2017-1-10(1).ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 更改了maxlen设计的函数，使得可以在main.py中直接定义maxlne的数值"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/M2015eliu/.pyenv/versions/anaconda2-4.1.1/lib/python2.7/site-packages/gensim/utils.py:1015: UserWarning: Pattern library is not installed, lemmatization won't be available.\n",
      "  warnings.warn(\"Pattern library is not installed, lemmatization won't be available.\")\n",
      "Using gpu device 0: GeForce GTX TITAN X (CNMeM is enabled with initial size: 95.0% of memory, cuDNN 4007)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting to training the model..., saving to weights/201701051610_e20_1k1k_l200.p\n",
      "Training\n",
      "the length of the training data is  1998\n",
      "Before trianing, the error is:\n",
      "Training error: 0.256023526192\n",
      "Training_correct error 0.333335667849\n",
      "Testing_correct error: 0.342972397804\n",
      "top- 1 = [1] : 1\n",
      "top- 5 = [17] : 17\n",
      "top- 10 = [26] : 26\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean     458.362362\n",
      "std      279.827936\n",
      "min        1.000000\n",
      "25%      224.000000\n",
      "50%      461.000000\n",
      "75%      693.750000\n",
      "max      978.000000\n",
      "dtype: float64\n",
      "top- 1 = [2] : 2\n",
      "top- 5 = [16] : 16\n",
      "top- 10 = [29] : 29\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean      475.959000\n",
      "std       282.316227\n",
      "min         1.000000\n",
      "25%       237.125000\n",
      "50%       476.500000\n",
      "75%       722.375000\n",
      "max       985.000000\n",
      "dtype: float64\n",
      "\n",
      "Epoch 0 ...\n",
      "Epoch  0 Update  21 Cost  0.251657098532\n",
      "Epoch  0 Update  42 Cost  0.24139970541\n",
      "Epoch  0 Update  63 Cost  0.272840440273\n",
      "Training error: 0.221390083432\n",
      "Training_correct error 0.190529316664\n",
      "Testing_correct error: 0.193975538015\n",
      "top- 1 = [1, 8] : 8\n",
      "top- 5 = [17, 37] : 37\n",
      "top- 10 = [26, 70] : 70\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean     254.861862\n",
      "std      216.797657\n",
      "min        1.000000\n",
      "25%       75.500000\n",
      "50%      194.000000\n",
      "75%      397.000000\n",
      "max      903.500000\n",
      "dtype: float64\n",
      "top- 1 = [2, 5] : 5\n",
      "top- 5 = [16, 23] : 23\n",
      "top- 10 = [29, 50] : 50\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean      291.273000\n",
      "std       235.115686\n",
      "min         1.000000\n",
      "25%        96.875000\n",
      "50%       229.500000\n",
      "75%       451.250000\n",
      "max       963.000000\n",
      "dtype: float64\n",
      "epoch took: 9.56556606293\n",
      "\n",
      "Epoch 1 ...\n",
      "Epoch  1 Update  84 Cost  0.173480719328\n",
      "Epoch  1 Update  105 Cost  0.227109074593\n",
      "Epoch  1 Update  126 Cost  0.209477975965\n",
      "Training error: 0.235600933433\n",
      "Training_correct error 0.153468713164\n",
      "Testing_correct error: 0.155288174748\n",
      "top- 1 = [1, 8, 4] : 4\n",
      "top- 5 = [17, 37, 35] : 35\n",
      "top- 10 = [26, 70, 67] : 67\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean     298.236236\n",
      "std      232.106766\n",
      "min        1.000000\n",
      "25%       95.500000\n",
      "50%      263.000000\n",
      "75%      471.000000\n",
      "max      897.500000\n",
      "dtype: float64\n",
      "top- 1 = [2, 5, 3] : 3\n",
      "top- 5 = [16, 23, 37] : 37\n",
      "top- 10 = [29, 50, 72] : 72\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean      317.561000\n",
      "std       241.350061\n",
      "min         1.000000\n",
      "25%       107.750000\n",
      "50%       291.000000\n",
      "75%       472.125000\n",
      "max       952.000000\n",
      "dtype: float64\n",
      "epoch took: 9.61420798302\n",
      "\n",
      "Epoch 2 ...\n",
      "Epoch  2 Update  147 Cost  0.152037858963\n",
      "Epoch  2 Update  168 Cost  0.211680442095\n",
      "Epoch  2 Update  189 Cost  0.233046606183\n",
      "Training error: 0.187974974513\n",
      "Training_correct error 0.242754876614\n",
      "Testing_correct error: 0.231702938676\n",
      "top- 1 = [1, 8, 4, 13] : 13\n",
      "top- 5 = [17, 37, 35, 61] : 61\n",
      "top- 10 = [26, 70, 67, 113] : 113\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean     174.578579\n",
      "std      177.242147\n",
      "min        1.000000\n",
      "25%       35.750000\n",
      "50%      115.000000\n",
      "75%      257.000000\n",
      "max      848.500000\n",
      "dtype: float64\n",
      "top- 1 = [2, 5, 3, 15] : 15\n",
      "top- 5 = [16, 23, 37, 63] : 63\n",
      "top- 10 = [29, 50, 72, 120] : 120\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.00000\n",
      "mean      216.51600\n",
      "std       200.48973\n",
      "min         1.00000\n",
      "25%        47.50000\n",
      "50%       154.75000\n",
      "75%       349.12500\n",
      "max       879.00000\n",
      "dtype: float64\n",
      "epoch took: 9.51830220222\n",
      "\n",
      "Epoch 3 ...\n",
      "Epoch  3 Update  210 Cost  0.128209650517\n",
      "Epoch  3 Update  231 Cost  0.189863979816\n",
      "Epoch  3 Update  252 Cost  0.143652036786\n",
      "Training error: 0.209973186255\n",
      "Training_correct error 0.144986107945\n",
      "Testing_correct error: 0.158282533288\n",
      "top- 1 = [1, 8, 4, 13, 6] : 6\n",
      "top- 5 = [17, 37, 35, 61, 33] : 33\n",
      "top- 10 = [26, 70, 67, 113, 70] : 70\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean     223.489489\n",
      "std      191.909551\n",
      "min        1.000000\n",
      "25%       65.000000\n",
      "50%      174.000000\n",
      "75%      331.500000\n",
      "max      888.000000\n",
      "dtype: float64\n",
      "top- 1 = [2, 5, 3, 15, 5] : 5\n",
      "top- 5 = [16, 23, 37, 63, 27] : 27\n",
      "top- 10 = [29, 50, 72, 120, 52] : 52\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean      261.356000\n",
      "std       219.097853\n",
      "min         1.000000\n",
      "25%        79.000000\n",
      "50%       199.250000\n",
      "75%       403.000000\n",
      "max       930.000000\n",
      "dtype: float64\n",
      "epoch took: 9.54189419746\n",
      "\n",
      "Epoch 4 ...\n",
      "Epoch  4 Update  273 Cost  0.199347779155\n",
      "Epoch  4 Update  294 Cost  0.297611951828\n",
      "Epoch  4 Update  315 Cost  0.225044250488\n",
      "Training error: 0.164968177676\n",
      "Training_correct error 0.234963208437\n",
      "Testing_correct error: 0.219791069627\n",
      "top- 1 = [1, 8, 4, 13, 6, 8] : 8\n",
      "top- 5 = [17, 37, 35, 61, 33, 67] : 67\n",
      "top- 10 = [26, 70, 67, 113, 70, 125] : 125\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean     119.911912\n",
      "std      120.956532\n",
      "min        1.000000\n",
      "25%       25.000000\n",
      "50%       79.000000\n",
      "75%      183.750000\n",
      "max      724.500000\n",
      "dtype: float64\n",
      "top- 1 = [2, 5, 3, 15, 5, 16] : 16\n",
      "top- 5 = [16, 23, 37, 63, 27, 73] : 73\n",
      "top- 10 = [29, 50, 72, 120, 52, 126] : 126\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean      141.915000\n",
      "std       146.831868\n",
      "min         1.000000\n",
      "25%        29.375000\n",
      "50%        91.000000\n",
      "75%       206.375000\n",
      "max       714.000000\n",
      "dtype: float64\n",
      "epoch took: 9.53586483002\n",
      "\n",
      "Epoch 5 ...\n",
      "Epoch  5 Update  336 Cost  0.113318875432\n",
      "Epoch  5 Update  357 Cost  0.128445804119\n",
      "Epoch  5 Update  378 Cost  0.139468535781\n",
      "Training error: 0.138220712543\n",
      "Training_correct error 0.0922933667898\n",
      "Testing_correct error: 0.0988744050264\n",
      "top- 1 = [1, 8, 4, 13, 6, 8, 20] : 20\n",
      "top- 5 = [17, 37, 35, 61, 33, 67, 90] : 90\n",
      "top- 10 = [26, 70, 67, 113, 70, 125, 140] : 140\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      92.508509\n",
      "std       87.854111\n",
      "min        1.000000\n",
      "25%       22.000000\n",
      "50%       65.000000\n",
      "75%      143.500000\n",
      "max      456.500000\n",
      "dtype: float64\n",
      "top- 1 = [2, 5, 3, 15, 5, 16, 13] : 13\n",
      "top- 5 = [16, 23, 37, 63, 27, 73, 67] : 67\n",
      "top- 10 = [29, 50, 72, 120, 52, 126, 118] : 118\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean      118.969000\n",
      "std       119.504093\n",
      "min         1.000000\n",
      "25%        29.500000\n",
      "50%        75.500000\n",
      "75%       173.500000\n",
      "max       720.000000\n",
      "dtype: float64\n",
      "epoch took: 9.53620696068\n",
      "\n",
      "Epoch 6 ...\n",
      "Epoch  6 Update  399 Cost  0.08879083395\n",
      "Epoch  6 Update  420 Cost  0.0998856425285\n",
      "Epoch  6 Update  441 Cost  0.205095753074\n",
      "Training error: 0.13291439414\n",
      "Training_correct error 0.156636953354\n",
      "Testing_correct error: 0.18975739181\n",
      "top- 1 = [1, 8, 4, 13, 6, 8, 20, 11] : 11\n",
      "top- 5 = [17, 37, 35, 61, 33, 67, 90, 81] : 81\n",
      "top- 10 = [26, 70, 67, 113, 70, 125, 140, 157] : 157\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.00000\n",
      "mean      92.85986\n",
      "std       86.06248\n",
      "min        1.00000\n",
      "25%       24.00000\n",
      "50%       71.00000\n",
      "75%      134.00000\n",
      "max      444.00000\n",
      "dtype: float64\n",
      "top- 1 = [2, 5, 3, 15, 5, 16, 13, 16] : 16\n",
      "top- 5 = [16, 23, 37, 63, 27, 73, 67, 76] : 76\n",
      "top- 10 = [29, 50, 72, 120, 52, 126, 118, 124] : 124\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean      128.016000\n",
      "std       120.254555\n",
      "min         1.000000\n",
      "25%        29.750000\n",
      "50%        88.000000\n",
      "75%       196.250000\n",
      "max       620.000000\n",
      "dtype: float64\n",
      "epoch took: 9.54475307465\n",
      "\n",
      "Epoch 7 ...\n",
      "Epoch  7 Update  462 Cost  0.158953100443\n",
      "Epoch  7 Update  483 Cost  0.150326967239\n",
      "Epoch  7 Update  504 Cost  0.208725348115\n",
      "Training error: 0.149434939027\n",
      "Training_correct error 0.0807736217976\n",
      "Testing_correct error: 0.0905024632812\n",
      "top- 1 = [1, 8, 4, 13, 6, 8, 20, 11, 18] : 18\n",
      "top- 5 = [17, 37, 35, 61, 33, 67, 90, 81, 79] : 79\n",
      "top- 10 = [26, 70, 67, 113, 70, 125, 140, 157, 151] : 151\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean     110.995996\n",
      "std      112.062444\n",
      "min        1.000000\n",
      "25%       21.500000\n",
      "50%       73.000000\n",
      "75%      169.500000\n",
      "max      543.000000\n",
      "dtype: float64\n",
      "top- 1 = [2, 5, 3, 15, 5, 16, 13, 16, 13] : 13\n",
      "top- 5 = [16, 23, 37, 63, 27, 73, 67, 76, 85] : 85\n",
      "top- 10 = [29, 50, 72, 120, 52, 126, 118, 124, 138] : 138\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean      159.422000\n",
      "std       159.991644\n",
      "min         1.000000\n",
      "25%        26.500000\n",
      "50%       100.750000\n",
      "75%       252.500000\n",
