 416  correct patches in 450.287806988 seconds.\n",
      "working on slice 228\n",
      "Loading z00000228.png\n",
      "Generated 300 split error patches and 300  correct patches in 438.957047939 seconds.\n",
      "working on slice 229\n",
      "Loading z00000229.png\n",
      "Generated 252 split error patches and 252  correct patches in 409.083827019 seconds.\n",
      "working on slice 230\n",
      "Loading z00000230.png\n",
      "Generated 268 split error patches and 268  correct patches in 407.127774 seconds.\n",
      "working on slice 231\n",
      "Loading z00000231.png\n",
      "Generated 400 split error patches and 400  correct patches in 420.710175991 seconds.\n",
      "working on slice 232\n",
      "Loading z00000232.png\n",
      "Generated 296 split error patches and 296  correct patches in 383.633507967 seconds.\n",
      "working on slice 233\n",
      "Loading z00000233.png\n",
      "Generated 316 split error patches and 316  correct patches in 376.427021027 seconds.\n",
      "working on slice 234\n",
      "Loading z00000234.png\n",
      "Generated 296 split error patches and 296  correct patches in 398.408967018 seconds.\n",
      "working on slice 235\n",
      "Loading z00000235.png\n",
      "Generated 208 split error patches and 208  correct patches in 376.357911825 seconds.\n",
      "working on slice 236\n",
      "Loading z00000236.png\n",
      "Generated 220 split error patches and 220  correct patches in 383.599498987 seconds.\n",
      "working on slice 237\n",
      "Loading z00000237.png\n",
      "Generated 240 split error patches and 240  correct patches in 370.776157141 seconds.\n",
      "working on slice 238\n",
      "Loading z00000238.png\n",
      "Generated 244 split error patches and 244  correct patches in 373.844875097 seconds.\n",
      "working on slice 239\n",
      "Loading z00000239.png\n",
      "Generated 248 split error patches and 248  correct patches in 367.531090975 seconds.\n",
      "working on slice 240\n",
      "Loading z00000240.png\n",
      "Generated 168 split error patches and 168  correct patches in 331.703675032 seconds.\n",
      "working on slice 241\n",
      "Loading z00000241.png\n",
      "Generated 164 split error patches and 164  correct patches in 332.139189959 seconds.\n",
      "working on slice 242\n",
      "Loading z00000242.png\n",
      "Generated 184 split error patches and 184  correct patches in 326.20072794 seconds.\n",
      "working on slice 243\n",
      "Loading z00000243.png\n",
      "Generated 116 split error patches and 116  correct patches in 317.571787119 seconds.\n",
      "working on slice 244\n",
      "Loading z00000244.png\n",
      "Generated 156 split error patches and 156  correct patches in 312.1851542 seconds.\n",
      "working on slice 245\n",
      "Loading z00000245.png\n",
      "Generated 208 split error patches and 208  correct patches in 306.746034861 seconds.\n",
      "working on slice 246\n",
      "Loading z00000246.png\n",
      "Generated 404 split error patches and 404  correct patches in 330.435755968 seconds.\n",
      "working on slice 247\n",
      "Loading z00000247.png\n",
      "Generated 312 split error patches and 312  correct patches in 312.728674889 seconds.\n",
      "working on slice 248\n",
      "Loading z00000248.png\n",
      "Generated 260 split error patches and 260  correct patches in 297.399206161 seconds.\n",
      "working on slice 249\n",
      "Loading z00000249.png\n",
      "Generated 276 split error patches and 276  correct patches in 299.098962069 seconds.\n",
      "We have a total of 44424 patches.\n",
      "Errors: 22212\n",
      "Correct: 22212\n",
