Initial weights XOR.ipynb

{
 "metadata": {
  "name": "",
  "signature": "sha256:6ba0b3837a14838bd025115466103ac54b26187188559f8e092f07f7bc675bc1"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "raw",
     "metadata": {},
     "source": [
      "Hi reader!\n",
      "As far as I understood, plots are a big deal for experimental people. Sometimes even more important than results themselves.\n",
      "So I decided to just fit in that atmosphere and plot!\n",
      "I always wanted to know what's happening inside these NNs so I want to find a way to plot weights for my favorite NN (XOR) and see if I can understand anything about how they behave during training process."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import matplotlib.pyplot as plt\n",
      "import theanets\n",
      "import time\n",
      "import numpy as np\n",
      "from numpy import float32\n",
      "dt = np.dtype(float32)\n",
      "trainer = theanets.feedforward.Regressor(layers=(2,2,(1,'logistic')))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Couldn't import dot_parser, loading of dot files will not be possible.\n"
       ]
      }
     ],
     "prompt_number": 1
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "dataset_input = np.array([[1,1],[0,1],[1,0],[0,0]],dt)\n",
      "dataset_output = np.array([[0],[1],[1],[0]],dt)\n",
      "np.save('network/XOR_Input.npy', dataset_input)\n",
      "np.save('network/XOR_Output.npy', dataset_output)\n",
      "np.save('network/XOR_Input.npy', np.concatenate((np.load('network/XOR_Input.npy', mmap_mode='r'),dataset_input)))\n",
      "np.save('network/XOR_Output.npy', np.concatenate((np.load('network/XOR_Output.npy', mmap_mode='r'),dataset_output)))\n",
      "BATCH_SIZE = 4\n",
      "training_data_ratio = 0.5\n",
      "def train_batch():\n",
      "    X = np.load('network/XOR_Input.npy', mmap_mode='r')\n",
      "    Y = np.load('network/XOR_Output.npy', mmap_mode='r')\n",
      "    cut = training_data_ratio * len(X)\n",
      "    if cut == BATCH_SIZE:\n",
      "        i = 0\n",
      "    else:\n",
      "        i = np.random.randint(cut-BATCH_SIZE)\n",
      "    return (X[i:i+BATCH_SIZE],Y[i:i+BATCH_SIZE])\n",
      "def valid_batch():\n",
      "    X = np.load('network/XOR_Input.npy', mmap_mode='r')\n",
      "    Y = np.load('network/XOR_Output.npy', mmap_mode='r')\n",
      "    cut = training_data_ratio * len(X)\n",
      "    if cut == len(X)-BATCH_SIZE:\n",
      "        i = cut\n",
      "    else:\n",
      "        i = np.random.randint(cut,len(X)-BATCH_SIZE)\n",
      "    return (X[i:i+BATCH_SIZE],Y[i:i+BATCH_SIZE])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 2
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print trainer.find('hid1', 'b').get_value()\n",
      "print trainer.find('hid1', 'w').get_value()\n",
      "print trainer.find('out', 'b').get_value()\n",
      "print trainer.find('out', 'w').get_value()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[-1.15093398  3.28787665]\n",
        "[[ 0.48318008  0.60989918]\n",
        " [ 0.92120928  0.43320099]]\n",
        "[-0.81879884]\n",
        "[[-0.04571631]\n",
        " [ 0.9527267 ]]\n"
       ]
      }
     ],
     "prompt_number": 3
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "trainer.find('hid1', 'b').set_value([5,5])\n",
      "trainer.find('hid1', 'w').set_value([[1,-1],[-1,1]])\n",
      "trainer.find('out', 'b').set_value([20])\n",
      "trainer.find('out', 'w').set_value([[-3],[-3]])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 4
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "for i in range (1,2) :\n",
      "    for j in range (1) :\n",
      "        start = time.mktime(time.gmtime())\n",
      "        print \"training started ... \"\n",
      "        epoch_counter = 0\n",
      "        LR = 1e-2\n",
      "        for trn, val in trainer.itertrain(train_batch(),valid_batch(), algo='rprop',learning_rate = 1,iteration_size = 1,batch_size=BATCH_SIZE,validate_every = 1,patience=500) :\n",
      "            if epoch_counter%100 == 0 :\n",
      "                if epoch_counter < 1000:\n",
      "                    plt.figure(1 + epoch_counter/100)\n",
      "                    plt.plot(trainer.find('hid1', 'w').get_value())\n",
      "                print \"epoch \",epoch_counter+1\n",
      "                print \"training_error = \\t\",trn['err']\n",
      "                print \"validation_error = \\t\",val['err']\n",
      "            epoch_counter += 1\n",
      "        print \"\\t training ended\"\n",
      "        stop = time.mktime (time.gmtime())\n",
      "        print \"training time (seconds) \\t\" , stop - start"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "training started ... \n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 1\n",
        "training_error = \t0.499999201203\n",
        "validation_error = \t0.499999201203\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 101\n",
        "training_error = \t2.77810722861e-07\n",
        "validation_error = \t2.77810722861e-07\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 201\n",
        "training_error = \t1.70373943755e-07\n",
        "validation_error = \t1.70373943755e-07\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 301\n",
        "training_error = \t1.98527866444e-07\n",
        "validation_error = \t1.98527866444e-07\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 401\n",
        "training_error = \t1.22334769496e-07\n",
        "validation_error = \t1.22334769496e-07\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 501\n",
        "training_error = \t9.44256090614e-08\n",
        "validation_error = \t9.44256090614e-08\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 601\n",
        "training_error = \t8.05428606517e-10\n",
        "validation_error = \t8.05428606517e-10\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 701\n",
        "training_error = \t2.08652396658e-15\n",
        "validation_error = \t2.08652396658e-15\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 801\n",
        "training_error = \t2.92304071074e-17\n",
        "validation_error = \t2.92304071074e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 901\n",
        "training_error = \t5.25644741952e-18\n",
        "validation_error = \t5.25644741952e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 1001\n",
        "training_error = \t2.88215911459e-17\n",
        "validation_error = \t2.88215911459e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 1101\n",
        "training_error = \t2.35613397425e-18\n",
        "validation_error = \t2.35613397425e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 1201\n",
        "training_error = \t5.21443427886e-18\n",
        "validation_error = \t5.21443427886e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 1301\n",
        "training_error = \t2.86730611764e-17\n",
        "validation_error = \t2.86730611764e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 1401\n",
        "training_error = \t2.3560395933e-18\n",
        "validation_error = \t2.3560395933e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 1501\n",
        "training_error = \t5.17317000793e-18\n",
        "validation_error = \t5.17317000793e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 1601\n",
        "training_error = \t2.85269490673e-17\n",
        "validation_error = \t2.85269490673e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 1701\n",
        "training_error = \t2.35594615523e-18\n",
        "validation_error = \t2.35594615523e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 1801\n",
