RGBA_image_prob_mergedarray_borderoverlap.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "mlproof/tifffile.py:1995: UserWarning: failed to import _tifffile.decodepackbits\n",
      "  warnings.warn(\"failed to import %s\" % module_function)\n",
      "mlproof/tifffile.py:1995: UserWarning: failed to import _tifffile.decodelzw\n",
      "  warnings.warn(\"failed to import %s\" % module_function)\n",
      "mlproof/tifffile.py:1995: UserWarning: failed to import _tifffile.unpackints\n",
      "  warnings.warn(\"failed to import %s\" % module_function)\n",
      "Using gpu device 0: GeForce GTX TITAN (CNMeM is disabled, CuDNN 4007)\n",
      "/home/d/nolearn/local/lib/python2.7/site-packages/theano/tensor/signal/downsample.py:6: UserWarning: downsample module has been moved to the theano.tensor.signal.pool module.\n",
      "  \"downsample module has been moved to the theano.tensor.signal.pool module.\")\n"
     ]
    }
   ],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "import mlproof as mlp\n",
    "import mlproof.nets as nets\n",
    "\n",
    "import numpy as np\n",
    "\n",
    "from nolearn.lasagne.visualize import plot_loss\n",
    "from nolearn.lasagne.visualize import plot_conv_weights\n",
    "from nolearn.lasagne.visualize import plot_conv_activity\n",
    "from nolearn.lasagne.visualize import plot_occlusion\n",
    "\n",
    "from matplotlib.pyplot import imshow\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CNN configuration: \n",
      "    Our CNN with image, prob, merged_array, border_overlap as RGBA.\n",
      "    \n"
     ]
    }
   ],
   "source": [
    "cnn = nets.RGBANet()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded /home/d/patches//cylinder_small1_rgba/ in 0.144218921661 seconds.\n"
     ]
    }
   ],
   "source": [
    "X_train, y_train, X_test, y_test = mlp.Patch.load('cylinder_small1_rgba')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# Neural Network with 1558642 learnable parameters\n",
      "\n",
      "## Layer information\n",
      "\n",
      "  #  name      size\n",
      "---  --------  --------\n",
      "  0  input     4x75x75\n",
      "  1  conv1     64x63x63\n",
      "  2  pool1     64x31x31\n",
      "  3  conv2     48x19x19\n",
      "  4  pool2     48x9x9\n",
      "  5  hidden3   256\n",
      "  6  dropout3  256\n",
      "  7  output    2\n",
      "\n",
      "  epoch    train loss    valid loss    train/val    valid acc  dur\n",
      "-------  ------------  ------------  -----------  -----------  -----\n",
      "      1       \u001b[36m0.70505\u001b[0m       \u001b[32m0.67564\u001b[0m      1.04353      0.68583  8.04s\n",
      "      2       0.79774       0.68894      1.15792      0.49935  8.00s\n",
      "      3       \u001b[36m0.61422\u001b[0m       \u001b[32m0.61823\u001b[0m      0.99352      0.71225  7.97s\n",
      "      4       \u001b[36m0.55081\u001b[0m       \u001b[32m0.51828\u001b[0m      1.06277      0.77054  7.97s\n",
      "      5       \u001b[36m0.46730\u001b[0m       \u001b[32m0.43158\u001b[0m      1.08277      0.80996  7.99s\n",
      "      6       \u001b[36m0.42587\u001b[0m       \u001b[32m0.37492\u001b[0m      1.13590      0.84105  8.01s\n",
      "      7       \u001b[36m0.35317\u001b[0m       \u001b[32m0.30767\u001b[0m      1.14789      0.87543  7.97s\n",
      "      8       \u001b[36m0.28840\u001b[0m       \u001b[32m0.26615\u001b[0m      1.08361      0.89087  7.98s\n",
      "      9       \u001b[36m0.24999\u001b[0m       \u001b[32m0.22105\u001b[0m      1.13092      0.91409  7.97s\n",
      "     10       \u001b[36m0.17840\u001b[0m       \u001b[32m0.16604\u001b[0m      1.07439      0.94134  8.04s\n",
      "     11       \u001b[36m0.14043\u001b[0m       \u001b[32m0.13218\u001b[0m      1.06244      0.96040  8.07s\n",