      "max       725.000000\n",
      "dtype: float64\n",
      "epoch took: 9.53625798225\n",
      "\n",
      "Epoch 8 ...\n",
      "Epoch  8 Update  525 Cost  0.0983357876539\n",
      "Epoch  8 Update  546 Cost  0.123154610395\n",
      "Epoch  8 Update  567 Cost  0.103565230966\n",
      "Training error: 0.110778510571\n",
      "Training_correct error 0.0948333144188\n",
      "Testing_correct error: 0.103301428258\n",
      "top- 1 = [1, 8, 4, 13, 6, 8, 20, 11, 18, 22] : 22\n",
      "top- 5 = [17, 37, 35, 61, 33, 67, 90, 81, 79, 110] : 110\n",
      "top- 10 = [26, 70, 67, 113, 70, 125, 140, 157, 151, 181] : 181\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      72.652653\n",
      "std       74.072055\n",
      "min        1.000000\n",
      "25%       15.000000\n",
      "50%       48.000000\n",
      "75%      105.000000\n",
      "max      435.000000\n",
      "dtype: float64\n",
      "top- 1 = [2, 5, 3, 15, 5, 16, 13, 16, 13, 17] : 17\n",
      "top- 5 = [16, 23, 37, 63, 27, 73, 67, 76, 85, 95] : 95\n",
      "top- 10 = [29, 50, 72, 120, 52, 126, 118, 124, 138, 162] : 162\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean      105.217000\n",
      "std       110.803052\n",
      "min         1.000000\n",
      "25%        19.000000\n",
      "50%        65.500000\n",
      "75%       155.000000\n",
      "max       638.000000\n",
      "dtype: float64\n",
      "epoch took: 9.56129789352\n",
      "\n",
      "Epoch 9 ...\n",
      "Epoch  9 Update  588 Cost  0.123290605843\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0mTraceback (most recent call last)",
      "\u001b[0;32m/home/M2015eliu/cas/2017.1.1~LiuSTM/main.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m    323\u001b[0m                 \u001b[0mdir_file\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m\"weights/201701051610_e20_1k1k_l200.p\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    324\u001b[0m                 \u001b[0;32mprint\u001b[0m \u001b[0;34m\"Starting to training the model..., saving to\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdir_file\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 325\u001b[0;31m                 \u001b[0msls\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlstm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mLSTM\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdir_file\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmaxlen\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mload\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtraining\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    326\u001b[0m                 \u001b[0msls\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain_lstm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrain\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m20\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtrain_1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtest_1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    327\u001b[0m                 \u001b[0msls\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msave_model\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/M2015eliu/cas/2017.1.1~LiuSTM/lstm.pyc\u001b[0m in \u001b[0;36mtrain_lstm\u001b[0;34m(self, train, max_epochs, correct, test_correct)\u001b[0m\n\u001b[1;32m    524\u001b[0m                 \u001b[0memb1\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mswapaxes\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrconv\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    525\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 526\u001b[0;31m                 \u001b[0mcst\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mf_grad_shared\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0memb2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmas2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0memb1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmas1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    527\u001b[0m                 \u001b[0ms\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mf_update\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlrate\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m# Not USED ???\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    528\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/M2015eliu/.pyenv/versions/anaconda2-4.1.1/lib/python2.7/site-packages/theano/compile/function_module.pyc\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m    857\u001b[0m         \u001b[0mt0_fn\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtime\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    858\u001b[0m         \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 859\u001b[0;31m             \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    860\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    861\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mhasattr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'position_of_error'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/M2015eliu/.pyenv/versions/anaconda2-4.1.1/lib/python2.7/site-packages/theano/scan_module/scan_op.pyc\u001b[0m in \u001b[0;36mrval\u001b[0;34m(p, i, o, n, allow_gc)\u001b[0m\n\u001b[1;32m    949\u001b[0m         def rval(p=p, i=node_input_storage, o=node_output_storage, n=node,\n\u001b[1;32m    950\u001b[0m                  allow_gc=allow_gc):\n\u001b[0;32m--> 951\u001b[0;31m             \u001b[0mr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mo\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    952\u001b[0m             \u001b[0;32mfor\u001b[0m \u001b[0mo\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mnode\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moutputs\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    953\u001b[0m                 \u001b[0mcompute_map\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mo\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/M2015eliu/.pyenv/versions/anaconda2-4.1.1/lib/python2.7/site-packages/theano/scan_module/scan_op.pyc\u001b[0m in \u001b[0;36m<lambda>\u001b[0;34m(node, args, outs)\u001b[0m\n\u001b[1;32m    938\u001b[0m                         \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    939\u001b[0m                         \u001b[0mouts\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 940\u001b[0;31m                         self, node)\n\u001b[0m\u001b[1;32m    941\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mImportError\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtheano\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgof\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcmodule\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mMissingGXX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    942\u001b[0m             \u001b[0mp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexecute\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "# maxlen = 200\n",
    "run main.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/M2015eliu/.pyenv/versions/anaconda2-4.1.1/lib/python2.7/site-packages/gensim/utils.py:1015: UserWarning: Pattern library is not installed, lemmatization won't be available.\n",
      "  warnings.warn(\"Pattern library is not installed, lemmatization won't be available.\")\n",
      "Using gpu device 0: GeForce GTX TITAN X (CNMeM is enabled with initial size: 95.0% of memory, cuDNN 4007)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('Now the maxlen =', 0)\n",
      "Starting to training the model..., saving to weights/201701051610_e20_1k1k_l200.p\n",
      "Training\n",
      "the length of the training data is  1998\n",
      "Before trianing, the error is:\n",
      "Training error: 0.257755994797\n",
      "Training_correct error 0.337905704975\n",
      "Testing_correct error: 0.335038632154\n",
      "top- 1 = [3] : 3\n",
      "top- 5 = [21] : 21\n",
      "top- 10 = [30] : 30\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean     501.754755\n",
      "std      264.648264\n",
      "min        1.000000\n",
      "25%      290.500000\n",
      "50%      541.500000\n",
      "75%      711.000000\n",
      "max      978.000000\n",
      "dtype: float64\n",
      "top- 1 = [1] : 1\n",
      "top- 5 = [12] : 12\n",
      "top- 10 = [24] : 24\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean      551.164000\n",
      "std       260.994743\n",
      "min         1.000000\n",
      "25%       362.625000\n",
      "50%       559.000000\n",
      "75%       775.500000\n",
      "max       985.000000\n",
      "dtype: float64\n",
      "\n",
      "Epoch 0 ...\n",
      "Epoch  0 Update  21 Cost  0.182155281305\n",
      "Epoch  0 Update  42 Cost  0.167487800121\n",
      "Epoch  0 Update  63 Cost  0.207024231553\n",
      "Training error: 0.125692605972\n",
      "Training_correct error 0.118412218988\n",
      "Testing_correct error: 0.114850081503\n",
      "top- 1 = [3, 48] : 48\n",
      "top- 5 = [21, 199] : 199\n",
      "top- 10 = [30, 304] : 304\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      52.060060\n",
      "std       68.136176\n",
      "min        1.000000\n",
      "25%        7.500000\n",
      "50%       28.000000\n",
      "75%       63.000000\n",
      "max      370.500000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46] : 46\n",
      "top- 5 = [12, 151] : 151\n",
      "top- 10 = [24, 232] : 232\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       61.457000\n",
      "std        73.761911\n",
      "min         1.000000\n",
      "25%        11.500000\n",
      "50%        34.000000\n",
      "75%        85.000000\n",
      "max       448.000000\n",
      "dtype: float64\n",
      "epoch took: 47.7021028996\n",
      "\n",
      "Epoch 1 ...\n",
      "Epoch  1 Update  84 Cost  0.10254971683\n",
      "Epoch  1 Update  105 Cost  0.11504471302\n",
      "Epoch  1 Update  126 Cost  0.0893101021647\n",
      "Training error: 0.117395535111\n",
      "Training_correct error 0.0951170623302\n",
      "Testing_correct error: 0.0903676599264\n",
      "top- 1 = [3, 48, 50] : 50\n",
      "top- 5 = [21, 199, 208] : 208\n",
      "top- 10 = [30, 304, 320] : 320\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      52.107608\n",
      "std       67.633366\n",
      "min        1.000000\n",
      "25%        7.000000\n",
      "50%       22.000000\n",
      "75%       72.500000\n",
      "max      324.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43] : 43\n",
      "top- 5 = [12, 151, 169] : 169\n",
      "top- 10 = [24, 232, 267] : 267\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       54.447000\n",
      "std        69.017665\n",
      "min         1.000000\n",
      "25%         9.500000\n",
      "50%        30.500000\n",
      "75%        61.250000\n",
      "max       425.000000\n",
      "dtype: float64\n",
      "epoch took: 48.3495190144\n",
      "\n",
      "Epoch 2 ...\n",
      "Epoch  2 Update  147 Cost  0.0750384479761\n",
      "Epoch  2 Update  168 Cost  0.0610469020903\n",
      "Epoch  2 Update  189 Cost  0.124674670398\n",
      "Training error: 0.0985486656427\n",
      "Training_correct error 0.101944170892\n",
      "Testing_correct error: 0.102379269898\n",
      "top- 1 = [3, 48, 50, 86] : 86\n",
      "top- 5 = [21, 199, 208, 293] : 293\n",
      "top- 10 = [30, 304, 320, 431] : 431\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      38.222723\n",