      "starting loop\n",
      "unshuffled saved\n",
      "saving..\n",
      "Done!\n",
      "working on slice 250\n",
      "Loading z00000250.png\n",
      "Generated 296 split error patches and 296  correct patches in 301.48516798 seconds.\n",
      "working on slice 251\n",
      "Loading z00000251.png\n",
      "Generated 280 split error patches and 280  correct patches in 298.132526159 seconds.\n",
      "working on slice 252\n",
      "Loading z00000252.png\n",
      "Generated 340 split error patches and 340  correct patches in 296.013477087 seconds.\n",
      "working on slice 253\n",
      "Loading z00000253.png\n",
      "Generated 212 split error patches and 212  correct patches in 275.51947403 seconds.\n",
      "working on slice 254\n",
      "Loading z00000254.png\n",
      "Generated 332 split error patches and 332  correct patches in 280.632390022 seconds.\n",
      "working on slice 255\n",
      "Loading z00000255.png\n",
      "Generated 300 split error patches and 300  correct patches in 281.837744951 seconds.\n",
      "working on slice 256\n",
      "Loading z00000256.png\n",
      "Generated 436 split error patches and 436  correct patches in 283.624251842 seconds.\n",
      "working on slice 257\n",
      "Loading z00000257.png\n",
      "Generated 420 split error patches and 420  correct patches in 282.023189068 seconds.\n",
      "working on slice 258\n",
      "Loading z00000258.png\n",
      "Generated 492 split error patches and 492  correct patches in 280.264985085 seconds.\n",
      "working on slice 259\n",
      "Loading z00000259.png\n",
      "Generated 448 split error patches and 448  correct patches in 274.082119942 seconds.\n",
      "We have a total of 7112 patches.\n",
      "Errors: 3556\n",
      "Correct: 3556\n",
      "starting loop\n",
      "unshuffled saved\n",
      "saving..\n",
      "Done!\n"
     ]
    }
   ],
   "source": [
    "run(200,250, 'train')\n",
    "run(250,260, 'test')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded /home/d/patches//cylinder2_rgba_small/ in 0.000495910644531 seconds.\n"
     ]
    }
   ],
   "source": [
    "X_train, y_train, X_test, y_test = mlp.Patch.load_rgba('cylinder2_rgba_small')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAABpCAYAAADyfL7tAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJztvWuMnPdZNn49c3rmmZlnzoc979prO3Y2cRLSWCEkQhUE\nVbSFNm3RWygSIJAqVAECFAkq+ISgb5FQQf9PKX+QoEhvVPohEtBXRZFKCyiqW3Ks7TiHXa/3NLNz\nfJ45H98P7nX7txMncZ21d8f7u6SV7d2d2WeeWd/X777v675uYzQaQUNDQ0NDYz/gOegL0NDQ0NC4\ne6BJRUNDQ0Nj36BJRUNDQ0Nj36BJRUNDQ0Nj36BJRUNDQ0Nj36BJRUNDQ0Nj36BJRUNDQ0Nj36BJ\nRUNDQ0Nj36BJRUNDQ0Nj36BJRUNDQ0Nj3+A76As4CHz7298eNZtN+Hw+dLtdxONxhEIhNBoNeL1e\n1Go19Ho95PN5FItF1Ot1OI6DbrcLr9cLr9cLn8+H+fl5LC0tYW5uDqFQCB6PB8PhEF6vF/1+H4PB\nAI7joFKpIBAIYGpqCj6fD/V6HQBgGAYAYDgcwuPxoNvtot1uYzAYwOv1IhAIIB6Pw7IsDIdDjEYj\n2LYNy7LQ6XTQ6XRgmiZ8Pp/8vMFgIK/Tdd09r6/T6WA4HKLf76Pb7aLb7cIwDAwGAxiGgVarhdXV\nVVy+fBndbhdf/vKXjVu5v4ZhjMb+DY/HA5/Ph2AwiEwmg9OnT+ORRx7BT/zET+DYsWNIpVIIhULw\ner3o9Xqo1+vI5/N4+eWX8V//9V/4n//5H1y9ehX1eh29Xg/D4fCW3//9hvra4vE4FhYWcOrUKdxz\nzz04fvw45ufnkc1mEY/HYZom/H4//H7/Ld1bALfsq/S9730Pjz322J7fkbsVo9Hox76/47+3GjfG\n+93bI0kqrVYLxWIRmUwGrVYLgUAAo9EIXq8XrVZLgm48Hkc0GkW9XsfW1hZef/11CeLtdhuvvvoq\nVldXJXDMzc0hEomg1WoJCViWhWAwCK/Xi2AwiH6/D9M0AVwLRv1+H81mE61WC81mE41GA5ZlIR6P\nwzAM+fD7/QCuEdBwOJTntG0b/X4f7XYbfr8fo9EIo9EIjUYD4XAY0WgUo9EIHo8HlmXBdV2USiUh\nMMdxUK1W4fV6Ua1WUSwWUalUYFnWvt3v0WiE4XAoZJDP5zEYDNDr9eTjxIkTSKfTCIfD8Pv9iMVi\nDL4IhUKIRqM4f/48VldXUa1WhSAP0rvO4/HA6/XC7/cjHA4jm83i1KlTeOihh3DfffdhcXERqVQK\nkUgEwWAQfr8fXq9XDhN3GufOncN//ud/4qd/+qfR7XYP5Bo07n4cSVLpdDoIh8OoVqvo9/vI5XIY\nDocIBALw+Xxy8vR4PGg2mwCAWCyGlZUVXLp0Ce12G51ORwJzKBRCMBhEvV5HMBhEKBRCKBQSAup2\nu0gkErBtGwAk6A+HQ1iWBdM0MRwO0e12EQgEYBiGPI/P55PPkRx8Ph9GoxE6nY5kNcx0iHA4jNFo\nhHa7Ddd10W63YRgGarUaut0ums0marUayuUyhsMhDMOA67oYDodIJBKSTe0XSHaj0QjNZhM7Ozto\nt9uo1+uoVCpwHAf33XcfcrkcbNsWMpmamoLf70ckEkEqlcL58+dx6dIl5PN5NJtN9Pv9O04shmEI\nmYRCISSTSSwuLuLBBx/EQw89hBMnTiCbzSISicA0TQQCAXmPDopQiEcffRTPP/88nnzySbTb7QO9\nFo27E0eSVK5cuQLTNDEajWCaJlqtFgAIgQwGA3S7XTkxJxIJBINB5PN5dLtdVCoVbG5uSglmNBqh\nXC7DNE3EYjF5fDQahWVZckKtVqswTROWZcG2bfR6PRiGgUgkAsMw4PP5EIlEUKvVUCwWkc1mkUgk\nYFmWnPZ9Ph/8fj+63a4QCcmLmdBoNEK325XXNRgM4Loutra24Lou+v0+Go0G+v0+AoEAer0eLMuS\n6wiFQvLY/QZfR6fTQblcRqfTQbVaRbVaRavVwgMPPIDp6WlEo1EEAgFYloVMJgO/3w/btpFMJhGP\nx/Hiiy9iY2NDXsedKIepZbxQKIRUKoWlpSWcPXsWDzzwAE6cOIFcLodIJCKHBZVM1HLnQZLL448/\njm9961v42Mc+JpmyhsZ+4UiSSjgcRqPRgN/vh2macmJmQGWAYpBnHyUcDuPee+/Fzs4ObNuWoOY4\nDnq9HtrtNkzTxMzMDGZmZpBIJKSk1mq1YJomTNNEr9dDLBZDIBCQPojf75cTLQDpzYxGIwlI7K00\nm00JzuyRBINBKWfxZxQKBbTbbZRKJRSLRTiOAwCIRCIIh8MIBoMYDAZCrMViEaZpSsZ1u8CMpdvt\nYjAYoNPpoNFoCLk8/PDDWFxcRDKZRDAYRDAYRDKZlKwtHA7DNE384Ac/wPr6OlzXRa/Xk+fdb/D+\n+3w+mKaJZDKJY8eO4YEHHsDZs2dx8uRJ5HI5xGIxWJb1jsxEJZTDgieeeAK1Wg3PP/88PvGJT+x7\nZqpxdHEkSYVN6k6nA6/XC8dxpEHcarVg2zYGgwHi8Tg8Hg/C4bAEfp/Ph3A4jGQyKUQBAL1eD4FA\nAMPhUE75DHRer1fKWwCQSCTg8Xjk68xsGo0G6vW6NO79fj+azeae3slgMJBeTCAQkBIZyc3r9UrJ\na21tDWtra+j1enBdF51OBx6PB6dOnZLMhK97e3tbmuDBYBCxWOy2vw987e12G/l8XgiUPa3l5WWk\nUikEg0ERLXg8HumBAdcC/tWrV4XY97vPwuwkEAggEolgenoa999/Px599FGsrKxgenr6HWRyIwLh\na+PfDwt+5md+Bt/4xjfwmc98Rg4dGhofBEeSVAqFgpSlOp0O/H4/gsHgnnKSbdsIhUISyNl8r9Vq\nomAKhUJYX18XEmEwZqArlUqIx+OYnp6G3++XBnswGJSTerfb3VPa8ng8EtxZPvF6veh0OtJLYbOf\nmUq73Uaz2US320W9Xke9XsfVq1dx6dIl9Pt9ISEA8Pv9aDQacnpm5uM4jpTpbNuGx3Nn1OYklk6n\ng0KhgBdffBHVahWFQgGPPfYYVlZWkMlkYFmWXJthGFhZWZGS3/nz57G2toZqtSrEvR+Bm414ZkrL\ny8v40Ic+hEceeQTLy8tIp9OIRCJy3/je8N7y8cxYDhOZqPi5n/s5PPvss/jsZz+LarV60JejMeE4\nkqRCaXAgEIBpmpidnZVyD4NrMBhEp9NBr9eThj4lvbFYDJFIBMlkEpFIREoww+EQg8EAPp8PnU4H\n7XZbspF0Oi3ZBk+Eo9EIgUAAnU4HjuPAcRw0m02RCvv9fiEEqqUoAW61WnvEBIVCAc1mE9VqFVeu\nXMHa2hoajYYoj+LxuGRh0WgUkUgEHo9HSn+pVEr6SIZh3PE6O/ssu7u7aDQaKJfLqFQqaDQaePDB\nBzE1NYVwOAyfz4doNAqv1ytNcAZzSrh7vZ7c31sFCSEUCiGXy+Hs2bN49NFHcf/992NhYUFKcyR6\nZr48JLBUFggE3pHBHEZy+chHPoJ/+qd/wq/+6q+iXC4f9OVoTDCOJKlwZqPb7YpKh5kJM4p2u412\nuy3ZCxvjoVBIpK/1eh2zs7MwDAObm5soFApwXReWZcHj8cBxnD2Bnf0Vyn+Z1aglMI/HI0HTMAx0\nOh0pnZBsAMiswXA4RLFYRLlcxtWrV1Eul9HtdkWR1m63hUAXFxeRyWQwMzMDn88H13VRLpcRiURQ\nKpUwHA5RrVYPrAdA2TFLd81mE/V6Ha1WCw8//DBmZmYQjUZFDca5oE6nI5mk2mMBbi2AU90VDAaR\ny+Xw0EMP4bHHHsP999+P2dlZEU+wVNdsNlGpVLCzs4NyuYxerwfbtpHNZpFMJhGLxaRMyYzmMOLn\nf/7n8Q//8A/4tV/7NRSLxYO+HI0JxZEklVwuh2q1isXFRQnaDNbRaBTdbldKKVRxkQB8Ph8GgwH6\n/T7C4TCGwyGmp6cRCoVgWZYMS3L2JBQKwXEc7OzsIB6PS2AhaXAWo9/vo9PpoNlsYjAY7Ol5MDPp\ndDpwXRemaSIUCqHdbqNYLKJUKkmTu9PpIBQKwbZtFAoFyXjS6TQymQyWlpaQTqcxGAyk/1Ov14VA\n4/E4OBh6EBjvs/zgBz9Ar9dDp9PBhz70ISwsLCAej4ucN5fLSRbX7/fR7/exubkpKjc+582AREoS\nTiaTuOeee3Du3Dncd999Qihq+dJ1XVy9ehUXLlzAa6+9hvX1dXQ6HWSzWdx77704e/Ys5ubmZLiT\nZdbDio9+9KP4u7/7O/zmb/4mCoXCQV+OxgTiSJJKLBbDsWPH0O/3USqVUK/Xkc1mJZCORiP4/X7U\najUpZ4VCIQCQWRa/3y/NeUpxgWsBidJjn88nJSpO5y8sLCCRSAgxtVotIZdQKCRBlSW0YDAo8yXM\nIkKhEMLhsMyW7OzsYHV1FYZhYHp6GpFIBK7rShBbXl7G6dOnYds2TNOEYRgiPmi329je3kYsFkOt\nVgMAmdg/SDBbJLE0m000m0385E/+JI4fP45EIiFDh9PT0yIr5p+DwQCNRgODweCmSGWcUKLRKI4d\nO4YHH3wQp0+fxvT0tEzEc/6nWCzijTfewAsvvIDz58/vKYOGQiG89dZbWF1dxSOPPIJ77rkHuVwO\nqVRqXwdLbwc+/vGP45lnnsHnP/957OzsHPTlaEwYjiSpzM7OIplMSv1+NBqhWq0im81KOaPX66HR\naMhkNxVZVCI5joN2u41+vw+fzwfbtuHz+eT7SBQ8OZOIKCsOBAKoVqsIhUKIRCKIRqMIBoOoVCoy\nlEZ5MQNlvV6X2Q2W0SgF5dxLr9eTk3AikUA2m8Xx48eRTqcRCoXQbDZl1sKyLMRiMczOzqJSqaDX\n6wlZsnx0kBiNRuj1eiiVSnj11VdF2cXyXDweF1Kfm5sT4QHLaFtbW2i1Wu8rNSahqDMoU1NTOHXq\nFE6dOoWpqSlEo1EhlFarhZ2dHbz22mv4zne+g/Pnz++ZmQEgh4Dd3V2USiVUKhU89NBDMAwD6XT6\n0BPLL/7iL2I0GuG3f/