        "training_error = \t5.13263803801e-18\n",
        "validation_error = \t5.13263803801e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 1901\n",
        "training_error = \t2.83832092655e-17\n",
        "validation_error = \t2.83832092655e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 2001\n",
        "training_error = \t2.35585365058e-18\n",
        "validation_error = \t2.35585365058e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 2101\n",
        "training_error = \t5.09282224124e-18\n",
        "validation_error = \t5.09282224124e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 2201\n",
        "training_error = \t2.82417931823e-17\n",
        "validation_error = \t2.82417931823e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 2301\n",
        "training_error = \t2.35576207e-18\n",
        "validation_error = \t2.35576207e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 2401\n",
        "training_error = \t5.05370691703e-18\n",
        "validation_error = \t5.05370691703e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 2501\n",
        "training_error = \t2.81026613683e-17\n",
        "validation_error = \t2.81026613683e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 2601\n",
        "training_error = \t2.3556714042e-18\n",
        "validation_error = \t2.3556714042e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 2701\n",
        "training_error = \t5.01527677897e-18\n",
        "validation_error = \t5.01527677897e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 2801\n",
        "training_error = \t2.79657643602e-17\n",
        "validation_error = \t2.79657643602e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 2901\n",
        "training_error = \t2.35558164403e-18\n",
        "validation_error = \t2.35558164403e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 3001\n",
        "training_error = \t4.9775169421e-18\n",
        "validation_error = \t4.9775169421e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 3101\n",
        "training_error = \t2.78310625077e-17\n",
        "validation_error = \t2.78310625077e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 3201\n",
        "training_error = \t2.3554927804e-18\n",
        "validation_error = \t2.3554927804e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 3301\n",
        "training_error = \t4.94041291076e-18\n",
        "validation_error = \t4.94041291076e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 3401\n",
        "training_error = \t2.76985132675e-17\n",
        "validation_error = \t2.76985132675e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 3501\n",
        "training_error = \t2.35540480432e-18\n",
        "validation_error = \t2.35540480432e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 3601\n",
        "training_error = \t4.90395056673e-18\n",
        "validation_error = \t4.90395056673e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 3701\n",
        "training_error = \t2.75680748787e-17\n",
        "validation_error = \t2.75680748787e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 3801\n",
        "training_error = \t2.35531770691e-18\n",
        "validation_error = \t2.35531770691e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 3901\n",
        "training_error = \t4.86811615783e-18\n",
        "validation_error = \t4.86811615783e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 4001\n",
        "training_error = \t2.74397087974e-17\n",
        "validation_error = \t2.74397087974e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 4101\n",
        "training_error = \t2.35523147934e-18\n",
        "validation_error = \t2.35523147934e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 4201\n",
        "training_error = \t4.83289839072e-18\n",
        "validation_error = \t4.83289839072e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 4301\n",
        "training_error = \t2.73133736983e-17\n",
        "validation_error = \t2.73133736983e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 4401\n",
        "training_error = \t2.3551461129e-18\n",
        "validation_error = \t2.3551461129e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 4501\n",
        "training_error = \t4.79827999009e-18\n",
        "validation_error = \t4.79827999009e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 4601\n",
        "training_error = \t2.71890351251e-17\n",
        "validation_error = \t2.71890351251e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 4701\n",
        "training_error = \t2.35506159896e-18\n",
        "validation_error = \t2.35506159896e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 4801\n",
        "training_error = \t4.76425254281e-18\n",
        "validation_error = \t4.76425254281e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 4901\n",
        "training_error = \t2.70666510149e-17\n",
        "validation_error = \t2.70666510149e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 5001\n",
        "training_error = \t2.35497792898e-18\n",
        "validation_error = \t2.35497792898e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 5101\n",
        "training_error = \t4.73079709517e-18\n",
        "validation_error = \t4.73079709517e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 5201\n",
        "training_error = \t2.69461893296e-17\n",
        "validation_error = \t2.69461893296e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 5301\n",
        "training_error = \t2.3548950945e-18\n",
        "validation_error = \t2.3548950945e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 5401\n",
        "training_error = \t4.69790812658e-18\n",
        "validation_error = \t4.69790812658e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 5501\n",
        "training_error = \t2.68276104333e-17\n",
        "validation_error = \t2.68276104333e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 5601\n",
        "training_error = \t2.35481308715e-18\n",
        "validation_error = \t2.35481308715e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 5701\n",
        "training_error = \t4.66556964482e-18\n",
        "validation_error = \t4.66556964482e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 5801\n",
        "training_error = \t2.67108809455e-17\n",
        "validation_error = \t2.67108809455e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 5901\n",
        "training_error = \t2.35473189864e-18\n",
        "validation_error = \t2.35473189864e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 6001\n",
        "training_error = \t4.63377424795e-18\n",
        "validation_error = \t4.63377424795e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 6101\n",
        "training_error = \t2.65959678093e-17\n",
        "validation_error = \t2.65959678093e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 6201\n",
        "training_error = \t2.35465152077e-18\n",
        "validation_error = \t2.35465152077e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 6301\n",
        "training_error = \t4.60250870016e-18\n",
        "validation_error = \t4.60250870016e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 6401\n",
        "training_error = \t2.64828367874e-17\n",
        "validation_error = \t2.64828367874e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 6501\n",
        "training_error = \t2.35457194541e-18\n",
        "validation_error = \t2.35457194541e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 6601\n",
        "training_error = \t4.57176209017e-18\n",