      "std       53.008296\n",
      "min        1.000000\n",
      "25%        4.000000\n",
      "50%       15.000000\n",
      "75%       50.250000\n",
      "max      398.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81] : 81\n",
      "top- 5 = [12, 151, 169, 246] : 246\n",
      "top- 10 = [24, 232, 267, 378] : 378\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       44.222000\n",
      "std        56.271624\n",
      "min         1.000000\n",
      "25%         5.500000\n",
      "50%        17.500000\n",
      "75%        62.000000\n",
      "max       432.000000\n",
      "dtype: float64\n",
      "epoch took: 47.8764941692\n",
      "\n",
      "Epoch 3 ...\n",
      "Epoch  3 Update  210 Cost  0.0899893343449\n",
      "Epoch  3 Update  231 Cost  0.0913775414228\n",
      "Epoch  3 Update  252 Cost  0.0783319026232\n",
      "Training error: 0.0997497886419\n",
      "Training_correct error 0.0821037888527\n",
      "Testing_correct error: 0.083479039371\n",
      "top- 1 = [3, 48, 50, 86, 76] : 76\n",
      "top- 5 = [21, 199, 208, 293, 310] : 310\n",
      "top- 10 = [30, 304, 320, 431, 411] : 411\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      37.868368\n",
      "std       57.583045\n",
      "min        1.000000\n",
      "25%        3.500000\n",
      "50%       16.000000\n",
      "75%       42.000000\n",
      "max      378.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66] : 66\n",
      "top- 5 = [12, 151, 169, 246, 242] : 242\n",
      "top- 10 = [24, 232, 267, 378, 344] : 344\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       43.143000\n",
      "std        60.577844\n",
      "min         1.000000\n",
      "25%         6.000000\n",
      "50%        21.250000\n",
      "75%        46.000000\n",
      "max       417.000000\n",
      "dtype: float64\n",
      "epoch took: 48.1992940903\n",
      "\n",
      "Epoch 4 ...\n",
      "Epoch  4 Update  273 Cost  0.102640353143\n",
      "Epoch  4 Update  294 Cost  0.0877730324864\n",
      "Epoch  4 Update  315 Cost  0.119521379471\n",
      "Training error: 0.086856238544\n",
      "Training_correct error 0.107121005654\n",
      "Testing_correct error: 0.120220825076\n",
      "top- 1 = [3, 48, 50, 86, 76, 118] : 118\n",
      "top- 5 = [21, 199, 208, 293, 310, 392] : 392\n",
      "top- 10 = [30, 304, 320, 431, 411, 533] : 533\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      24.363864\n",
      "std       37.543915\n",
      "min        1.000000\n",
      "25%        3.000000\n",
      "50%        9.000000\n",
      "75%       26.500000\n",
      "max      279.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92] : 92\n",
      "top- 5 = [12, 151, 169, 246, 242, 285] : 285\n",
      "top- 10 = [24, 232, 267, 378, 344, 455] : 455\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       35.453000\n",
      "std        50.465682\n",
      "min         1.000000\n",
      "25%         4.000000\n",
      "50%        12.500000\n",
      "75%        41.250000\n",
      "max       324.000000\n",
      "dtype: float64\n",
      "epoch took: 49.2182090282\n",
      "\n",
      "Epoch 5 ...\n",
      "Epoch  5 Update  336 Cost  0.0586893036962\n",
      "Epoch  5 Update  357 Cost  0.0592846162617\n",
      "Epoch  5 Update  378 Cost  0.172971352935\n",
      "Training error: 0.0897238701582\n",
      "Training_correct error 0.0772184580564\n",
      "Testing_correct error: 0.0757179409266\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93] : 93\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359] : 359\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495] : 495\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      28.575576\n",
      "std       44.863448\n",
      "min        1.000000\n",
      "25%        3.000000\n",
      "50%       10.500000\n",
      "75%       31.000000\n",
      "max      277.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79] : 79\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279] : 279\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443] : 443\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       35.241500\n",
      "std        50.618402\n",
      "min         1.000000\n",
      "25%         4.500000\n",
      "50%        13.750000\n",
      "75%        41.000000\n",
      "max       273.000000\n",
      "dtype: float64\n",
      "epoch took: 49.8195390701\n",
      "\n",
      "Epoch 6 ...\n",
      "Epoch  6 Update  399 Cost  0.0754013508558\n",
      "Epoch  6 Update  420 Cost  0.0876249596477\n",
      "Epoch  6 Update  441 Cost  0.165213420987\n",
      "Training error: 0.0925588086247\n",
      "Training_correct error 0.075753249228\n",
      "Testing_correct error: 0.0882906317711\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92] : 92\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406] : 406\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582] : 582\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      24.823824\n",
      "std       42.170965\n",
      "min        1.000000\n",
      "25%        3.000000\n",
      "50%        7.500000\n",
      "75%       21.000000\n",
      "max      344.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80] : 80\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306] : 306\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470] : 470\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       30.191000\n",
      "std        47.151795\n",
      "min         1.000000\n",
      "25%         4.000000\n",
      "50%        11.000000\n",
      "75%        25.000000\n",
      "max       333.000000\n",
      "dtype: float64\n",
      "epoch took: 50.368283987\n",
      "\n",
      "Epoch 7 ...\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0mTraceback (most recent call last)",
      "\u001b[0;32m/home/M2015eliu/cas/2017.1.1~LiuSTM/main.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m    324\u001b[0m                 \u001b[0mdir_file\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m\"weights/201701102308_e40_1k1k_l0.p\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    325\u001b[0m                 \u001b[0;32mprint\u001b[0m \u001b[0;34m\"Starting to training the model..., saving to\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdir_file\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 326\u001b[0;31m                 \u001b[0msls\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlstm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mLSTM\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdir_file\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmaxlen\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mload\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtraining\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    327\u001b[0m                 \u001b[0msls\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain_lstm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrain\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mepoch\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtrain_1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtest_1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    328\u001b[0m                 \u001b[0msls\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msave_model\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/M2015eliu/cas/2017.1.1~LiuSTM/lstm.pyc\u001b[0m in \u001b[0;36mtrain_lstm\u001b[0;34m(self, train, max_epochs, correct, test_correct)\u001b[0m\n\u001b[1;32m    524\u001b[0m                 \u001b[0memb1\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mswapaxes\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrconv\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    525\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 526\u001b[0;31m                 \u001b[0mcst\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mf_grad_shared\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0memb2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmas2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0memb1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmas1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    527\u001b[0m                 \u001b[0ms\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mf_update\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlrate\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m# Not USED ???\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    528\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/M2015eliu/.pyenv/versions/anaconda2-4.1.1/lib/python2.7/site-packages/theano/compile/function_module.pyc\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m    857\u001b[0m         \u001b[0mt0_fn\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtime\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    858\u001b[0m         \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 859\u001b[0;31m             \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    860\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    861\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mhasattr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'position_of_error'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/M2015eliu/.pyenv/versions/anaconda2-4.1.1/lib/python2.7/site-packages/theano/scan_module/scan_op.pyc\u001b[0m in \u001b[0;36mrval\u001b[0;34m(p, i, o, n, allow_gc)\u001b[0m\n\u001b[1;32m    949\u001b[0m         def rval(p=p, i=node_input_storage, o=node_output_storage, n=node,\n\u001b[1;32m    950\u001b[0m                  allow_gc=allow_gc):\n\u001b[0;32m--> 951\u001b[0;31m             \u001b[0mr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mo\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    952\u001b[0m             \u001b[0;32mfor\u001b[0m \u001b[0mo\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mnode\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moutputs\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    953\u001b[0m                 \u001b[0mcompute_map\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mo\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/M2015eliu/.pyenv/versions/anaconda2-4.1.1/lib/python2.7/site-packages/theano/scan_module/scan_op.pyc\u001b[0m in \u001b[0;36m<lambda>\u001b[0;34m(node, args, outs)\u001b[0m\n\u001b[1;32m    938\u001b[0m                         \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    939\u001b[0m                         \u001b[0mouts\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 940\u001b[0;31m                         self, node)\n\u001b[0m\u001b[1;32m    941\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mImportError\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtheano\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgof\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcmodule\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mMissingGXX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    942\u001b[0m             \u001b[0mp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexecute\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "# maxlen = 0 (shoul the max. length appeared)\n",
    "run main.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/M2015eliu/.pyenv/versions/anaconda2-4.1.1/lib/python2.7/site-packages/gensim/utils.py:1015: UserWarning: Pattern library is not installed, lemmatization won't be available.\n",