      "     12       \u001b[36m0.10137\u001b[0m       \u001b[32m0.11410\u001b[0m      0.88839      0.96457  8.08s\n",
      "     13       \u001b[36m0.07534\u001b[0m       \u001b[32m0.10487\u001b[0m      0.71846      0.97442  8.07s\n",
      "     14       \u001b[36m0.06356\u001b[0m       \u001b[32m0.08386\u001b[0m      0.75796      0.98047  8.07s\n",
      "     15       \u001b[36m0.04265\u001b[0m       \u001b[32m0.07235\u001b[0m      0.58952      0.98047  8.07s\n",
      "     16       \u001b[36m0.03351\u001b[0m       \u001b[32m0.06619\u001b[0m      0.50628      0.98728  8.08s\n",
      "     17       \u001b[36m0.02121\u001b[0m       \u001b[32m0.06116\u001b[0m      0.34680      0.98652  8.07s\n",
      "     18       \u001b[36m0.01648\u001b[0m       \u001b[32m0.05501\u001b[0m      0.29953      0.98486  8.07s\n",
      "     19       0.01669       0.06112      0.27303      0.98819  8.10s\n",
      "     20       \u001b[36m0.01324\u001b[0m       0.06310      0.20986      0.98986  8.09s\n",
      "     21       \u001b[36m0.01265\u001b[0m       0.05787      0.21862      0.99159  8.07s\n",
      "     22       \u001b[36m0.00760\u001b[0m       0.05723      0.13285      0.98826  8.07s\n",
      "     23       0.00859       0.06927      0.12405      0.98993  8.07s\n",
      "     24       0.02294       \u001b[32m0.04708\u001b[0m      0.48723      0.99250  8.07s\n",
      "     25       0.01525       0.06795      0.22446      0.98826  8.07s\n",
      "     26       0.01623       0.05706      0.28450      0.98312  8.17s\n",
      "     27       0.01374       0.06158      0.22315      0.98736  8.27s\n",
      "     28       0.01375       \u001b[32m0.04675\u001b[0m      0.29402      0.99326  8.27s\n",
      "     29       \u001b[36m0.00619\u001b[0m       0.05188      0.11937      0.98743  8.28s\n",
      "     30       0.00649       0.05330      0.12175      0.99159  8.27s\n",
      "     31       \u001b[36m0.00338\u001b[0m       0.06405      0.05279      0.98993  8.28s\n",
      "     32       0.00435       0.06829      0.06368      0.99159  8.28s\n",
      "     33       0.00522       0.06041      0.08641      0.98993  8.27s\n",
      "     34       0.00530       0.05063      0.10478      0.98993  8.28s\n",
      "     35       0.00473       0.06904      0.06845      0.98993  8.28s\n",
      "     36       \u001b[36m0.00314\u001b[0m       0.05567      0.05644      0.98986  8.28s\n",
      "     37       0.00468       0.07018      0.06672      0.99159  8.27s\n",
      "     38       0.00479       0.05064      0.09468      0.98993  8.28s\n",
      "     39       0.00319       0.05333      0.05989      0.98993  8.28s\n",
      "     40       \u001b[36m0.00261\u001b[0m       0.05304      0.04917      0.98993  8.28s\n",
      "     41       0.00440       0.06538      0.06736      0.98993  8.28s\n",
      "     42       0.00461       0.05721      0.08061      0.98993  8.28s\n",
      "     43       0.00476       0.06468      0.07356      0.99159  8.28s\n",
      "     44       0.00472       0.06291      0.07505      0.99159  8.28s\n",
      "     45       0.00324       0.05169      0.06271      0.99159  8.37s\n",
      "     46       0.00376       0.06363      0.05907      0.99159  8.39s\n",
      "     47       \u001b[36m0.00183\u001b[0m       0.06631      0.02764      0.99159  8.39s\n",
      "     48       \u001b[36m0.00100\u001b[0m       0.07053      0.01418      0.99159  8.39s\n",
      "     49       \u001b[36m0.00080\u001b[0m       0.06875      0.01158      0.99159  8.39s\n",
      "     50       \u001b[36m0.00062\u001b[0m       0.07440      0.00838      0.98993  8.39s\n",
      "     51       0.00131       0.06800      0.01924      0.98993  8.38s\n",