        "validation_error = \t4.57176209017e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 6701\n",
        "training_error = \t2.63714545155e-17\n",
        "validation_error = \t2.63714545155e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 6801\n",
        "training_error = \t2.35449316453e-18\n",
        "validation_error = \t2.35449316453e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 6901\n",
        "training_error = \t4.54152377788e-18\n",
        "validation_error = \t4.54152377788e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 7001\n",
        "training_error = \t2.62617895536e-17\n",
        "validation_error = \t2.62617895536e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 7101\n",
        "training_error = \t2.35441517017e-18\n",
        "validation_error = \t2.35441517017e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 7201\n",
        "training_error = \t4.51178338649e-18\n",
        "validation_error = \t4.51178338649e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 7301\n",
        "training_error = \t2.61538092245e-17\n",
        "validation_error = \t2.61538092245e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 7401\n",
        "training_error = \t2.35433795444e-18\n",
        "validation_error = \t2.35433795444e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 7501\n",
        "training_error = \t4.48253079496e-18\n",
        "validation_error = \t4.48253079496e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 7601\n",
        "training_error = \t2.60474843198e-17\n",
        "validation_error = \t2.60474843198e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 7701\n",
        "training_error = \t2.35426150956e-18\n",
        "validation_error = \t2.35426150956e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 7801\n",
        "training_error = \t4.45375613064e-18\n",
        "validation_error = \t4.45375613064e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 7901\n",
        "training_error = \t2.59427859359e-17\n",
        "validation_error = \t2.59427859359e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 8001\n",
        "training_error = \t2.35418582781e-18\n",
        "validation_error = \t2.35418582781e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 8101\n",
        "training_error = \t4.42544976217e-18\n",
        "validation_error = \t4.42544976217e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 8201\n",
        "training_error = \t2.58396823541e-17\n",
        "validation_error = \t2.58396823541e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 8301\n",
        "training_error = \t2.35411090153e-18\n",
        "validation_error = \t2.35411090153e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 8401\n",
        "training_error = \t4.39760229262e-18\n",
        "validation_error = \t4.39760229262e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 8501\n",
        "training_error = \t2.57381452854e-17\n",
        "validation_error = \t2.57381452854e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 8601\n",
        "training_error = \t2.35403672317e-18\n",
        "validation_error = \t2.35403672317e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 8701\n",
        "training_error = \t4.37020455283e-18\n",
        "validation_error = \t4.37020455283e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 8801\n",
        "training_error = \t2.56381467292e-17\n",
        "validation_error = \t2.56381467292e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 8901\n",
        "training_error = \t2.35396328523e-18\n",
        "validation_error = \t2.35396328523e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 9001\n",
        "training_error = \t4.343247595e-18\n",
        "validation_error = \t4.343247595e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 9101\n",
        "training_error = \t2.55396605154e-17\n",
        "validation_error = \t2.55396605154e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 9201\n",
        "training_error = \t2.35389058031e-18\n",
        "validation_error = \t2.35389058031e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 9301\n",
        "training_error = \t4.31672268641e-18\n",
        "validation_error = \t4.31672268641e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 9401\n",
        "training_error = \t2.54426576514e-17\n",
        "validation_error = \t2.54426576514e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 9501\n",
        "training_error = \t2.35381860106e-18\n",
        "validation_error = \t2.35381860106e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 9601\n",
        "training_error = \t4.29062130343e-18\n",
        "validation_error = \t4.29062130343e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 9701\n",
        "training_error = \t2.5347112512e-17\n",
        "validation_error = \t2.5347112512e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 9801\n",
        "training_error = \t2.35374734022e-18\n",
        "validation_error = \t2.35374734022e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 9901\n",
        "training_error = \t4.26493512567e-18\n",
        "validation_error = \t4.26493512567e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 10001\n",
        "training_error = \t2.52529997271e-17\n",
        "validation_error = \t2.52529997271e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 10101\n",
        "training_error = \t2.3536767906e-18\n",
        "validation_error = \t2.3536767906e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 10201\n",
        "training_error = \t4.23965603033e-18\n",
        "validation_error = \t4.23965603033e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 10301\n",
        "training_error = \t2.51602941776e-17\n",
        "validation_error = \t2.51602941776e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 10401\n",
        "training_error = \t2.35360694508e-18\n",
        "validation_error = \t2.35360694508e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 10501\n",
        "training_error = \t4.21477608669e-18\n",
        "validation_error = \t4.21477608669e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 10601\n",
        "training_error = \t2.50689694598e-17\n",
        "validation_error = \t2.50689694598e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 10701\n",
        "training_error = \t2.3535377966e-18\n",
        "validation_error = \t2.3535377966e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 10801\n",
        "training_error = \t4.19028755085e-18\n",
        "validation_error = \t4.19028755085e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 10901\n",
        "training_error = \t2.49790040072e-17\n",
        "validation_error = \t2.49790040072e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 11001\n",
        "training_error = \t2.35346933821e-18\n",
        "validation_error = \t2.35346933821e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 11101\n",
        "training_error = \t4.16618286056e-18\n",
        "validation_error = \t4.16618286056e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 11201\n",
        "training_error = \t2.48903703681e-17\n",
        "validation_error = \t2.48903703681e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 11301\n",
        "training_error = \t2.35340156298e-18\n",
        "validation_error = \t2.35340156298e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 11401\n",
        "training_error = \t4.14245463026e-18\n",
        "validation_error = \t4.14245463026e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 11501\n",