      "  warnings.warn(\"Pattern library is not installed, lemmatization won't be available.\")\n",
      "Using gpu device 0: GeForce GTX TITAN X (CNMeM is enabled with initial size: 95.0% of memory, cuDNN 4007)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('Now the maxlen =', 0)\n",
      "Starting to training the model..., saving to weights/201701102308_e40_1k1k_l0.p\n",
      "Training\n",
      "the length of the training data is  1998\n",
      "Before trianing, the error is:\n",
      "Training error: 0.257755994797\n",
      "Training_correct error 0.337905704975\n",
      "Testing_correct error: 0.335038632154\n",
      "top- 1 = [3] : 3\n",
      "top- 5 = [21] : 21\n",
      "top- 10 = [30] : 30\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean     501.754755\n",
      "std      264.648264\n",
      "min        1.000000\n",
      "25%      290.500000\n",
      "50%      541.500000\n",
      "75%      711.000000\n",
      "max      978.000000\n",
      "dtype: float64\n",
      "top- 1 = [1] : 1\n",
      "top- 5 = [12] : 12\n",
      "top- 10 = [24] : 24\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean      551.164000\n",
      "std       260.994743\n",
      "min         1.000000\n",
      "25%       362.625000\n",
      "50%       559.000000\n",
      "75%       775.500000\n",
      "max       985.000000\n",
      "dtype: float64\n",
      "\n",
      "Epoch 0 ...\n",
      "Epoch  0 Update  21 Cost  0.182155281305\n",
      "Epoch  0 Update  42 Cost  0.167487800121\n",
      "Epoch  0 Update  63 Cost  0.207024231553\n",
      "Training error: 0.125692605972\n",
      "Training_correct error 0.118412218988\n",
      "Testing_correct error: 0.114850081503\n",
      "top- 1 = [3, 48] : 48\n",
      "top- 5 = [21, 199] : 199\n",
      "top- 10 = [30, 304] : 304\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      52.060060\n",
      "std       68.136176\n",
      "min        1.000000\n",
      "25%        7.500000\n",
      "50%       28.000000\n",
      "75%       63.000000\n",
      "max      370.500000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46] : 46\n",
      "top- 5 = [12, 151] : 151\n",
      "top- 10 = [24, 232] : 232\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       61.457000\n",
      "std        73.761911\n",
      "min         1.000000\n",
      "25%        11.500000\n",
      "50%        34.000000\n",
      "75%        85.000000\n",
      "max       448.000000\n",
      "dtype: float64\n",
      "epoch took: 49.6009230614\n",
      "\n",
      "Epoch 1 ...\n",
      "Epoch  1 Update  84 Cost  0.10254971683\n",
      "Epoch  1 Update  105 Cost  0.11504471302\n",
      "Epoch  1 Update  126 Cost  0.0893101021647\n",
      "Training error: 0.117395535111\n",
      "Training_correct error 0.0951170623302\n",
      "Testing_correct error: 0.0903676599264\n",
      "top- 1 = [3, 48, 50] : 50\n",
      "top- 5 = [21, 199, 208] : 208\n",
      "top- 10 = [30, 304, 320] : 320\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      52.107608\n",
      "std       67.633366\n",
      "min        1.000000\n",
      "25%        7.000000\n",
      "50%       22.000000\n",
      "75%       72.500000\n",
      "max      324.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43] : 43\n",
      "top- 5 = [12, 151, 169] : 169\n",
      "top- 10 = [24, 232, 267] : 267\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       54.447000\n",
      "std        69.017665\n",
      "min         1.000000\n",
      "25%         9.500000\n",
      "50%        30.500000\n",
      "75%        61.250000\n",
      "max       425.000000\n",
      "dtype: float64\n",
      "epoch took: 50.0520739555\n",
      "\n",
      "Epoch 2 ...\n",
      "Epoch  2 Update  147 Cost  0.0750384479761\n",
      "Epoch  2 Update  168 Cost  0.0610469020903\n",
      "Epoch  2 Update  189 Cost  0.124674670398\n",
      "Training error: 0.0985486656427\n",
      "Training_correct error 0.101944170892\n",
      "Testing_correct error: 0.102379269898\n",
      "top- 1 = [3, 48, 50, 86] : 86\n",
      "top- 5 = [21, 199, 208, 293] : 293\n",
      "top- 10 = [30, 304, 320, 431] : 431\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      38.222723\n",
      "std       53.008296\n",
      "min        1.000000\n",
      "25%        4.000000\n",
      "50%       15.000000\n",
      "75%       50.250000\n",
      "max      398.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81] : 81\n",
      "top- 5 = [12, 151, 169, 246] : 246\n",
      "top- 10 = [24, 232, 267, 378] : 378\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       44.222000\n",
      "std        56.271624\n",
      "min         1.000000\n",
      "25%         5.500000\n",
      "50%        17.500000\n",
      "75%        62.000000\n",
      "max       432.000000\n",
      "dtype: float64\n",
      "epoch took: 50.0570180416\n",
      "\n",
      "Epoch 3 ...\n",
      "Epoch  3 Update  210 Cost  0.0899893343449\n",
      "Epoch  3 Update  231 Cost  0.0913775414228\n",
      "Epoch  3 Update  252 Cost  0.0783319026232\n",
      "Training error: 0.0997497886419\n",
      "Training_correct error 0.0821037888527\n",
      "Testing_correct error: 0.083479039371\n",
      "top- 1 = [3, 48, 50, 86, 76] : 76\n",
      "top- 5 = [21, 199, 208, 293, 310] : 310\n",
      "top- 10 = [30, 304, 320, 431, 411] : 411\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      37.868368\n",
      "std       57.583045\n",
      "min        1.000000\n",
      "25%        3.500000\n",
      "50%       16.000000\n",
      "75%       42.000000\n",
      "max      378.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66] : 66\n",
      "top- 5 = [12, 151, 169, 246, 242] : 242\n",
      "top- 10 = [24, 232, 267, 378, 344] : 344\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       43.143000\n",
      "std        60.577844\n",
      "min         1.000000\n",
      "25%         6.000000\n",
      "50%        21.250000\n",
      "75%        46.000000\n",
      "max       417.000000\n",
      "dtype: float64\n",
      "epoch took: 50.3379600048\n",
      "\n",
      "Epoch 4 ...\n",
      "Epoch  4 Update  273 Cost  0.102640353143\n",
      "Epoch  4 Update  294 Cost  0.0877730324864\n",
      "Epoch  4 Update  315 Cost  0.119521379471\n",
      "Training error: 0.086856238544\n",
      "Training_correct error 0.107121005654\n",
      "Testing_correct error: 0.120220825076\n",
      "top- 1 = [3, 48, 50, 86, 76, 118] : 118\n",
      "top- 5 = [21, 199, 208, 293, 310, 392] : 392\n",
      "top- 10 = [30, 304, 320, 431, 411, 533] : 533\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      24.363864\n",
      "std       37.543915\n",
      "min        1.000000\n",
      "25%        3.000000\n",
      "50%        9.000000\n",
      "75%       26.500000\n",
      "max      279.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92] : 92\n",
      "top- 5 = [12, 151, 169, 246, 242, 285] : 285\n",
      "top- 10 = [24, 232, 267, 378, 344, 455] : 455\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       35.453000\n",
      "std        50.465682\n",
      "min         1.000000\n",
      "25%         4.000000\n",
      "50%        12.500000\n",
      "75%        41.250000\n",
      "max       324.000000\n",
      "dtype: float64\n",
      "epoch took: 50.2852768898\n",
      "\n",
      "Epoch 5 ...\n",
      "Epoch  5 Update  336 Cost  0.0586893036962\n",
      "Epoch  5 Update  357 Cost  0.0592846162617\n",
      "Epoch  5 Update  378 Cost  0.172971352935\n",
      "Training error: 0.0897238701582\n",
      "Training_correct error 0.0772184580564\n",
      "Testing_correct error: 0.0757179409266\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93] : 93\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359] : 359\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495] : 495\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      28.575576\n",
      "std       44.863448\n",
      "min        1.000000\n",
      "25%        3.000000\n",
      "50%       10.500000\n",
      "75%       31.000000\n",
      "max      277.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79] : 79\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279] : 279\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443] : 443\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       35.241500\n",
      "std        50.618402\n",
      "min         1.000000\n",
      "25%         4.500000\n",
      "50%        13.750000\n",
      "75%        41.000000\n",
      "max       273.000000\n",
      "dtype: float64\n",
      "epoch took: 50.6158559322\n",
      "\n",
      "Epoch 6 ...\n",
      "Epoch  6 Update  399 Cost  0.0754013508558\n",
      "Epoch  6 Update  420 Cost  0.0876249596477\n",
      "Epoch  6 Update  441 Cost  0.165213420987\n",
      "Training error: 0.0925588086247\n",
      "Training_correct error 0.075753249228\n",
      "Testing_correct error: 0.0882906317711\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92] : 92\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406] : 406\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582] : 582\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      24.823824\n",
      "std       42.170965\n",
      "min        1.000000\n",
      "25%        3.000000\n",
      "50%        7.500000\n",
      "75%       21.000000\n",
      "max      344.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80] : 80\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306] : 306\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470] : 470\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       30.191000\n",
      "std        47.151795\n",
      "min         1.000000\n",
      "25%         4.000000\n",
      "50%        11.000000\n",
      "75%        25.000000\n",
      "max       333.000000\n",
      "dtype: float64\n",
      "epoch took: 50.237580061\n",
      "\n",
      "Epoch 7 ...\n",
      "Epoch  7 Update  462 Cost  0.0874141305685\n",
      "Epoch  7 Update  483 Cost  0.0745683759451\n",
      "Epoch  7 Update  504 Cost  0.084833085537\n",
      "Training error: 0.0767989233136\n",
      "Training_correct error 0.0734879598022\n",
      "Testing_correct error: 0.0873811766505\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202] : 202\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515] : 515\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690] : 690\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      18.065065\n",
      "std       33.711923\n",
      "min        1.000000\n",
      "25%        1.500000\n",
      "50%        5.000000\n",
      "75%       14.500000\n",
      "max      263.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135] : 135\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423] : 423\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557] : 557\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       27.840000\n",
      "std        46.303525\n",
      "min         1.000000\n",
      "25%         2.500000\n",
      "50%         8.000000\n",
      "75%        27.250000\n",
      "max       291.000000\n",
      "dtype: float64\n",
      "epoch took: 50.0159208775\n",
      "\n",
      "Epoch 8 ...\n",
      "Epoch  8 Update  525 Cost  0.0818914100528\n",
      "Epoch  8 Update  546 Cost  0.102885998785\n",
      "Epoch  8 Update  567 Cost  0.124444670975\n",