u3sb29fdCXozFBOJKkkk6nRaUTiURk8rnVaolfFifNB4MBms0mstksYrGY\nqKUY5B3HkXkWAFJ64gxGsViUGRb2MSqVCubm5rC4uCj1djb02eBl2YplFmYXVBq1Wi3s7u5id3cX\n/X4ftm3Dtm1p+qdSKSQSCUSjUWQyGXi9XhEoUPXG0h4/NxgMEAqF4Lou8vn8Ab9L10BiqVaruHDh\ngqiqPB4Pjh8/LqWwaDQKwzDQ6/WkXMhM52Zkxnxe9roWFxdxzz33YH5+XnoowDWLn+3tbbz00kt4\n/vnncf78eWxvb++x5hmXRlPizQMIhROHHZ/4xCcwHA7xO7/zO9jc3Dzoy9GYEBxJUmEZyu/3o1gs\nwrIsMZFkCUiV4HLOhOUUznbQ6oQ1/UQiIaUTBnF6erEf0mq10Ov14PV6EYvFREHGXg3JiT9nOBzC\nNE0Eg0GEw2F4vV40m00Ui0Xs7OyIoovT9BxcpEqNdi0MtuwLUXlGCxcAyGQyctpmj+mwoN/vo1qt\n4uLFi5Jl8b2JxWKSLc7OzmIwGKBaraJUKkkwv5nBSPqjZbNZLC4uygCranRZqVRw8eJF/Md//AfO\nnz+Pzc1NtNvtdxCX2hsqFApSEqNLwtzc3O2+ZfuCp556CqPRCL/3e7+HjY2Ng74cjQnAkSQVzm5Q\nikoPrsFgIGUkSlXD4TBSqRSAa3JjSk0Z9BnILcsS7y5KXwOBAILBIFZXV1GpVGRYksFpY2ND+jiJ\nRAKmaYqfFeXEqv0IiadWq0m5JRqNwrZtabYbhiHlGroXu64r6rNAICBS5+FwKNkQy3Ms2x2G8tc4\n+v0+KpUKLl26hGg0Kv5qLIFRbpzNZrGysoL19XUUCoU9WcS7gaUv27YxMzODpaUlzMzMiPMBy14b\nGxv4/ve/jxdffBHb29viKv1uUKXSL7/8MjKZDObn52/H7blt+NSnPoWdnR184QtfOOhL0ZgAHElS\n+da3voV+v490Oi3T6Dz9srFLWS7VWixPAdc8whiYVX8nj8cjtitswM/MzGA4HIrMlX2Vfr+PWq0G\nx3FQr9fFhdh1XZnBoLTVMAx0u13pdXDQkWQYjUZRLpexvb0tJ2vLsvao0aLRKJLJJPr9vpgxhkIh\nZDIZeQylznQTOGxQeyw//OEPYVmWiBZYVmKmmMvlcObMGbz99tsolUrvauPCfophGDK1PzU1hUwm\nI4154NpsU6lUwoULF/Dyyy9jc3MTrVbrpmzked27u7t47bXX8OCDD+LJJ5/c/xt0G5HL5bC4uIgr\nV64c9KVoHHIcSVL52te+hp2dHTzxxBNYXFyUzCQQCEjQJzlQpcXyEevlbJJ7PB6x6VB7FVRzUYHV\n6XRw+fJlFItFeDweTE9PS++lUChgNBqJ7xgzB/ZOVI8sltAASDZEu/5EIoFTp04hkUgAuB4wAexR\nRrGBnUwmMTMzA9M0EQ6HZdiTPYzDCBL+9vY2XnnlFcTjcZkb4j0j0c7Pz+PMmTNYX1+XgwCfQwWz\nT8uykEqlMDU1hWQyKYcKOiuvr6/jpZdewltvvQXXdX+svSSUb29sbODSpUv7ek/uBD796U8DAJ5+\n+mmsrq4e8NVoHGYcSVLZ2trCN7/5TQDXXXOTyaTYbnA6mztJKA9mcGejPplMijKMA4/sf3DeBABM\n05SZmF6vJ6fqYDAoAdJ1XczMzMikNkmAfmQsXVFMQNkzp58HgwGmp6eRy+VkoJPOABzko88ZSY7K\nM071s7mcSCQOLakA1wP09vY2Ll68iIWFBVFUkQhs20Y6ncbJkydx6dIlyVbGiUA1i6TzcDKZFJdk\n3pNyuYwLFy7gwoULKJVKt2S5PxwO4TgOLl++vJ+3447h05/+NNbX1/EHf/AHB30pGocYR5JU5ubm\nUC6X8cILL6BYLGJ+fh733nuvqMDomuvxeFCtViU4M5h1u12R4wKQvki/35etkAzcHKL0er0yT1Gp\nVKShzmZuo9EAgD0+ZCQVANLHoXEkAMl0+v0+QqGQqL24GIrSWm6LbDabct3q5D4l1OynGIaBXC53\nAO/MzYG9sGazic3NTVy+fBnHjx9HJpORbJOLzqamprC8vIwrV66IWeeNSmB0oeY95E4XAGi329jY\n2MBrr72Gq1evSmP+Vq672+1O9OzH/Pw8lpeX8dZbbx30pWgcUhxJUvnZn/1ZfPe738Xm5iY2Nzdx\n9uxZmRXhnAe9nOhkTFlpIpEQt99wOLznxEo1EgA5/aumgpQJBwIBANebuOzpkLSazab0Ubg8jKaP\nDGgc3FRJjHtaGHT5b25JpKFlJpMR+3g25Tkjw3LPYeypqGCfolKp4O2335YNnCQDyoyTySQWFhaQ\nyWSwtbW1R4Cg2tIzm4tEIiIAoPrOcRysrq7i7bff/rHLXuPXzN02k4rPfOYzWFtbw9NPP33Ql6Jx\nSHF4/SJuI1KpFM6ePYvp6WkAwMbGBt544w2sr6+jXC7LPAGHCNUdJiwTcdCOYFYAQLY2cvqekmNa\no7TbbRnK42wFVURcP8zPk2B4PbRhZ++Dklr2UTiQqZoZBgIByXpoL5NOp5FKpcSlmcTyXvbthw1c\nNpbP53HlyhVUKhUhUMqDOaczPT0t4odxbzOSikrOPp9PSl806czn82Lro6GhcWMcSVLpdrtIpVJY\nWFiQtcL5fB7r6+vI5/PS+A4Gg7IGmNkLVV/jz8dZFM6C0Mo+EAjsKWWRHBzHkeFDltxojUIiYg/G\ncRwJ/mrWw6FFXhfJg+IC+okBkKDJa2RPh0qyarWK7e1tUYtNglMty0mVSgXb29solUqiyCIRc/fN\nzMyM9L3GMb6Ei6QCQJ4/n8+LBcutggOW0Wj0lp/jMIC2PxoaN8KRLH+98soroqAKhULweDzS/yiV\nSsjlchLsuRHS7/fDdV1RGHF2gad89lOoAOM0Pt1sucmRJSfg+mbJWCwmj+n1emKfzkVe3PHCk3et\nVhOFGMtaNKCktxRFBZyv4M+m7Ller8uMDbOwarUqMx+3WuK506BwYmdnR+z/1UzTNE1Eo1ERYRSL\nxT0lSzVj4awKiRuAzPnQrPODkoplWZidnf1gL/qA8dRTT+H111/HH//xHx/0pWgcQhxJUrEsC67r\nSl8kGAyKrxN3jHAtb6PRgGmaEugByBZHWtFTTqxa1Hc6HQluqoFjqVQSO3xOvFNlRGJj34XDiSSP\nWq2GWq22p3RGKTJXF7NR7fF40Gg00Gg0JMshUTabTTQaDVQqlXcMRrqui2azKcR32MGSZLlcxtbW\nlswDsW9FaTdtbHw+3x4VGN9TlhbV0hh7U1wH/UGIlplQIpHAsWPHPuCrPnicPn0an/rUpz7Qc/z3\nf/+39hVTcPLkSQSDQbz66qsHfSkfCEeSVM6dO4dKpYKtrS2ZKbEsC9PT06LQomzU7/eLbTnLUSw5\ncSqfmQxJiIGf+98pM1Z/ViQSkW2G6gmZdXyqvVT/KGYX0WhUZk5IdnyMuoMFgFjbx2IxGQB0XRfb\n29tYX1+Hz+fDwsKCLCmjxT7tag472PxuNBooFouo1WrodDri1cU+CeXTfr9fnAfU3og6BMkshXJz\n1cjyVmEYBoLBIObm5rC8vPwBXvHhwCc/+Ul88pOf/EDP8fGPfxz/8i//sk9XNPn42Mc+hg9/+MP4\n4he/ONHEciRJhTs5stmsBOF4PC6lqkKhIME6Go0ikUjscceloouNeGYltFVRm++qyWCr1ZIFU3yc\n6oBMixdOxDO4c2aEe1ToTMyGPZ+PZKc25VnaYvmNH6prAL3CGo2GBOBJkr2qvZVSqSSiB3UdcCQS\nEfNJlUD4oZbDqOBTsxj2s25lLTAPHdzPcuLEiX189Rp3Cy5duoTPfe5z+LM/+zM8++yzAK6V6l97\n7bUDvrIfD0eSVKamphCJRKTXwNP/aDTC9vb2npW+nU4H+XwePp8PyWQS4XBYhuzYTFe9xDidbhgG\nqtWqTMCz+U5bdTbHB4MBAoGA7FLhYi/KiGliyTIZA2ij0YDX65XZGpbfuAWRYgKqwbiAi8HNcRws\nLS3J62EPgtJizsJMCjj/UywWZe6GPSQSOdcYqDvj+adq4cKvq5P20WhUZn1+nL4Ky17RaBQnT57E\nww8/jKWlpdt0FzQmGd/85jexsrKCv/zLv8Qv/MIvAAD+5E/+RJPKJMDn8yGTycicw/jJFMAeWW+p\nVBKJajwel/4FAJELq7vIAUhZi30U9m84R0Li4nONRiNpmqfTabTbbZkCj0ajCIfDcm2s/zML4qmc\nZaB2uy2NfPZjWE4DrvUZZmZmEI/HhZwqlQqazSa8Xq8MbE4K1PmPcrkspKqSCu3mOVg6btbJ5yDR\n8HtM00Q8Hkcul0M8HhcF2PhjbwQq7mKxGE6dOoXHH38cZ8+eRTqdvu33REPjoHAkJcVsrLPUxJII\nCWI0GskEvbpTvNFoyL4Vfg2ANPZZZuLnaNDIpjtLKxx6ZO+F4K4VNoZV4YAa5GlWyT4I1WXMSMLh\nsGyj5Ck5kUhIIO31erLamE7GxWIRrutia2sLlUrl0A8/qlBt5h3HEQUYcL3ERaNJ9rHUhrxa1rzR\nQGQmk8Hx48cxPz8vw5X8Xj5eBR9rmiYymQxWVlbw4Q9/GI8//jjm5+dlvYGGxt2II5mpcMcId1yY\npikDiexRsBTEvSPANaLgWluWs3hqbbfbKJfLUiKh6oolL86ssCzFJnu325USFBVbtHChcowmlTz5\n8lp4CqYyjM1+via+TpZxOBDJDIwiAM7M1Ot1VCoVdLvdibNnJ1E3m015H9Wsgy7U47MqKokwOwGu\ny4u5qfHMmTMoFAqo1+tYW1uD4ziyPmC84c/3IJvN4syZMzh37hzOnTuH5eVlsdLX0LhbcSRJpdls\nysIry7JkMr1Wq4l9SaVSkZM9Awe3PVL2yxMvre77/T6SySQA7CGdSCQiDXQKABjs+aca8NSv0woG\nuH4iZ0+FDsncF0KFGEs9qVRqT8+Ag5AMvHwtDJDcycJtlJMEknS73RaVHrNDEgSFDuOOASqxjM+t\neL1eRKNRLC8vy+/Jiy++iLW1tXfMrlBsEY1GMTc3h5WVFTz00EM4c+YMZmdnZYkbM9yjjo9+9KN4\n5ZVXsL6+ftCXorGPOJKkEg6HhSzo0dVutyX4MAjT+p5BgFlDsViUwUVmI2oGoAYOmjnW63XZMU/V\nl2qrz5KXz+dDpVKRwM7ym9frlfkXypmZzXDgslQqoVgsIhgMIp1Oo9FoiOkke0SWZQn5lEol1Go1\nkdh6vV4xYZwkUiHh8z6yUc/3TiUWtXxF0rnRcwHXsw7LspBOp7GysoJwOIzp6WlcunQJGxsbslmS\n9zYWi2F2dhYnT57E8vIy5ubmROBBF4RJsMC5E/j85z+P5557TpPKXYYjSSrRaHTPGl3uKaEdynA4\nlNIRAxPLYyxdqJ5cJBE2hllqYalD3YmiEgl7MyyzMIMAri+NYmZDcjFNE5FIBPV6XWZduGiMQ50c\n7OQmSvYG2AegZ1a5XJZsampqCrOzs7AsC4FAQGTLkwIaTFL9RrsaVSrs9/th27bck3EVF10QxstZ\ntMWnjDyVSmF5eRmFQgGO48jviG3bSCQSyGQyMsFPJZ+a2Wpcx1NPPYWLFy/q5V8/wve//31897vf\nxRNPPAEA+Kmf+imcO3cO3/ve9w74ym4eN0UqhmF8BMBXcK2x//+PRqP/fVuv6jaDjW4OAlar1T1N\neNVRWB2kY8Dv9/syg8KSBk+/NGU0TXNP34SlFJaYeLpls5yiAQ5IcukUS1/qNRmGIfMl7PtwAyKH\nIjmjwufz+/3yOfYcKC/2+/1YXFzE8vIyOp2ObKOcFDC7YGmQMm91Al5VgY036m/0fCoZjTfvTdNE\nLBbD4uLiHtGGZVni0EDi4se4CEDjGn7rt34LX//61zWp/Ajf/va38W//9m9CKh/5yEfwne985+4i\nFcMwPAD+PwA/A2ALwHnDMJ4bjUaTt77uR+CueFWSS9df9jcYBAKBgDTDAexRfLFcxRkPTtCzDKY2\nciklZqDQFFiaAAAdYklEQVRjLweAlGFUA8RkMim2KQxcAKTUBlyXPauZz2g0kiVT7AdwTa7f78fu\n7q5srAyHw8jlcvB4PLJSuF6vixx5ksB7y/Lg+OwJS2Cq/cz4FD3fc97rcX8wvvcsZfH95feoGYna\n9Ndk8t747Gc/i8uXL2tiuUtwM5HjHIA3RqPRFQAwDOP/APhFABNLKrQ0YU8lFArJzvpQKCT/+Tl4\nyKxCVU+xHMXZDgYTBhYGGRoeMsBw5zlJgM7FDFi0dRkOh6IEUwml0WiIOICKMA5FsuEej8eRTCbh\n8Xhg27b8m4uoXNcVTyzKWzkn0+/3JdOaJKiyYs7pkBTUoUZiXFKslgbHbVnUv6s9GtXvjb8Taolr\nXAww/jmNa/j1X/91/OM//qMmlbsEN0MqswCuKv/ewDWimVhQxcP+BwcL1dKIz+dDuVze03thVkIS\n4U55ALI+mCUWy7JEcUVXXLoJU8KsWqgwcxgOh9Kb4fez1EZJcavVkgHFXC6HTCYjz6kq2ljPN01T\nGvqcX1FLaXQtZrmOwXLSQFmx6n02bsHyXtb3bParr/1G90E9VAB7J/BvNLNyo79raNytmKwaxz6B\nTrXsc7DZzuAy3lQlcdC7i0GamQtLLiy/dDod2W3P2j57KXy8Wn5jIAuFQrKvhc3fRCKxZy99OByG\nx+NBvV6Xfk8ymZRA5/F4xM04kUiIQowlvWazCeA6gXIBGedV1BmcSYParGdJke8hA75qbT9u0aLK\nx98N75Z9vB95aEJ5b/zGb/wGVldXsba2dtCXovEBcTOksglgQfn33I8+N7GguonyYTbYWVpiOYn9\nlEajgd3dXbFDYU2dZOP3+xGLxaQcpW6AZJBjoOJEN8tfLKMwmHEGJRKJSLOeDsl8HjaBme3wtfCx\n3AHT7/fluljCY8mN0mee3Gu1Gq5cuYJWq4VkMikzLZMEujXztarNcXVrJqW9zCBV4mEv6f0ytfci\nifHHjmdLGu/E5z73ORiGgS9+8Yu6DDbhuBlSOQ/ghGEYiwC2AfwvAJ+9rVd1mzEYDMRAslQqyUme\nLr/cY8LAyjIRgxEzCTbM6TNl27ac8FXvKXURGO3x6WJcq9XQaDTg9/vRbDbhui4MwxCrFZ6i6d3F\nPfPJZFKyDgoGOp0O6vU6dnd3RXrM7IQeZ5TckmAAoFQq4cqVKygUCpJt8bknBWqmAextvjMbJakw\ne1NJhdkqDxqquzShZiY3Ip13I6Lx/gyAG5bhjjp+5Vd+BYZh4I/+6I/07MoE431JZTQaDQzD+AKA\nb+G6pPjibb+y2wjHcRCPx0V+OhgMhFjY36D8lk392dlZ6YEQnAWh6isajaLRaACAEAQDGiW/LNFQ\nEOA4DlzXlW2LXDJFZZaq7KIU2DAMGeKjuKDZbErpjOU5DmoOBgNUKhUA2OMRxsDbaDRkfwwHCMdX\nJk8C1IFVFaoai5Jgkgw/1NXPJG9mpuMEcDNZyruRDj8/aUKIO4Vf/uVfxt/8zd9oUplg3FRPZTQa\n/V8A99zma7lj6Ha7WF1dFe8v/kd3HEeMJFlu4omVTW8qi9Ttgeyt9Pt9yWY8Hg+q1aqUt5jpBINB\nyVhCoZCUohqNxp59KxxW5EbKdruNSqUifRleE6XCnU5HmvQcpCyXy/B4PDJfQzGBuhCMYgCe2geD\nAWzbntigR9JUwRIYiYMzRfza+MbMcTeFm1FxjTf3VWsffvB3ZjQaYWpq6nbdgonH7//+7+MP//AP\ncfXq1ff/Zo1DhyPZqGcpand3F7Ozs7KXncOQjuOIPJQT6wzEVGOpwZkN7263i2g0KoHDsiyRJJOA\nmAVRylytVlGv12GapjxnOBxGPB4Xo0lOv1cqFfT7fdi2DY/Hg0gkgn6/j1qtJvvZSYhbW1vw+Xzi\n3Ov1eiXTYlY1Go0QiUQkuzJNE+l0Gul0WhyQJw2qd5oK9qJY4lL7KCQVfg2AZCt8rnEyUctZ481+\nyr3VZWntdlv6XZ1OR5PKe+CXfumX8OUvf1mTyoTiSJIKg0Sv15PSE00IS6USXNeV8lM0GkW9Xodl\nWWi1WrLbhM1xNRNhEGFQoaKMga7Vaklw4c9rt9uS6ViWJRP4rNu3Wi04jiOnb05rU3rcarVw5coV\n7O7uSvAaDAbSV2Fg9Pv9sjuF5Z5YLIZcLodCoSC7XkKhENLpNGKx2AG/Sz8+1Kl5te+h9qXUaXq+\nL7wn46qwGzkQq3+Oq8ZoGNrpdIRAGo0GqtUqdnd3kc/nUSwW0Wg08OSTT975GzQh+NKXvqSb9ROM\nI0kq9OVyXRfAtUY8jQhZAqNvEwBp7lIpZhgGYrGYZB0slagqL+5B4UwMMwgqvCqVCgqFAvr9PizL\ngm3bMhNTrVZRqVREfUYRAC1W6AgwGAxgGIasAKbdDK3smenYti2WLv1+X3a8cEAynU6LUSKXgk0a\nqZAkuBp5fAmXChII3xc+ViUVPuf431VCYVbCwwQFDo7jYHd3F1tbW9jc3MTm5ia2t7dRKpXE2Vjj\nxvjSl76Ev/qrv0KxWDzoS9G4RRxJUuHwIpv1lPTylG+aJpLJpJx4mXHw3xxEpHyVDscApMzFSXuq\ntpjJMFthqSoajSIajcLv98NxHDGtbDQaUpbiNVqWBcdxxISSTX2qxThv0mg0EI1GkUqlkMlkYNs2\nUqmUkBp7B7ZtyzVEIhEhLtb/Jw0kFc7yqHMoBP/u9/tlHoeDoeryLfVDJRo+H8mEdvvMTKrVKra3\nt3H58mW8+eab2NjYwO7uLhzHkWVrkzgDdCfwF3/xF/jKV76C3d3dg74UjQ+AI0kq3W5Xyl4sLTGA\ncD7EMAwJwpQEU8mlrv8Frq/3VZVCPL2y4c6+DQclASASiSAejyMSiQhJ0U9K3b3CxzDQcx4DuCYH\nZimN0/uzs7NIp9Oy1yMWi8m0vdfrRSaTQTwex8zMjEzUU2LL5jQJbZIwbho5vreEfRIOuJqmKf0t\n9lNUpdj4hLxa6mLpkpmh4zioVCpYX1/HhQsXcPHiRWxtbaFarYqa7kYuyBrXQEIpFAoHfSmHAs89\n9xz++Z//GQDw0ksvHfDV/Hg4kqRSKBTQarVEPkrVFlVPtF+h8SJPwJy0ByDZB61V1MYulVTMCqgo\nI2lxToRWKcx4+L30C6tWqxLkea0sudXr9T17XoBrQW92dhbJZBKxWEzmWagIY8ZlWRYymYw4FLuu\nC9M0Ua1WMRqN0Gg0Ji7wqQqv8b0lqncaiZkScva+KDMmGbFfxucdz1CYbbqui1qthnK5jI2NDVy4\ncAGvvvoq1tfXZZXCuA3/pN3bO4F///d/14TyIzz33HP413/9V7z66qsHfSm3hCNJKjz5c/CRzVoA\nMrDIhnu/35caOBv8/LoqzWVpo9froVaryQZGPrdpmqjX61KiYpBhr4UlLWYtVJqRYBjoqFBjFhSJ\nROQUzLJdLBaTfgpP1plMRtRgqVQKU1NTiMVie5RLHI6sVCoTGfjUDY9UeJEQqMLin2rvi/f2vTIV\ntSnPQwBLX67rolQqYXNzE+vr67J2mAowtbcziff13fCNb3wDzz777L481w9/+MN9eZ67ARcvTvQY\n4NEkFcplOdhIu/dgMCgT9XS7paRXndZmwOHn1DILS110zKWtCy3XG42GnILZUOfzMdthvyMajYp0\nmDMm/Fn8PgZM/mwSFhdxMcPiz+Sa4Xg8DsuyUK1WZbpeNcRULeInAapTMEtZLEUye+RMkDr0qkqG\nVV+3G5W/1Ma86tvWaDRQqVSQz+dRKpVkaPbdhAJ3Cy5evIivf/3rB30ZGocMR5JU6NQLXAsU3PHO\njITZCZdVjUYjxGIx2Rzo9XqlRKb6ZzFIsewE7FUJ0XqeJ1medmnTwsVZtm3L0i5eJ0kKgAQ/VTig\nOiAzYBqGIb0R+ojxdVC2TM8wddFULBabyGYyeyrqjhvg+v56kgkzSvbS+L6xWa+WOYG9vRR17kjN\nNrvdrqjzjgKhaGi8G44kqXBaXFVm0W1Y3e5IOxOWMbjel1P4nLLnXApP+RwwbDQaIgqIRqMAIGUu\nlrJo4cLAxH3xvI5wOLzH90vNoNSFYVyBrPZ3SBiO44goALjmFVYul0UlBUCCKQMlxQSTADVDYZNe\nXbjVbrdFmVWpVPZkc8z22ONSd9uMT8TzEMDfDfbJQqEQEokEstksqtWqqAj5XvL9GF/+paFxN+JI\nkkq73UYkEtlzmux0OmKtwkHEdDot1vBszLZaLSQSCViWhWAwKL2SWq0mAco0TZm8pypLHToMBoOI\nxWJiBMmvsT7P0hqDHN2OaQXD5jKJiFYu7A+oJokMjK1WC41GA/l8HrlcDul0Gn6/H7lcDrVaTfo6\nbC5PYuBT+x+qpJjvgeu6YnXDBjrLXKpC60bW+CRrlSD4c7heOJvNolQqYXd3F7VaTa7l3YwpNTTu\nRhxJUmGGwaBCCe1gMJAAzpMrHYtN0xS3X25LZHmp2WzKY1Q7Fiq5qBZS5yDY4G+1WtLUj0Qi4kGl\nTtur9vi8ZroYc9qfAgCW3sLhsDgg8zW3Wi0Ui0XUajXE43EhS9u2YRiGCAk4ADpJUJVfqvqLZc1m\nsyl7ZliiUnfqqCudx8tf49PyqtCC80OxWAxTU1OoVqvY2tqSWQvV6kWrvzSOAo4kqcTjcZimKQGG\nPZJ+v49IJIJisYhKpYLhcCgDgmzicwUvJbjNZlM8tTjPwn0sDMxcA8wyFs0hOV3Nxjit6pkpdTod\nKaFQ6kt/MdrsMytJpVLweDyIRqPibtztduXzJEueyovFIjyea7vpWfqKRqMol8syyzIpIEkHg0FE\no9E9pUHKiB3HEZJlP0XNVJlFqoo+Ve3Fnol6gFBVc5ZlIZVKYW5uDuvr69ja2pI1BhoaRwlHklRM\n05Q+CgcMqYBSbe85cR0KheR0yv0kVPgwWFmWJWSi9iMYeFqtljxeDVa9Xg+VSkVUXpyT4ZR9s9mU\nUg1nTGzbRiaTQb1elwl9klokEoHP54Nt24jFYhI8aUNDAUC9Xsfs7Ky8Hl4PBwIpNJgEULkViUSQ\nTCYRCoX2kArLiuVyWYj53ZrpanOfcynMdOr1OlzXlfIgs1KKAlgynZmZwcbGBlzXFRHE3Yavfe1r\n+Nu//duDvgyNQ4gjSSrqMi7+h+eCK7UfUa1WMRwOMTU1JT0Jzoe0223x4Wq1WpKVcKaEC7xoK+/3\n+6Vh3O123zEXoe5lUZvNvB6W3jweD2KxGOLxOEajESqViljA0Kus3W4jEAggHA5L74UEpM7A0KnZ\n6/UiGo3K12zbnih/KtWxgKTCMiNFFJT98j1Up+tV/y/2syg3ZxmShMJsR93bopJ5PB7H0tKS9G7Y\np7qRQeUk4+rVq9r0UeOGOJKkwrXB427FAGRXSbvdRiwWQ6fTwc7ODkajEWq1mlie7O7uyjItzpOM\nRiPYti12+Wz+M/thCatSqch6Yroek0g490BlGPsD7O0AgG3bYga5ubmJYrGIYDCIqakpVCoV6Rkw\ny2I5zTTNPQ3/t956CxsbGwiHwzh16pSQISfFJwEkg0AggEQigUwmg1AoJBb2LF25riu282qzXd17\nosrMSSpqP6ZSqQgxqTNKLHWapolwOIyFhQUpi/l8Ply5cgWO46DX6901pKKh8W440qTCchTNIqnG\n8vl8EmRIBI7jiLsvy0n8Gr2jmFH4fD4JIMwK2NB3XVca5FwaxayCP8/n88npljJhAOI55bouhsMh\nkskkMpkMXn/9dbRaLezu7iIajcpJneTEXgoAUYrt7u6K8WSv18PW1hYsywJw3SNrEsB+SigUElWb\nZVlC8jTxHHczGG+g84ChzgVxUJKlr2KxiHK5LEOtzOwCgYDcY64xoCs0Z4muXLki5bBJnAHS0LhZ\nHElSYeOdHlrsRxiGIb0EWr8Xi0VphHORFxvk7Mk4joNGo4FkMim26wwcJB6SCnsxLFmpvRTulSfZ\nsS8QDAb3+HFRTUb/rnA4jLW1Nbiui5MnT4pcuFarYTgcIp1Oy8mb8y+u68rQZywWE9GA1+tFrVZD\nLpe702/LLYEqvXQ6jaWlJaRSKcnumK1xpwn7KeokPctY6kCjahrJNc6O46BYLKJYLMrKZv4cZoA8\nnPAgwZUE7Pmsra2hVqvJEKuGxt2II0kqLAGxTKKeygeDgTSzu90ubNuWOjwAIQbKjFkqUfeWsBxC\nyTG3STJw80TLvkWv14NlWbIoy+fzSf+D8yyqVT2DXDabhW3biMfj0qynvQsdlNk/Uu1m1BJQp9NB\noVAQ4UIsFpuYJj17X5FIBPPz85ifn4dt21ImZJbG7ZdqP4VfZ8bKOSTVsVpVfXH/TaFQkEFTy7Jk\n4JXlTW7u9Hg8CIVCIuCIRCKwLAtvvPHGxFu7//3f/z3++q//+qAvQ+OQ4kiSCstY7GGwdk4pKqfk\naQnPiXfureeCLPp4RaNRmdKv1+uywGs0GsnaXg7EdTqdPaaUqryVQZ1KIg5S9vt9mTthA9g0TSQS\nCcmYKKcFgHK5LEOUoVBIpvvL5bJImFWzRboTh0IhhMNhIafDDmYAyWQSS0tLmJ6eFjsa3v/BYCDl\nL9XvC7iuzFNLZONrn0lKVI+VSiXJcNlPU21t1H33zGYo6Wb/hdnipKJYLCKfzx/0ZWgcUhxJUmEp\niKACiKTCYUMG3Wg0Ctd1ZdaDJQ86E7MRTndgBjYqw0gkAPZMrJumKeUzyoe73e4e52HW6rneuFAo\nYH19HVeuXNkzM7OwsIB6vY5SqSRDeuFwGJVKBZ1OR4gUgBAam/L8OwUB6tKxwwr2UizLwvT0NJaW\nlsTmn016kgr3nvCeqs1yTspT4cVMhgIL1dtLXQVNqTdl5yQn/l1dNc3fF6r8SqXSQd02DY3bjiNJ\nKixtqcGdvlhsxFOKyhIZA1QgEBD7Fc6UMDjX6/U9y5/Yw6hUKrKvnGaV9XpdghTVXNx7AkB2sKhN\nesMwkEgkUC6XAUBMIWlqSWt7ns759VarBdu2ZdCRbsQUEsRiMSEWBsHDvvmRrzuRSODkyZNYWFgQ\nR4JxMmAJUm3Qq+BsEYmCZTG1Gc8yGf3h+G86FfDwwIxFlSgz+xsOhzh37pzeG6JxV+NIkgpr7ZyE\nt20btm3vKYGwTs7Zj1AohFKpJLYntm2LcovBh6fcwWAgq3+LxSJ2d3exvb2NWq0mBpa8BpZv4vE4\nQqGQBHvW5elTxRIMVV80lqxWq+L/xcVenNwHrs3fcMrc7/cjFotJduXxeESavL29jXQ6DdM04TiO\nvO7DCBKobdtYXl7GysoKcrmcrBhg5slGO/sp76a6Yl+MGzRJOuockZqtUOBAW5xSqYRMJiNEzpKm\nqg6j19yJEyfwxBNP3Mnbta/46le/ij//8z8/6MvQOMQ4kqTCMofjOKLcASANdboA0zalWCzKACMX\nZ9EWBID0KcLhMLrdLmq1GtbX14W4ODfCTZGqIy4XbqkrgVWjQl5bOByW7IlZFk/MtVoNxWIR4XAY\nyWQSjuPgzTffRKFQwNTUlPRaqOwaDAaSmXESfX5+XkwlO50O1tfXb/n+sox2O2YyWPaybRtzc3M4\nc+YMZmdnheTVXgpnfjgF/27XRILggCgzFc6ZcJaJe+hVUjFNU1RhuVxuD6mwDMfeD/tUKysr+35f\n7hRc10W1Wj3oy9A4xDiSpFKr1aTUwyBUKpUkYHAhFnsmak3c6/WKDQuNINXd6J1OB7VaDbu7u1Jy\nUa3zOfHtOI4ELKrFuAIYgEhZWd9PJBJi/MgASIWXaZqYmZlBMpnE9PQ01tfXkc/nsbq6KoQ3MzMj\nsmYAYtnOuZVEIgHTNMXzrFgs3vL9DYVCewQI+zXzwkAfDocxPT2NM2fO4MSJE0ilUtIP4/2k1Q6b\n67TjGQeb9cxomA2qP4+ZCjNGkkq/30etVkOhUEAqlUK5XEYul9uzUVN9nkAgIASvoXG34kiSCgcL\nuQMlmUxK0N/c3JQSCGvo6XRaJqZpdd9oNKRcxSDPGQhOcNNOndkG95iML3LiKZvzM+yJMLDPzs6K\nwqjRaKBYLEr9n4QTiUTk5/Jn0lfMdV0po/l8Pgm06XQapVJJAl0ulxMl2wc5jS4tLYnPFvtPHzRz\nUYccs9ksTp8+jXvuuQdTU1NiR0MpsWrNQin3e02zqw19ZrHMVnhYoDQYgDTkVft727aRSqWQyWQQ\ni8VEXq5ePw8Rk0oszzzzDP70T//0oC9D45DjSJJKuVxGvV6XxjUDKwMpm+6RSEQkwVRP0cOLAYx9\nGZ/PJ3V5nlJpMMlGLocgTdOEbduyh4P7WagO4sxJJpOB3++XqXqWcLgWuNFoSKaVz+dliM8wDJw5\ncwaVSgVXr17dU1LjXE6r1ZJMynVdKd3w+RcWFm75/j788MPY2trCzs6OZGzsN41Psb8fKH32+Xyw\nLAuZTAanTp3CysoKlpeXkc1mhVSYpZCUW60WSqUSqtXqe06x81o6