        "training_error = \t2.48030489559e-17\n",
        "validation_error = \t2.48030489559e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 11601\n",
        "training_error = \t2.35333446408e-18\n",
        "validation_error = \t2.35333446408e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 11701\n",
        "training_error = \t4.11909564624e-18\n",
        "validation_error = \t4.11909564624e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 11801\n",
        "training_error = \t2.47170143029e-17\n",
        "validation_error = \t2.47170143029e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 11901\n",
        "training_error = \t2.35326803475e-18\n",
        "validation_error = \t2.35326803475e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 12001\n",
        "training_error = \t4.096098862e-18\n",
        "validation_error = \t4.096098862e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 12101\n",
        "training_error = \t2.46322457251e-17\n",
        "validation_error = \t2.46322457251e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 12201\n",
        "training_error = \t2.35320226827e-18\n",
        "validation_error = \t2.35320226827e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 12301\n",
        "training_error = \t4.07345739365e-18\n",
        "validation_error = \t4.07345739365e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 12401\n",
        "training_error = \t2.45487212235e-17\n",
        "validation_error = \t2.45487212235e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 12501\n",
        "training_error = \t2.35313715803e-18\n",
        "validation_error = \t2.35313715803e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 12601\n",
        "training_error = \t4.05116451559e-18\n",
        "validation_error = \t4.05116451559e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 12701\n",
        "training_error = \t2.44664174954e-17\n",
        "validation_error = \t2.44664174954e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 12801\n",
        "training_error = \t2.35307269745e-18\n",
        "validation_error = \t2.35307269745e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 12901\n",
        "training_error = \t4.02921365623e-18\n",
        "validation_error = \t4.02921365623e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 13001\n",
        "training_error = \t2.43853159715e-17\n",
        "validation_error = \t2.43853159715e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 13101\n",
        "training_error = \t2.35300888004e-18\n",
        "validation_error = \t2.35300888004e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 13201\n",
        "training_error = \t4.00759839386e-18\n",
        "validation_error = \t4.00759839386e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 13301\n",
        "training_error = \t2.43053937471e-17\n",
        "validation_error = \t2.43053937471e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 13401\n",
        "training_error = \t2.35294569936e-18\n",
        "validation_error = \t2.35294569936e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 13501\n",
        "training_error = \t3.98631245263e-18\n",
        "validation_error = \t3.98631245263e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 13601\n",
        "training_error = \t2.42266326263e-17\n",
        "validation_error = \t2.42266326263e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 13701\n",
        "training_error = \t2.35288314904e-18\n",
        "validation_error = \t2.35288314904e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 13801\n",
        "training_error = \t3.96534969871e-18\n",
        "validation_error = \t3.96534969871e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 13901\n",
        "training_error = \t2.41490115867e-17\n",
        "validation_error = \t2.41490115867e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 14001\n",
        "training_error = \t2.35282122277e-18\n",
        "validation_error = \t2.35282122277e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 14101\n",
        "training_error = \t3.9447041365e-18\n",
        "validation_error = \t3.9447041365e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 14201\n",
        "training_error = \t2.40725127847e-17\n",
        "validation_error = \t2.40725127847e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 14301\n",
        "training_error = \t2.35275991432e-18\n",
        "validation_error = \t2.35275991432e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 14401\n",
        "training_error = \t3.92436990499e-18\n",
        "validation_error = \t3.92436990499e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 14501\n",
        "training_error = \t2.39971155549e-17\n",
        "validation_error = \t2.39971155549e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 14601\n",
        "training_error = \t2.3526992175e-18\n",
        "validation_error = \t2.3526992175e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 14701\n",
        "training_error = \t3.90434127421e-18\n",
        "validation_error = \t3.90434127421e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 14801\n",
        "training_error = \t2.39228009041e-17\n",
        "validation_error = \t2.39228009041e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 14901\n",
        "training_error = \t2.3526391262e-18\n",
        "validation_error = \t2.3526391262e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 15001\n",
        "training_error = \t3.88461264183e-18\n",
        "validation_error = \t3.88461264183e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 15101\n",
        "training_error = \t2.38495529813e-17\n",
        "validation_error = \t2.38495529813e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 15201\n",
        "training_error = \t2.35257963435e-18\n",
        "validation_error = \t2.35257963435e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 15301\n",
        "training_error = \t3.86517852977e-18\n",
        "validation_error = \t3.86517852977e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 15401\n",
        "training_error = \t2.37773516261e-17\n",
        "validation_error = \t2.37773516261e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 15501\n",
        "training_error = \t2.35252073598e-18\n",
        "validation_error = \t2.35252073598e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 15601\n",
        "training_error = \t3.84603358103e-18\n",
        "validation_error = \t3.84603358103e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 15701\n",
        "training_error = \t2.37061812972e-17\n",
        "validation_error = \t2.37061812972e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 15801\n",
        "training_error = \t2.35246242513e-18\n",
        "validation_error = \t2.35246242513e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 15901\n",
        "training_error = \t3.82717255651e-18\n",
        "validation_error = \t3.82717255651e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 16001\n",
        "training_error = \t2.36360221538e-17\n",
        "validation_error = \t2.36360221538e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 16101\n",
        "training_error = \t2.35240469594e-18\n",
        "validation_error = \t2.35240469594e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 16201\n",
        "training_error = \t3.80859033197e-18\n",
        "validation_error = \t3.80859033197e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 16301\n",
        "training_error = \t2.35668604299e-17\n",