      "Training error: 0.0649016872048\n",
      "Training_correct error 0.0729025006294\n",
      "Testing_correct error: 0.0839079767466\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205] : 205\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564] : 564\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738] : 738\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      12.638639\n",
      "std       22.141127\n",
      "min        1.000000\n",
      "25%        1.500000\n",
      "50%        4.000000\n",
      "75%       11.000000\n",
      "max      162.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156] : 156\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468] : 468\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596] : 596\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       19.704000\n",
      "std        34.629317\n",
      "min         1.000000\n",
      "25%         2.000000\n",
      "50%         6.000000\n",
      "75%        19.625000\n",
      "max       206.000000\n",
      "dtype: float64\n",
      "epoch took: 49.9009859562\n",
      "\n",
      "Epoch 9 ...\n",
      "Epoch  9 Update  588 Cost  0.08970028162\n",
      "Epoch  9 Update  609 Cost  0.0695525407791\n",
      "Epoch  9 Update  630 Cost  0.0319570749998\n",
      "Training error: 0.0614430606365\n",
      "Training_correct error 0.0697435438633\n",
      "Testing_correct error: 0.0887727290392\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225] : 225\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610] : 610\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736] : 736\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      12.092092\n",
      "std       22.614483\n",
      "min        1.000000\n",
      "25%        1.500000\n",
      "50%        3.500000\n",
      "75%       11.250000\n",
      "max      208.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166] : 166\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479] : 479\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614] : 614\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.00000\n",
      "mean       21.41400\n",
      "std        38.95274\n",
      "min         1.00000\n",
      "25%         2.00000\n",
      "50%         6.00000\n",
      "75%        22.50000\n",
      "max       271.00000\n",
      "dtype: float64\n",
      "epoch took: 48.964138031\n",
      "\n",
      "Epoch 10 ...\n",
      "Epoch  10 Update  651 Cost  0.063362672925\n",
      "Epoch  10 Update  672 Cost  0.0694444775581\n",
      "Epoch  10 Update  693 Cost  0.0757331997156\n",
      "Training error: 0.064365811646\n",
      "Training_correct error 0.0788674727082\n",
      "Testing_correct error: 0.0984400510788\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200] : 200\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598] : 598\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745] : 745\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      12.458458\n",
      "std       22.218322\n",
      "min        1.000000\n",
      "25%        1.500000\n",
      "50%        3.500000\n",
      "75%       11.000000\n",
      "max      173.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148] : 148\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475] : 475\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605] : 605\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       21.715000\n",
      "std        38.724454\n",
      "min         1.000000\n",
      "25%         2.000000\n",
      "50%         6.000000\n",
      "75%        19.000000\n",
      "max       289.000000\n",
      "dtype: float64\n",
      "epoch took: 49.38337183\n",
      "\n",
      "Epoch 11 ...\n",
      "Epoch  11 Update  714 Cost  0.0437170118093\n",
      "Epoch  11 Update  735 Cost  0.114784456789\n",
      "Epoch  11 Update  756 Cost  0.0514871068299\n",
      "Training error: 0.0676115751266\n",
      "Training_correct error 0.0683713331819\n",
      "Testing_correct error: 0.0836198031902\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161] : 161\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536] : 536\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698] : 698\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      13.920921\n",
      "std       23.795582\n",
      "min        1.000000\n",
      "25%        2.000000\n",
      "50%        4.500000\n",
      "75%       14.000000\n",
      "max      294.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115] : 115\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452] : 452\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622] : 622\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.00000\n",
      "mean       20.02600\n",
      "std        33.57005\n",
      "min         1.00000\n",
      "25%         2.00000\n",
      "50%         6.25000\n",
      "75%        17.50000\n",
      "max       217.00000\n",
      "dtype: float64\n",
      "epoch took: 48.763133049\n",
      "\n",
      "Epoch 12 ...\n",
      "Epoch  12 Update  777 Cost  0.0336507894099\n",
      "Epoch  12 Update  798 Cost  0.0849446207285\n",
      "Epoch  12 Update  819 Cost  0.0798107907176\n",
      "Training error: 0.0687031969428\n",
      "Training_correct error 0.0725525915623\n",
      "Testing_correct error: 0.0905041620135\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169] : 169\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532] : 532\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702] : 702\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      13.523524\n",
      "std       23.272234\n",
      "min        1.000000\n",
      "25%        2.000000\n",
      "50%        5.000000\n",
      "75%       13.000000\n",
      "max      288.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127] : 127\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485] : 485\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672] : 672\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       17.531000\n",
      "std        29.637196\n",
      "min         1.000000\n",
      "25%         2.000000\n",
      "50%         5.500000\n",
      "75%        15.000000\n",
      "max       190.000000\n",
      "dtype: float64\n",
      "epoch took: 49.0213570595\n",
      "\n",
      "Epoch 13 ...\n",
      "Epoch  13 Update  840 Cost  0.072304174304\n",
      "Epoch  13 Update  861 Cost  0.0690368041396\n",
      "Epoch  13 Update  882 Cost  0.0413294769824\n",
      "Training error: 0.0671595260501\n",
      "Training_correct error 0.0689151287079\n",
      "Testing_correct error: 0.0870760679245\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186] : 186\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595] : 595\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716] : 716\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean      12.617618\n",
      "std       21.336281\n",
      "min        1.000000\n",
      "25%        1.500000\n",
      "50%        3.500000\n",
      "75%       12.500000\n",
      "max      245.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143] : 143\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529] : 529\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665] : 665\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       15.855000\n",
      "std        25.879059\n",
      "min         1.000000\n",
      "25%         2.000000\n",
      "50%         5.000000\n",
      "75%        16.000000\n",
      "max       181.000000\n",
      "dtype: float64\n",
      "epoch took: 50.0139009953\n",
      "\n",
      "Epoch 14 ...\n",
      "Epoch  14 Update  903 Cost  0.0230688545853\n",
      "Epoch  14 Update  924 Cost  0.0697318464518\n",
      "Epoch  14 Update  945 Cost  0.0603713318706\n",
      "Training error: 0.055987522006\n",
      "Training_correct error 0.0552660189569\n",
      "Testing_correct error: 0.0641145408154\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233] : 233\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653] : 653\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779] : 779\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       8.545546\n",
      "std       13.568803\n",
      "min        1.000000\n",
      "25%        1.500000\n",
      "50%        3.000000\n",
      "75%        8.500000\n",
      "max      157.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162] : 162\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530] : 530\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698] : 698\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       11.858000\n",
      "std        17.776947\n",
      "min         1.000000\n",
      "25%         2.000000\n",
      "50%         4.750000\n",
      "75%        13.125000\n",
      "max       120.000000\n",
      "dtype: float64\n",
      "epoch took: 49.6344349384\n",
      "\n",
      "Epoch 15 ...\n",
      "Epoch  15 Update  966 Cost  0.0660632997751\n",
      "Epoch  15 Update  987 Cost  0.0347093641758\n",
      "Epoch  15 Update  1008 Cost  0.039114356041\n",
      "Training error: 0.0527125261724\n",
      "Training_correct error 0.0673270151019\n",
      "Testing_correct error: 0.07998406142\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259] : 259\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714] : 714\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831] : 831\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       7.254254\n",
      "std       13.664529\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        2.500000\n",
      "75%        6.000000\n",
      "max      155.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193] : 193\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630] : 630\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753] : 753\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean       11.165000\n",
      "std        18.126012\n",
      "min         1.000000\n",
      "25%         1.500000\n",
      "50%         3.000000\n",
      "75%        10.000000\n",
      "max       151.000000\n",
      "dtype: float64\n",
      "epoch took: 48.8395800591\n",
      "\n",
      "Epoch 16 ...\n",
      "Epoch  16 Update  1029 Cost  0.0442724451423\n",
      "Epoch  16 Update  1050 Cost  0.0652790665627\n",
      "Epoch  16 Update  1071 Cost  0.0747208446264\n",
      "Training error: 0.0483071468771\n",
      "Training_correct error 0.0516675412655\n",
      "Testing_correct error: 0.0678435266018\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272] : 272\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735] : 735\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843] : 843\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       6.183183\n",
      "std       11.160784\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        2.500000\n",
      "75%        6.000000\n",
      "max      164.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194] : 194\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636] : 636\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802] : 802\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        9.038000\n",
      "std        14.761859\n",
      "min         1.000000\n",
      "25%         1.500000\n",
      "50%         3.500000\n",
      "75%         8.000000\n",
      "max       135.000000\n",
      "dtype: float64\n",
      "epoch took: 48.9036560059\n",
      "\n",