      "     52       0.00124       0.06839      0.01810      0.98993  8.39s\n",
      "     53       0.00096       0.06493      0.01472      0.99159  8.43s\n",
      "     54       0.00097       0.05952      0.01626      0.99159  8.39s\n",
      "     55       0.00097       0.05944      0.01636      0.98993  8.39s\n",
      "     56       0.00067       0.06172      0.01094      0.98993  8.42s\n",
      "     57       \u001b[36m0.00052\u001b[0m       0.06294      0.00823      0.99159  8.39s\n",
      "     58       \u001b[36m0.00032\u001b[0m       0.06276      0.00510      0.98993  8.38s\n",
      "     59       \u001b[36m0.00027\u001b[0m       0.06728      0.00403      0.98993  8.38s\n",
      "     60       0.00032       0.07125      0.00448      0.99159  8.39s\n",
      "     61       0.00039       0.07173      0.00541      0.98993  8.41s\n",
      "     62       0.00078       0.06029      0.01291      0.99326  8.38s\n",
      "     63       0.00048       0.06333      0.00763      0.99326  8.39s\n",
      "     64       0.00033       0.06868      0.00484      0.99159  15.98s\n",
      "     65       0.00034       0.07010      0.00488      0.99159  16.89s\n",
      "     66       0.00066       0.06936      0.00954      0.99159  16.46s\n",
      "     67       0.00031       0.06509      0.00477      0.99159  16.45s\n",
      "     68       0.00030       0.07212      0.00410      0.99159  16.45s\n",
      "     69       0.00309       0.12720      0.02427      0.98228  16.44s\n",
      "     70       0.00822       0.06176      0.13316      0.98978  16.46s\n",
      "     71       0.00390       0.05943      0.06565      0.99326  16.46s\n",
      "     72       0.00403       0.05901      0.06829      0.99159  16.47s\n",
      "     73       0.00247       0.05539      0.04450      0.99333  16.47s\n",
      "     74       0.00235       0.06631      0.03545      0.98902  16.46s\n",
      "     75       0.00292       0.06648      0.04391      0.98993  16.47s\n",
      "     76       0.00156       0.07572      0.02064      0.99076  16.47s\n",
      "     77       0.00235       0.07230      0.03244      0.98993  16.48s\n",
      "     78       0.00239       0.08320      0.02873      0.98826  16.46s\n",
      "Early stopping.\n",
      "Best valid loss was 0.046753 at epoch 28.\n",
      "Loaded parameters to layer 'conv1' (shape 64x4x13x13).\n",
      "Loaded parameters to layer 'conv1' (shape 64).\n",
      "Loaded parameters to layer 'conv2' (shape 48x64x13x13).\n",
      "Loaded parameters to layer 'conv2' (shape 48).\n",
      "Loaded parameters to layer 'hidden3' (shape 3888x256).\n",
      "Loaded parameters to layer 'hidden3' (shape 256).\n",
      "Loaded parameters to layer 'output' (shape 256x2).\n",
      "Loaded parameters to layer 'output' (shape 2).\n"
     ]
    }
   ],
   "source": [
    "cnn = cnn.fit(X_train['rgba'], y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8017638036809815"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# this is the score on cylinder_small1 with regularization and patience counter 50 and data augmentation\n",
    "cnn.score(X_test['rgba'], y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<module 'matplotlib.pyplot' from '/home/d/nolearn/local/lib/python2.7/site-packages/matplotlib/pyplot.pyc'>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAEPCAYAAABGP2P1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nqhgtDEoppYr5f3Vi5C4hbBrEAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fa6b3871a50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# this is the plot on cylinder_small1 with regularization and patience counter 50 and data augmentation\n",
    "plot_loss(cnn)"
   ]
  },
  {
   "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
}