        "validation_error = \t2.35668604299e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 16401\n",
        "training_error = \t2.35234754258e-18\n",
        "validation_error = \t2.35234754258e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 16501\n",
        "training_error = \t3.79028189511e-18\n",
        "validation_error = \t3.79028189511e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 16601\n",
        "training_error = \t2.34986765862e-17\n",
        "validation_error = \t2.34986765862e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 16701\n",
        "training_error = \t2.35229095931e-18\n",
        "validation_error = \t2.35229095931e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 16801\n",
        "training_error = \t3.77224234266e-18\n",
        "validation_error = \t3.77224234266e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 16901\n",
        "training_error = \t2.34314556604e-17\n",
        "validation_error = \t2.34314556604e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 17001\n",
        "training_error = \t2.35223494042e-18\n",
        "validation_error = \t2.35223494042e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 17101\n",
        "training_error = \t3.7544668776e-18\n",
        "validation_error = \t3.7544668776e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 17201\n",
        "training_error = \t2.33651828199e-17\n",
        "validation_error = \t2.33651828199e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 17301\n",
        "training_error = \t2.35217948027e-18\n",
        "validation_error = \t2.35217948027e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 17401\n",
        "training_error = \t3.7369508065e-18\n",
        "validation_error = \t3.7369508065e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 17501\n",
        "training_error = \t2.3299841892e-17\n",
        "validation_error = \t2.3299841892e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 17601\n",
        "training_error = \t2.35212457327e-18\n",
        "validation_error = \t2.35212457327e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 17701\n",
        "training_error = \t3.71968953683e-18\n",
        "validation_error = \t3.71968953683e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 17801\n",
        "training_error = \t2.32354168386e-17\n",
        "validation_error = \t2.32354168386e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 17901\n",
        "training_error = \t2.3520702139e-18\n",
        "validation_error = \t2.3520702139e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 18001\n",
        "training_error = \t3.70267857445e-18\n",
        "validation_error = \t3.70267857445e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 18101\n",
        "training_error = \t2.31718932162e-17\n",
        "validation_error = \t2.31718932162e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 18201\n",
        "training_error = \t2.35201639668e-18\n",
        "validation_error = \t2.35201639668e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 18301\n",
        "training_error = \t3.68591352115e-18\n",
        "validation_error = \t3.68591352115e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 18401\n",
        "training_error = \t2.31092567019e-17\n",
        "validation_error = \t2.31092567019e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 18501\n",
        "training_error = \t2.35196311618e-18\n",
        "validation_error = \t2.35196311618e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 18601\n",
        "training_error = \t3.66939007217e-18\n",
        "validation_error = \t3.66939007217e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 18701\n",
        "training_error = \t2.30474916335e-17\n",
        "validation_error = \t2.30474916335e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 18801\n",
        "training_error = \t2.35191036705e-18\n",
        "validation_error = \t2.35191036705e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 18901\n",
        "training_error = \t3.65310401396e-18\n",
        "validation_error = \t3.65310401396e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 19001\n",
        "training_error = \t2.29865853853e-17\n",
        "validation_error = \t2.29865853853e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 19101\n",
        "training_error = \t2.35185814398e-18\n",
        "validation_error = \t2.35185814398e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 19201\n",
        "training_error = \t3.63705122185e-18\n",
        "validation_error = \t3.63705122185e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 19301\n",
        "training_error = \t2.29265225308e-17\n",
        "validation_error = \t2.29265225308e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 19401\n",
        "training_error = \t2.3518064417e-18\n",
        "validation_error = \t2.3518064417e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 19501\n",
        "training_error = \t3.62122765785e-18\n",
        "validation_error = \t3.62122765785e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 19601\n",
        "training_error = \t2.28672906665e-17\n",
        "validation_error = \t2.28672906665e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 19701\n",
        "training_error = \t2.35175525501e-18\n",
        "validation_error = \t2.35175525501e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 19801\n",
        "training_error = \t3.60562936852e-18\n",
        "validation_error = \t3.60562936852e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 19901\n",
        "training_error = \t2.28088745936e-17\n",
        "validation_error = \t2.28088745936e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 20001\n",
        "training_error = \t2.35170457876e-18\n",
        "validation_error = \t2.35170457876e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 20101\n",
        "training_error = \t3.59025248289e-18\n",
        "validation_error = \t3.59025248289e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 20201\n",
        "training_error = \t2.27512635697e-17\n",
        "validation_error = \t2.27512635697e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 20301\n",
        "training_error = \t2.35165440786e-18\n",
        "validation_error = \t2.35165440786e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 20401\n",
        "training_error = \t3.57509321041e-18\n",
        "validation_error = \t3.57509321041e-18\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 20501\n",
        "training_error = \t2.26944440535e-17\n",
        "validation_error = \t2.26944440535e-17\n",
        "epoch "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 20601\n",
        "training_error = \t2.35160473723e-18\n",
        "validation_error = \t2.35160473723e-18\n"
       ]
      },
      {
       "ename": "KeyboardInterrupt",
       "evalue": "",
       "output_type": "pyerr",
       "traceback": [
        "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m\n\u001b[1;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