      "Epoch 17 ...\n",
      "Epoch  17 Update  1092 Cost  0.0414074324071\n",
      "Epoch  17 Update  1113 Cost  0.071006372571\n",
      "Epoch  17 Update  1134 Cost  0.0419829301536\n",
      "Training error: 0.0467457361519\n",
      "Training_correct error 0.0499443151057\n",
      "Testing_correct error: 0.0650392621756\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263] : 263\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749] : 749\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864] : 864\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       5.903904\n",
      "std       10.514103\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        2.500000\n",
      "75%        5.250000\n",
      "max      145.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198] : 198\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685] : 685\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807] : 807\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        8.325000\n",
      "std        13.669537\n",
      "min         1.000000\n",
      "25%         1.500000\n",
      "50%         3.000000\n",
      "75%         7.000000\n",
      "max       123.000000\n",
      "dtype: float64\n",
      "epoch took: 50.1264619827\n",
      "\n",
      "Epoch 18 ...\n",
      "Epoch  18 Update  1155 Cost  0.0281577594578\n",
      "Epoch  18 Update  1176 Cost  0.0279857199639\n",
      "Epoch  18 Update  1197 Cost  0.031181935221\n",
      "Training error: 0.0448585934937\n",
      "Training_correct error 0.0507593862712\n",
      "Testing_correct error: 0.0675635188818\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297] : 297\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776] : 776\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887] : 887\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       5.160160\n",
      "std        9.167385\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        2.000000\n",
      "75%        5.000000\n",
      "max       96.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225] : 225\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701] : 701\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817] : 817\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        8.575000\n",
      "std        15.275134\n",
      "min         1.000000\n",
      "25%         1.500000\n",
      "50%         2.500000\n",
      "75%         7.000000\n",
      "max       134.000000\n",
      "dtype: float64\n",
      "epoch took: 49.9880919456\n",
      "\n",
      "Epoch 19 ...\n",
      "Epoch  19 Update  1218 Cost  0.0584981888533\n",
      "Epoch  19 Update  1239 Cost  0.0524895340204\n",
      "Epoch  19 Update  1260 Cost  0.0314953140914\n",
      "Training error: 0.0442051738501\n",
      "Training_correct error 0.0425166003406\n",
      "Testing_correct error: 0.0482059717178\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274] : 274\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778] : 778\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881] : 881\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       4.926927\n",
      "std        7.862557\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        2.000000\n",
      "75%        5.000000\n",
      "max      100.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206] : 206\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689] : 689\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814] : 814\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        6.672000\n",
      "std         9.635105\n",
      "min         1.000000\n",
      "25%         1.500000\n",
      "50%         3.000000\n",
      "75%         7.000000\n",
      "max       109.000000\n",
      "dtype: float64\n",
      "epoch took: 49.970911026\n",
      "\n",
      "Epoch 20 ...\n",
      "Epoch  20 Update  1281 Cost  0.0426433384418\n",
      "Epoch  20 Update  1302 Cost  0.0587750151753\n",
      "Epoch  20 Update  1323 Cost  0.0450344607234\n",
      "Training error: 0.0559212006629\n",
      "Training_correct error 0.0429740361869\n",
      "Testing_correct error: 0.0454378426075\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255] : 255\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691] : 691\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808] : 808\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       6.494494\n",
      "std        8.808143\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        2.500000\n",
      "75%        7.500000\n",
      "max       59.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170] : 170\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616] : 616\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751] : 751\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        8.008000\n",
      "std        10.683542\n",
      "min         1.000000\n",
      "25%         1.500000\n",
      "50%         4.000000\n",
      "75%        10.000000\n",
      "max        79.000000\n",
      "dtype: float64\n",
      "epoch took: 49.3655118942\n",
      "\n",
      "Epoch 21 ...\n",
      "Epoch  21 Update  1344 Cost  0.0263530723751\n",
      "Epoch  21 Update  1365 Cost  0.0522510185838\n",
      "Epoch  21 Update  1386 Cost  0.0789667367935\n",
      "Training error: 0.055720359087\n",
      "Training_correct error 0.0555658824742\n",
      "Testing_correct error: 0.0678117498755\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212] : 212\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669] : 669\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815] : 815\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       7.540541\n",
      "std       13.140623\n",
      "min        1.000000\n",
      "25%        1.500000\n",
      "50%        3.000000\n",
      "75%        7.500000\n",
      "max      167.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155] : 155\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550] : 550\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738] : 738\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.00000\n",
      "mean       11.61400\n",
      "std        18.44666\n",
      "min         1.00000\n",
      "25%         2.00000\n",
      "50%         4.50000\n",
      "75%        11.00000\n",
      "max       124.00000\n",
      "dtype: float64\n",
      "epoch took: 48.9510321617\n",
      "\n",
      "Epoch 22 ...\n",
      "Epoch  22 Update  1407 Cost  0.0580897480249\n",
      "Epoch  22 Update  1428 Cost  0.0394529700279\n",
      "Epoch  22 Update  1449 Cost  0.0814228877425\n",
      "Training error: 0.0426833406091\n",
      "Training_correct error 0.0509922988713\n",
      "Testing_correct error: 0.0691846162081\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357] : 357\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826] : 826\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928] : 928\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       4.130130\n",
      "std        7.809838\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        2.000000\n",
      "75%        3.500000\n",
      "max       79.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239] : 239\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741] : 741\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846] : 846\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.00000\n",
      "mean        8.28100\n",
      "std        16.69739\n",
      "min         1.00000\n",
      "25%         1.50000\n",
      "50%         2.00000\n",
      "75%         5.50000\n",
      "max       122.00000\n",
      "dtype: float64\n",
      "epoch took: 48.2691612244\n",
      "\n",
      "Epoch 23 ...\n",
      "Epoch  23 Update  1470 Cost  0.0264135282487\n",
      "Epoch  23 Update  1491 Cost  0.0362154617906\n",
      "Epoch  23 Update  1512 Cost  0.0264660809189\n",
      "Training error: 0.0416636988521\n",
      "Training_correct error 0.0405298173428\n",
      "Testing_correct error: 0.0468727312982\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337] : 337\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827] : 827\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909] : 909\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       3.782783\n",
      "std        5.498439\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.500000\n",
      "75%        3.500000\n",
      "max       75.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234] : 234\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746] : 746\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839] : 839\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        5.429000\n",
      "std         7.801399\n",
      "min         1.000000\n",
      "25%         1.500000\n",
      "50%         2.000000\n",
      "75%         5.500000\n",
      "max        93.000000\n",
      "dtype: float64\n",
      "epoch took: 49.2957229614\n",
      "\n",
      "Epoch 24 ...\n",
      "Epoch  24 Update  1533 Cost  0.0491187348962\n",
      "Epoch  24 Update  1554 Cost  0.0457423888147\n",
      "Epoch  24 Update  1575 Cost  0.0264707300812\n",
      "Training error: 0.0377053432167\n",
      "Training_correct error 0.0344945862889\n",
      "Testing_correct error: 0.043886654079\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394] : 394\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875] : 875\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939] : 939\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       3.208208\n",
      "std        4.979637\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.500000\n",
      "75%        3.000000\n",
      "max       54.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274] : 274\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783] : 783\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872] : 872\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        5.048000\n",
      "std         7.617199\n",
      "min         1.000000\n",
      "25%         1.000000\n",
      "50%         2.000000\n",
      "75%         4.500000\n",
      "max        85.000000\n",
      "dtype: float64\n",
      "epoch took: 50.686259985\n",
      "\n",
      "Epoch 25 ...\n",
      "Epoch  25 Update  1596 Cost  0.0414323508739\n",
      "Epoch  25 Update  1617 Cost  0.0281491354108\n",
      "Epoch  25 Update  1638 Cost  0.0174471158534\n",
      "Training error: 0.0395827367902\n",
      "Training_correct error 0.0427967011929\n",
      "Testing_correct error: 0.0637759938836\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399] : 399\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867] : 867\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940] : 940\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       3.638639\n",
      "std        7.016660\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.500000\n",
      "75%        3.000000\n",
      "max       56.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268] : 268\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750] : 750\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855] : 855\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.00000\n",
      "mean        7.44400\n",
      "std        14.25012\n",
      "min         1.00000\n",
      "25%         1.00000\n",
      "50%         2.00000\n",
      "75%         5.12500\n",
      "max       100.00000\n",