nY6ou3gf6ODMQwWAPfYutO/Z\n3NxEMplENptFIpHY08dSJ/VZUptEUFiiofFemMzf7g8INsm5w51lLu7b6PV6SCQSmJ6eRjAYFHko\nsxQOE6qNfJ502Zxn0OE0PEtPlPqyjKNmB2ppjXMNwWDwHQu5WOrh9XBmwLZteb7Lly+LzT2vORKJ\n7Alq+XwetVoNU1NT2NzcFNuWUCiE6enpW76/KysrIj6wLAvFYlHmRHjKV/e8j5eKCCqoOCuUTqdx\n8uRJrKys4MSJE5idnUU6nZZMjw16NevI5/NwHOcdJTj156mu0ePGofQV4/ug+oPxT84hvfXWW0gk\nEpiamhI1n7puWn1Nk4ZnnnkGTz/99EFfhsYE4EiSSq1WkxKW1+sVs0FmIpy2b7fbMgthGAZc1xWr\nj2q1KhPu3GdP00AqqNhQp8UKsxCqlADIECOt7tlHicfjyGazaDQactrv9/solUpiUEhX41QqBeC6\nIzJJErhGdpubm8jn83jkkUcwPT0Nr9eLF154QaTG+XxeTtW0ffkgU/WnTp1CPB5HMplENBrFzs4O\nyuWyWJQwi7hR34WnerUPQbI8fvw4Tp8+jTNnzmBxcRHpdBqxWGzPbIq6lKtYLGJnZ0fcgtWhxxth\nMBhIlsr3Se2XUVyhEh/7N41GAxsbG0LIJDtmoPwAMJGkQhGJhsb74UiSCu3iI5GInEapqhofWgSu\nn2Y56FatVsUqRQ1mDC4sX9XrdVQqFbH/GA6HiMfjiMfjiEajMnDIoNfpdBAIBODz+dBsNpHP52VG\nhtkVyzsbGxswDAPHjx+XWRSqolqtFtbW1rC5uQnLspBMJrG8vCzNfo/HgwcffBAXLlyA4zjwer2o\n1+uSGXGT5K0im83KThiSZrFYRKlUkvtWq9XEWFMtjXm9XiF2y7IQj8cxMzODhYUFLC4u4tixY1hY\nWEAmkxHrE3VQVQ3y5XIZxWJRylk3AnsnvP+1Wk3Ko7FYTIiNbgMksPE+0XA4hOu6uHr1Kl555RWk\n02mZYeJMDYUEk0YqX/3qV/G7v/u7B30ZGhOCI0kq5XJZSia0uE8kEmJXwgn6drstg5Ku64pais1v\nBhtOrTMIMRNqNptoNBpiyxEOh6WpXq/X0e/3MT09jYWFBZlr4UlY3YkSi8UQiUTEO4wOx+12G1ev\nXpUMh5src7kczp49i52dHaytrcHj8aBSqYjlfSwWQ71eRzablUHISCQiROY4DhKJBJaWlm7p/mYy\nGSn9UEqbzWZl0ya3XlJtpQ5qkkj4kclkMD09jZmZGWSzWclOwuGwZA7qYq3x9cGc43m/LEV1gC4W\ni9IHUlVbzD52d3ffsdGRpF+tVrG6uoo333wTiURCvOB4gBnPdCYBLClqaNwMDL3fQUNDQ0NjvzBZ\nebiGhoaGxqGGJhUNDQ0NjX2DJhUNDQ0NjX2DJhUNDQ0NjX2DJhUNDQ0NjX2DJhUNDQ0NjX2DJhUN\nDQ0NjX2DJhUNDQ0NjX2DJhUNDQ0NjX2DJhUNDQ0NjX2DJhUNDQ0NjX2DJhUNDQ0NjX2DJhUNDQ0N\njX2DJhUNDQ0NjX2DJhUNDQ0NjX2DJhUNDQ0NjX2DJhUNDQ0NjX2DJhUNDQ0NjX2DJhUNDQ0NjX2D\nJhUNDQ0NjX3D/wNHKvsLeiw/nQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5644a82990>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAABpCAYAAADyfL7tAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5643089710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAABpCAYAAADyfL7tAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5642560a50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAABpCAYAAADyfL7tAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5659f856d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAABpCAYAAADyfL7tAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nGzSpaGhoaGhsGzSpaGhoaGhsG/4v66OYOtSyMIMAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f564a41c090>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAABpCAYAAADyfL7tAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n0Dgy/D/dlhsgaa7oPQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f563f9b1710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAABpCAYAAADyfL7tAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\noqGhoaFxavg/f6h2sX72oxoAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f56411f9510>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAABpCAYAAADyfL7tAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQ2PfoElFQ0NDQ2Pf8P8ATwzbq0w0g/UAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f565b7f53d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAABpCAYAAADyfL7tAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\naGhoaJwa/h8Vi/Ug8ch74AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f563f8fb5d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAABpCAYAAADyfL7tAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQyoGBgYGBheG/wcPPHj0f4WFtgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5653e50f90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for key in range(10):\n",
    "    mlpu.Util.view_rgba(X_train[key], text=y_train[key])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "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": 0
}