        "\u001b[1;32m<ipython-input-5-fe2d8a5233c5>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[0;32m      5\u001b[0m         \u001b[0mepoch_counter\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;36m0\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      6\u001b[0m         \u001b[0mLR\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;36m1e-2\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 7\u001b[1;33m         \u001b[1;32mfor\u001b[0m \u001b[0mtrn\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mval\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mtrainer\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mitertrain\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mtrain_batch\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mvalid_batch\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0malgo\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;34m'rprop'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mlearning_rate\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0miteration_size\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mbatch_size\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mBATCH_SIZE\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mvalidate_every\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mpatience\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m500\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      8\u001b[0m             \u001b[1;32mif\u001b[0m \u001b[0mepoch_counter\u001b[0m\u001b[1;33m%\u001b[0m\u001b[1;36m100\u001b[0m \u001b[1;33m==\u001b[0m \u001b[1;36m0\u001b[0m \u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      9\u001b[0m                 \u001b[1;32mif\u001b[0m \u001b[0mepoch_counter\u001b[0m \u001b[1;33m<\u001b[0m \u001b[1;36m1000\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
        "\u001b[1;32m/tmp/theanets/theanets/graph.pyc\u001b[0m in \u001b[0;36mitertrain\u001b[1;34m(self, train, valid, algo, subalgo, save_every, save_progress, **kwargs)\u001b[0m\n\u001b[0;32m    324\u001b[0m         \u001b[1;31m# train it!\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    325\u001b[0m         \u001b[0mstart\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mdatetime\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdatetime\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mutcnow\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 326\u001b[1;33m         \u001b[1;32mfor\u001b[0m \u001b[0mi\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmonitors\u001b[0m \u001b[1;32min\u001b[0m \u001b[0menumerate\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0malgo\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mitertrain\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mtrain\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mvalid\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    327\u001b[0m             \u001b[1;32myield\u001b[0m \u001b[0mmonitors\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    328\u001b[0m             \u001b[0mnow\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mdatetime\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdatetime\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mutcnow\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
        "\u001b[1;32m/tmp/theanets/theanets/trainer.pyc\u001b[0m in \u001b[0;36mitertrain\u001b[1;34m(self, train, valid, **kwargs)\u001b[0m\n\u001b[0;32m     64\u001b[0m                 \u001b[0mparams\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mnetwork\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mparams\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     65\u001b[0m                 \u001b[0mmonitor_gradients\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mget\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'monitor_gradients'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mFalse\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 66\u001b[1;33m         ).iterate(train, valid=valid, **kwargs):\n\u001b[0m\u001b[0;32m     67\u001b[0m             \u001b[1;32myield\u001b[0m \u001b[0mmonitors\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     68\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
        "\u001b[1;32m/usr/local/lib/python2.7/site-packages/downhill-0.2.0-py2.7.egg/downhill/base.pyc\u001b[0m in \u001b[0;36miterate\u001b[1;34m(self, train, valid, patience, validate_every, min_improvement, max_gradient_norm, max_gradient_elem, learning_rate, momentum, nesterov, **kwargs)\u001b[0m\n\u001b[0;32m    408\u001b[0m                 \u001b[1;32mbreak\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    409\u001b[0m             \u001b[0miteration\u001b[0m \u001b[1;33m+=\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 410\u001b[1;33m             \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_log\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mtraining\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0miteration\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    411\u001b[0m             \u001b[1;32myield\u001b[0m \u001b[0mtraining\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mvalidation\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    412\u001b[0m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mset_params\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'best'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
        "\u001b[1;32m/usr/local/lib/python2.7/site-packages/downhill-0.2.0-py2.7.egg/downhill/base.pyc\u001b[0m in \u001b[0;36m_log\u001b[1;34m(self, monitors, iteration, label, suffix)\u001b[0m\n\u001b[0;32m    224\u001b[0m         \u001b[0mlabel\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mlabel\u001b[0m \u001b[1;32mor\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m__class__\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m__name__\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    225\u001b[0m         \u001b[0mfields\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m(\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'{}={:.6f}'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mk\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mv\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mk\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mv\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mmonitors\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mitems\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 226\u001b[1;33m         \u001b[0mlogging\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0minfo\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'%s %i %s%s'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mlabel\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0miteration\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m' '\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mjoin\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mfields\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0msuffix\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    227\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    228\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mevaluate\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mdataset\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
        "\u001b[1;32m/usr/local/lib/python2.7/logging/__init__.pyc\u001b[0m in \u001b[0;36minfo\u001b[1;34m(self, msg, *args, **kwargs)\u001b[0m\n\u001b[0;32m   1157\u001b[0m         \u001b[0mlogger\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0minfo\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"Houston, we have a %s\"\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m\"interesting problem\"\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mexc_info\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1158\u001b[0m         \"\"\"\n\u001b[1;32m-> 1159\u001b[1;33m         \u001b[1;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0misEnabledFor\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mINFO\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   1160\u001b[0m             \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_log\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mINFO\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmsg\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0margs\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1161\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