      "dtype: float64\n",
      "epoch took: 50.023966074\n",
      "\n",
      "Epoch 26 ...\n",
      "Epoch  26 Update  1659 Cost  0.0460323989391\n",
      "Epoch  26 Update  1680 Cost  0.0329669043422\n",
      "Epoch  26 Update  1701 Cost  0.0176671352237\n",
      "Training error: 0.0378140434623\n",
      "Training_correct error 0.0409294292331\n",
      "Testing_correct error: 0.0505609735847\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399, 373] : 373\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867, 859] : 859\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940, 941] : 941\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       3.520521\n",
      "std        5.985439\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.500000\n",
      "75%        3.500000\n",
      "max       80.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268, 269] : 269\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750, 777] : 777\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855, 853] : 853\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        5.603000\n",
      "std         9.090546\n",
      "min         1.000000\n",
      "25%         1.000000\n",
      "50%         2.000000\n",
      "75%         5.000000\n",
      "max        69.000000\n",
      "dtype: float64\n",
      "epoch took: 50.0605611801\n",
      "\n",
      "Epoch 27 ...\n",
      "Epoch  27 Update  1722 Cost  0.0237131565809\n",
      "Epoch  27 Update  1743 Cost  0.0538401976228\n",
      "Epoch  27 Update  1764 Cost  0.0231092013419\n",
      "Training error: 0.0398889742792\n",
      "Training_correct error 0.0451489761472\n",
      "Testing_correct error: 0.0619098581374\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399, 373, 397] : 397\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867, 859, 841] : 841\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940, 941, 934] : 934\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       3.738739\n",
      "std        6.686009\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.500000\n",
      "75%        3.000000\n",
      "max       54.500000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268, 269, 257] : 257\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750, 777, 747] : 747\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855, 853, 852] : 852\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        7.371000\n",
      "std        14.550432\n",
      "min         1.000000\n",
      "25%         1.000000\n",
      "50%         2.000000\n",
      "75%         5.500000\n",
      "max       107.000000\n",
      "dtype: float64\n",
      "epoch took: 49.7377181053\n",
      "\n",
      "Epoch 28 ...\n",
      "Epoch  28 Update  1785 Cost  0.0307862199843\n",
      "Epoch  28 Update  1806 Cost  0.0532463639975\n",
      "Epoch  28 Update  1827 Cost  0.023118449375\n",
      "Training error: 0.043498147279\n",
      "Training_correct error 0.0411802940071\n",
      "Testing_correct error: 0.0437696017325\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399, 373, 397, 332] : 332\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867, 859, 841, 836] : 836\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940, 941, 934, 943] : 943\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       3.345345\n",
      "std        4.253062\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        2.000000\n",
      "75%        3.500000\n",
      "max       48.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268, 269, 257, 239] : 239\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750, 777, 747, 766] : 766\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855, 853, 852, 879] : 879\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        4.801000\n",
      "std         6.670524\n",
      "min         1.000000\n",
      "25%         1.500000\n",
      "50%         2.000000\n",
      "75%         5.000000\n",
      "max        46.000000\n",
      "dtype: float64\n",
      "epoch took: 48.7383549213\n",
      "\n",
      "Epoch 29 ...\n",
      "Epoch  29 Update  1848 Cost  0.0345920696855\n",
      "Epoch  29 Update  1869 Cost  0.0392372980714\n",
      "Epoch  29 Update  1890 Cost  0.0204216353595\n",
      "Training error: 0.0357829034328\n",
      "Training_correct error 0.0366004817188\n",
      "Testing_correct error: 0.0494484007359\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399, 373, 397, 332, 426] : 426\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867, 859, 841, 836, 886] : 886\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940, 941, 934, 943, 957] : 957\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       2.897898\n",
      "std        4.603528\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.500000\n",
      "75%        2.500000\n",
      "max       46.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268, 269, 257, 239, 290] : 290\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750, 777, 747, 766, 793] : 793\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855, 853, 852, 879, 875] : 875\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        5.596000\n",
      "std         9.861746\n",
      "min         1.000000\n",
      "25%         1.000000\n",
      "50%         2.000000\n",
      "75%         4.000000\n",
      "max        86.000000\n",
      "dtype: float64\n",
      "epoch took: 48.9504799843\n",
      "\n",
      "Epoch 30 ...\n",
      "Epoch  30 Update  1911 Cost  0.0358876697719\n",
      "Epoch  30 Update  1932 Cost  0.034554567188\n",
      "Epoch  30 Update  1953 Cost  0.0349995717406\n",
      "Training error: 0.033931940794\n",
      "Training_correct error 0.0362541563809\n",
      "Testing_correct error: 0.045201279223\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399, 373, 397, 332, 426, 420] : 420\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867, 859, 841, 836, 886, 912] : 912\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940, 941, 934, 943, 957, 969] : 969\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       2.510511\n",
      "std        3.581871\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.500000\n",
      "75%        2.500000\n",
      "max       65.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268, 269, 257, 239, 290, 290] : 290\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750, 777, 747, 766, 793, 810] : 810\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855, 853, 852, 879, 875, 893] : 893\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        4.524000\n",
      "std         6.976395\n",
      "min         1.000000\n",
      "25%         1.000000\n",
      "50%         2.000000\n",
      "75%         4.000000\n",
      "max        62.000000\n",
      "dtype: float64\n",
      "epoch took: 49.0416247845\n",
      "\n",
      "Epoch 31 ...\n",
      "Epoch  31 Update  1974 Cost  0.0188236683607\n",
      "Epoch  31 Update  1995 Cost  0.0324405953288\n",
      "Epoch  31 Update  2016 Cost  0.0351068265736\n",
      "Training error: 0.0341457985342\n",
      "Training_correct error 0.037388574332\n",
      "Testing_correct error: 0.0550444424152\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399, 373, 397, 332, 426, 420, 437] : 437\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867, 859, 841, 836, 886, 912, 902] : 902\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940, 941, 934, 943, 957, 969, 966] : 966\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       3.173173\n",
      "std        6.846553\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.500000\n",
      "75%        2.500000\n",
      "max       74.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268, 269, 257, 239, 290, 290, 302] : 302\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750, 777, 747, 766, 793, 810, 795] : 795\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855, 853, 852, 879, 875, 893, 884] : 884\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        6.820000\n",
      "std        15.057368\n",
      "min         1.000000\n",
      "25%         1.000000\n",
      "50%         2.000000\n",
      "75%         4.000000\n",
      "max       113.000000\n",
      "dtype: float64\n",
      "epoch took: 48.8438000679\n",
      "\n",
      "Epoch 32 ...\n",
      "Epoch  32 Update  2037 Cost  0.0197140984237\n",
      "Epoch  32 Update  2058 Cost  0.0404128469527\n",
      "Epoch  32 Update  2079 Cost  0.0363337732852\n",
      "Training error: 0.0400777868927\n",
      "Training_correct error 0.0407038629055\n",
      "Testing_correct error: 0.0435023233294\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399, 373, 397, 332, 426, 420, 437, 345] : 345\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867, 859, 841, 836, 886, 912, 902, 858] : 858\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940, 941, 934, 943, 957, 969, 966, 944] : 944\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.00000\n",
      "mean       3.23023\n",
      "std        4.42433\n",
      "min        1.00000\n",
      "25%        1.00000\n",
      "50%        2.00000\n",
      "75%        3.50000\n",
      "max       77.00000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268, 269, 257, 239, 290, 290, 302, 265] : 265\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750, 777, 747, 766, 793, 810, 795, 790] : 790\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855, 853, 852, 879, 875, 893, 884, 887] : 887\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        4.762000\n",
      "std         7.138229\n",
      "min         1.000000\n",
      "25%         1.000000\n",
      "50%         2.000000\n",
      "75%         4.500000\n",
      "max        42.000000\n",
      "dtype: float64\n",
      "epoch took: 48.5661759377\n",
      "\n",
      "Epoch 33 ...\n",
      "Epoch  33 Update  2100 Cost  0.0225670337677\n",
      "Epoch  33 Update  2121 Cost  0.0422090739012\n",
      "Epoch  33 Update  2142 Cost  0.0197534430772\n",
      "Training error: 0.0357491523027\n",
      "Training_correct error 0.0343536958098\n",
      "Testing_correct error: 0.0373701564968\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399, 373, 397, 332, 426, 420, 437, 345, 395] : 395\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867, 859, 841, 836, 886, 912, 902, 858, 921] : 921\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940, 941, 934, 943, 957, 969, 966, 944, 982] : 982\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       2.395395\n",
      "std        2.664116\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.500000\n",
      "75%        2.500000\n",
      "max       36.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268, 269, 257, 239, 290, 290, 302, 265, 295] : 295\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750, 777, 747, 766, 793, 810, 795, 790, 840] : 840\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855, 853, 852, 879, 875, 893, 884, 887, 919] : 919\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        3.757000\n",
      "std         4.943722\n",
      "min         1.000000\n",
      "25%         1.000000\n",
      "50%         2.000000\n",
      "75%         4.000000\n",
      "max        30.000000\n",
      "dtype: float64\n",
      "epoch took: 50.380546093\n",
      "\n",
      "Epoch 34 ...\n",
      "Epoch  34 Update  2163 Cost  0.0274707898498\n",