        "\u001b[1;32m/usr/local/lib/python2.7/logging/__init__.pyc\u001b[0m in \u001b[0;36misEnabledFor\u001b[1;34m(self, level)\u001b[0m\n\u001b[0;32m   1351\u001b[0m         \u001b[1;32mreturn\u001b[0m \u001b[0mNOTSET\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1352\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1353\u001b[1;33m     \u001b[1;32mdef\u001b[0m \u001b[0misEnabledFor\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mlevel\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   1354\u001b[0m         \"\"\"\n\u001b[0;32m   1355\u001b[0m         \u001b[0mIs\u001b[0m \u001b[0mthis\u001b[0m \u001b[0mlogger\u001b[0m \u001b[0menabled\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mlevel\u001b[0m \u001b[1;34m'level'\u001b[0m\u001b[0;31m?\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
        "\u001b[1;31mKeyboardInterrupt\u001b[0m: "
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAEACAYAAACnJV25AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJ7eyAAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe143f8e390>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAXsAAAEACAYAAABS29YJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe141f67550>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAEACAYAAABVtcpZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nd4h/2hI6AAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe141eb1d90>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAEACAYAAABVtcpZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe141e783d0>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAEACAYAAABVtcpZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYhoCSinlYhoCSinlYhoCSinlYv8HcUmf6BSG2c4AAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe141dbf690>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAEACAYAAABVtcpZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe141d659d0>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAEACAYAAABVtcpZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nARGRDPt/qMs0KJvnQXcAAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe141ccae90>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAEACAYAAABVtcpZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe141cdb610>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAEACAYAAABVtcpZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nXwAAAABJRU5ErkJggg==\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe141b58c90>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAD7CAYAAACMlyg3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n5JiKgIhIjqkIiIjk2P8HVBcPBuGPqCwAAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe141ad5850>"
       ]
      }
     ],
     "prompt_number": 5
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print trainer.predict(valid_batch()[0])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[[  2.66849946e-09]\n",
        " [  1.00000000e+00]\n",
        " [  1.00000000e+00]\n",
        " [  2.66849946e-09]]\n"
       ]
      }
     ],
     "prompt_number": 6
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print trainer.find('hid1', 'b').get_value()\n",
      "print trainer.find('hid1', 'w').get_value()\n",
      "print trainer.find('out', 'b').get_value()\n",
      "print trainer.find('out', 'w').get_value()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[ 19.00046059  19.00046059]\n",
        "[[ 729.17617326 -728.03676278]\n",
        " [-728.03676278  729.17617326]]\n",
        "[ 58.03619787]\n",
        "[[-38.8889737]\n",
        " [-38.8889737]]\n"
       ]
      }
     ],
     "prompt_number": 7
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "plt.plot(trainer.find('hid1', 'w').get_value())"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 6,
       "text": [
        "[<matplotlib.lines.Line2D at 0x7f636512bc90>]"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEACAYAAABCl1qQAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nxcLMzCpysTAzs4pcLMzMrKL/DzjYz4MW+FY/AAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7f636f408a50>"
       ]
      }
     ],
     "prompt_number": 6
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "plt.plot([[1,-1],[-1,1]])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 9,
       "text": [
        "[<matplotlib.lines.Line2D at 0x7fe0c0710450>,\n",
        " <matplotlib.lines.Line2D at 0x7fe0c0710610>]"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAEACAYAAAC08h1NAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe0c06d1dd0>"
       ]
      }
     ],
     "prompt_number": 9
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "cp /afs/cern.ch/user/a/ajofrehe/public/network/5C-5L-500K network/5C-5L-500K"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import matplotlib.pyplot as plt\n",
      "import theanets\n",
      "import time\n",
      "import numpy as np\n",
      "from numpy import float32\n",
      "dt = np.dtype(float32)\n",
      "ncells = 100\n",
      "trainer = theanets.Experiment(theanets.feedforward.Regressor(layers=(ncells,ncells,ncells,ncells,ncells,(ncells,'logistic'))))\n",
      "net = trainer.network\n",
      "trainer.load('network/5C-5L-500K')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "ename": "ImportError",
       "evalue": "Cuda not found. Cannot unpickle CudaNdarray",
       "output_type": "pyerr",
       "traceback": [
        "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m\n\u001b[1;31mImportError\u001b[0m                               Traceback (most recent call last)",
        "\u001b[1;32m<ipython-input-3-fb8aeacd8837>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[0;32m      8\u001b[0m \u001b[0mtrainer\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mtheanets\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mExperiment\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mtheanets\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfeedforward\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mRegressor\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mlayers\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mncells\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mncells\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mncells\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mncells\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0mncells\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mncells\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'logistic'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      9\u001b[0m \u001b[0mnet\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mtrainer\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mnetwork\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 10\u001b[1;33m \u001b[0mtrainer\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mload\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'network/5C-5L-500K'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
        "\u001b[1;32m/tmp/theanets/theanets/main.pyc\u001b[0m in \u001b[0;36mload\u001b[1;34m(self, path)\u001b[0m\n\u001b[0;32m    106\u001b[0m             \u001b[1;32mfrom\u001b[0m \u001b[0mthe\u001b[0m \u001b[0mgiven\u001b[0m \u001b[0mpickle\u001b[0m \u001b[0mfile\u001b[0m\u001b[1;33m.\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    107\u001b[0m         '''\n\u001b[1;32m--> 108\u001b[1;33m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mnetwork\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mgraph\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mNetwork\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mload\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mpath\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    109\u001b[0m         \u001b[1;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mnetwork\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
        "\u001b[1;32m/tmp/theanets/theanets/graph.pyc\u001b[0m in \u001b[0;36mload\u001b[1;34m(cls, filename)\u001b[0m\n\u001b[0;32m    606\u001b[0m         \u001b[0mopener\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mgzip\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mopen\u001b[0m \u001b[1;32mif\u001b[0m \u001b[0mfilename\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mlower\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mendswith\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'.gz'\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32melse\u001b[0m \u001b[0mopen\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    607\u001b[0m         \u001b[0mhandle\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mopener\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mfilename\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'rb'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 608\u001b[1;33m         \u001b[0mmodel\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mpickle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mload\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mhandle\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    609\u001b[0m         \u001b[0mhandle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mclose\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    610\u001b[0m         \u001b[0mlogging\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0minfo\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'%s: loaded model'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfilename\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
        "\u001b[1;32m/usr/local/lib/python2.7/pickle.pyc\u001b[0m in \u001b[0;36mload\u001b[1;34m(file)\u001b[0m\n\u001b[0;32m   1376\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1377\u001b[0m \u001b[1;32mdef\u001b[0m \u001b[0mload\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mfile\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1378\u001b[1;33m     \u001b[1;32mreturn\u001b[0m \u001b[0mUnpickler\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mfile\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mload\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   1379\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1380\u001b[0m \u001b[1;32mdef\u001b[0m \u001b[0mloads\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mstr\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
        "\u001b[1;32m/usr/local/lib/python2.7/pickle.pyc\u001b[0m in \u001b[0;36mload\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m    856\u001b[0m             \u001b[1;32mwhile\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    857\u001b[0m                 \u001b[0mkey\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mread\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 858\u001b[1;33m                 \u001b[0mdispatch\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mkey\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    859\u001b[0m         \u001b[1;32mexcept\u001b[0m \u001b[0m_Stop\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mstopinst\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    860\u001b[0m             \u001b[1;32mreturn\u001b[0m \u001b[0mstopinst\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mvalue\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
        "\u001b[1;32m/usr/local/lib/python2.7/pickle.pyc\u001b[0m in \u001b[0;36mload_reduce\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m   1131\u001b[0m         \u001b[0margs\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mstack\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mpop\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1132\u001b[0m         \u001b[0mfunc\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mstack\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;33m-\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1133\u001b[1;33m         \u001b[0mvalue\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mfunc\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m*\u001b[0m\u001b[0margs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   1134\u001b[0m         \u001b[0mstack\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;33m-\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mvalue\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1135\u001b[0m     \u001b[0mdispatch\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mREDUCE\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mload_reduce\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
        "\u001b[1;32m/usr/local/lib/python2.7/site-packages/theano/sandbox/cuda/type.pyc\u001b[0m in \u001b[0;36mCudaNdarray_unpickler\u001b[1;34m(npa)\u001b[0m\n\u001b[0;32m    544\u001b[0m         \u001b[1;32mreturn\u001b[0m \u001b[0mcuda\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mCudaNdarray\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mnpa\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    545\u001b[0m     \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 546\u001b[1;33m         \u001b[1;32mraise\u001b[0m \u001b[0mImportError\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"Cuda not found. Cannot unpickle CudaNdarray\"\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    547\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    548\u001b[0m \u001b[0mcopy_reg\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mconstructor\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mCudaNdarray_unpickler\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
        "\u001b[1;31mImportError\u001b[0m: Cuda not found. Cannot unpickle CudaNdarray"
       ]
      }
     ],
     "prompt_number": 3
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "cd .."
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "/data/notebooks/Arash\n"
       ]
      }
     ],
     "prompt_number": 6
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "ls"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "100K3L                  500K4L                 deep-1L1000N50K-C\r\n",
        "10by10-1L-1000N100K     500K4L-relu-rprop      deep-1L-500N50K\r\n",
        "10by10-1L-2000N200K     500K4L-rprop           deep-2L-1000-100N100K\r\n",
        "10by10-1L-2000N400K     500K4L-rprop-adadelta  deep-2L1000N1M\r\n",
        "10by10-2L-100K          50K2L-rprop            deep-2L1000N50K\r\n",
        "15by15-1L-2000N130K     5C-5L-500K             deep-2L-100-1000N100K\r\n",
        "15by15-1L-3KN20K-batch  deep-1L10000N100K      deep-maxout-1L1000N50K\r\n",
        "1M5L-rprop              deep-1L10000N50K       filename\r\n",
        "200K3L-relu-rprop       deep-1L1000N100K       XOR_Data.npy\r\n",
        "20K1L-relu-rprop        deep-1L1000N1M         XOR_Input.npy\r\n",
        "500K3L                  deep-1L1000N50K        XOR_Output.npy\r\n"
       ]
      }
     ],
     "prompt_number": 5
    }
   ],
   "metadata": {}
  }
 ]
}