      "Epoch  34 Update  2184 Cost  0.0142586715519\n",
      "Epoch  34 Update  2205 Cost  0.0413485430181\n",
      "Training error: 0.0308684203774\n",
      "Training_correct error 0.0349571034312\n",
      "Testing_correct error: 0.0484385229647\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399, 373, 397, 332, 426, 420, 437, 345, 395, 454] : 454\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867, 859, 841, 836, 886, 912, 902, 858, 921, 931] : 931\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940, 941, 934, 943, 957, 969, 966, 944, 982, 969] : 969\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       2.552553\n",
      "std        4.316103\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.500000\n",
      "75%        2.500000\n",
      "max       41.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268, 269, 257, 239, 290, 290, 302, 265, 295, 327] : 327\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750, 777, 747, 766, 793, 810, 795, 790, 840, 822] : 822\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855, 853, 852, 879, 875, 893, 884, 887, 919, 895] : 895\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        5.188000\n",
      "std         9.699523\n",
      "min         1.000000\n",
      "25%         1.000000\n",
      "50%         2.000000\n",
      "75%         4.000000\n",
      "max        81.000000\n",
      "dtype: float64\n",
      "epoch took: 48.7266769409\n",
      "\n",
      "Epoch 35 ...\n",
      "Epoch  35 Update  2226 Cost  0.0219029150903\n",
      "Epoch  35 Update  2247 Cost  0.0226015094668\n",
      "Epoch  35 Update  2268 Cost  0.0214089415967\n",
      "Training error: 0.0279977619648\n",
      "Training_correct error 0.0266631599516\n",
      "Testing_correct error: 0.0349534787238\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399, 373, 397, 332, 426, 420, 437, 345, 395, 454, 500] : 500\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867, 859, 841, 836, 886, 912, 902, 858, 921, 931, 944] : 944\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940, 941, 934, 943, 957, 969, 966, 944, 982, 969, 981] : 981\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       2.123123\n",
      "std        2.839804\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.000000\n",
      "75%        2.000000\n",
      "max       34.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268, 269, 257, 239, 290, 290, 302, 265, 295, 327, 344] : 344\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750, 777, 747, 766, 793, 810, 795, 790, 840, 822, 840] : 840\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855, 853, 852, 879, 875, 893, 884, 887, 919, 895, 907] : 907\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        4.095000\n",
      "std         6.649225\n",
      "min         1.000000\n",
      "25%         1.000000\n",
      "50%         1.500000\n",
      "75%         3.500000\n",
      "max        42.000000\n",
      "dtype: float64\n",
      "epoch took: 48.885571003\n",
      "\n",
      "Epoch 36 ...\n",
      "Epoch  36 Update  2289 Cost  0.0116978958249\n",
      "Epoch  36 Update  2310 Cost  0.0345799550414\n",
      "Epoch  36 Update  2331 Cost  0.0285079330206\n",
      "Training error: 0.0314490869641\n",
      "Training_correct error 0.0349974296987\n",
      "Testing_correct error: 0.0378754511476\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399, 373, 397, 332, 426, 420, 437, 345, 395, 454, 500, 423] : 423\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867, 859, 841, 836, 886, 912, 902, 858, 921, 931, 944, 931] : 931\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940, 941, 934, 943, 957, 969, 966, 944, 982, 969, 981, 976] : 976\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       2.336336\n",
      "std        2.749657\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.500000\n",
      "75%        2.500000\n",
      "max       43.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268, 269, 257, 239, 290, 290, 302, 265, 295, 327, 344, 294] : 294\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750, 777, 747, 766, 793, 810, 795, 790, 840, 822, 840, 820] : 820\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855, 853, 852, 879, 875, 893, 884, 887, 919, 895, 907, 913] : 913\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        3.736000\n",
      "std         4.752567\n",
      "min         1.000000\n",
      "25%         1.000000\n",
      "50%         2.000000\n",
      "75%         4.000000\n",
      "max        38.000000\n",
      "dtype: float64\n",
      "epoch took: 48.8142700195\n",
      "\n",
      "Epoch 37 ...\n",
      "Epoch  37 Update  2352 Cost  0.0145929036662\n",
      "Epoch  37 Update  2373 Cost  0.0225160494447\n",
      "Epoch  37 Update  2394 Cost  0.0285577271134\n",
      "Training error: 0.0435521751642\n",
      "Training_correct error 0.04953674227\n",
      "Testing_correct error: 0.0480021797121\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399, 373, 397, 332, 426, 420, 437, 345, 395, 454, 500, 423, 303] : 303\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867, 859, 841, 836, 886, 912, 902, 858, 921, 931, 944, 931, 832] : 832\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940, 941, 934, 943, 957, 969, 966, 944, 982, 969, 981, 976, 960] : 960\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       3.258258\n",
      "std        3.441156\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        2.000000\n",
      "75%        4.000000\n",
      "max       46.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268, 269, 257, 239, 290, 290, 302, 265, 295, 327, 344, 294, 227] : 227\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750, 777, 747, 766, 793, 810, 795, 790, 840, 822, 840, 820, 772] : 772\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855, 853, 852, 879, 875, 893, 884, 887, 919, 895, 907, 913, 918] : 918\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        4.225000\n",
      "std         4.697919\n",
      "min         1.000000\n",
      "25%         1.500000\n",
      "50%         2.500000\n",
      "75%         5.000000\n",
      "max        37.000000\n",
      "dtype: float64\n",
      "epoch took: 49.220813036\n",
      "\n",
      "Epoch 38 ...\n",
      "Epoch  38 Update  2415 Cost  0.029201740399\n",
      "Epoch  38 Update  2436 Cost  0.0323530398309\n",
      "Epoch  38 Update  2457 Cost  0.0410589314997\n",
      "Training error: 0.0332983061671\n",
      "Training_correct error 0.036471247673\n",
      "Testing_correct error: 0.0384236089885\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399, 373, 397, 332, 426, 420, 437, 345, 395, 454, 500, 423, 303, 424] : 424\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867, 859, 841, 836, 886, 912, 902, 858, 921, 931, 944, 931, 832, 918] : 918\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940, 941, 934, 943, 957, 969, 966, 944, 982, 969, 981, 976, 960, 981] : 981\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       2.400400\n",
      "std        2.663932\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.500000\n",
      "75%        3.000000\n",
      "max       43.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268, 269, 257, 239, 290, 290, 302, 265, 295, 327, 344, 294, 227, 301] : 301\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750, 777, 747, 766, 793, 810, 795, 790, 840, 822, 840, 820, 772, 825] : 825\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855, 853, 852, 879, 875, 893, 884, 887, 919, 895, 907, 913, 918, 922] : 922\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        3.652000\n",
      "std         4.525636\n",
      "min         1.000000\n",
      "25%         1.000000\n",
      "50%         2.000000\n",
      "75%         4.000000\n",
      "max        33.000000\n",
      "dtype: float64\n",
      "epoch took: 48.625797987\n",
      "\n",
      "Epoch 39 ...\n",
      "Epoch  39 Update  2478 Cost  0.0360216870904\n",
      "Epoch  39 Update  2499 Cost  0.0418747924268\n",
      "Epoch  39 Update  2520 Cost  0.0241149105132\n",
      "Training error: 0.0285047963262\n",
      "Training_correct error 0.0271630845964\n",
      "Testing_correct error: 0.0319587923586\n",
      "top- 1 = [3, 48, 50, 86, 76, 118, 93, 92, 202, 205, 225, 200, 161, 169, 186, 233, 259, 272, 263, 297, 274, 255, 212, 357, 337, 394, 399, 373, 397, 332, 426, 420, 437, 345, 395, 454, 500, 423, 303, 424, 473] : 473\n",
      "top- 5 = [21, 199, 208, 293, 310, 392, 359, 406, 515, 564, 610, 598, 536, 532, 595, 653, 714, 735, 749, 776, 778, 691, 669, 826, 827, 875, 867, 859, 841, 836, 886, 912, 902, 858, 921, 931, 944, 931, 832, 918, 945] : 945\n",
      "top- 10 = [30, 304, 320, 431, 411, 533, 495, 582, 690, 738, 736, 745, 698, 702, 716, 779, 831, 843, 864, 887, 881, 808, 815, 928, 909, 939, 940, 941, 934, 943, 957, 969, 966, 944, 982, 969, 981, 976, 960, 981, 992] : 992\n",
      "Discription of evaluation (ranking) for training data:\n",
      "count    999.000000\n",
      "mean       2.008008\n",
      "std        2.085938\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.500000\n",
      "75%        2.000000\n",
      "max       27.000000\n",
      "dtype: float64\n",
      "top- 1 = [1, 46, 43, 81, 66, 92, 79, 80, 135, 156, 166, 148, 115, 127, 143, 162, 193, 194, 198, 225, 206, 170, 155, 239, 234, 274, 268, 269, 257, 239, 290, 290, 302, 265, 295, 327, 344, 294, 227, 301, 329] : 329\n",
      "top- 5 = [12, 151, 169, 246, 242, 285, 279, 306, 423, 468, 479, 475, 452, 485, 529, 530, 630, 636, 685, 701, 689, 616, 550, 741, 746, 783, 750, 777, 747, 766, 793, 810, 795, 790, 840, 822, 840, 820, 772, 825, 845] : 845\n",
      "top- 10 = [24, 232, 267, 378, 344, 455, 443, 470, 557, 596, 614, 605, 622, 672, 665, 698, 753, 802, 807, 817, 814, 751, 738, 846, 839, 872, 855, 853, 852, 879, 875, 893, 884, 887, 919, 895, 907, 913, 918, 922, 913] : 913\n",
      "Discription of evaluation (ranking) for testing data:\n",
      "count    1000.000000\n",
      "mean        3.608000\n",
      "std         4.799309\n",
      "min         1.000000\n",
      "25%         1.000000\n",
      "50%         2.000000\n",
      "75%         3.500000\n",
      "max        28.500000\n",
      "dtype: float64\n",
      "epoch took: 48.8241209984\n",
      "saving the model...\n",
      "Evaluate the model using fast estimation...\n",
      "count    999.000000\n",
      "mean       2.008008\n",
      "std        2.085938\n",
      "min        1.000000\n",
      "25%        1.000000\n",
      "50%        1.500000\n",
      "75%        2.000000\n",
      "max       27.000000\n",
      "dtype: float64\n",
      "count    1000.000000\n",
      "mean        3.608000\n",
      "std         4.799309\n",
      "min         1.000000\n",
      "25%         1.000000\n",
      "50%         2.000000\n",
      "75%         3.500000\n",
      "max        28.500000\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "# Set epoch = 40, maxlen = 0\n",
    "run main.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4MAAAHcCAYAAABs0H/xAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKaWUUkqpPEiDQaWUUkoppZTKg/4fGBA9sa84UsQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f143742ee10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "matplotlib.rcParams['figure.figsize'] = (15.0, 7.5)\n",
    "plt.title('MSE distribution')\n",
    "plt.plot(sls.mse,'ro-', label=\"Training data (similarity = 1 + similarity = 2)\")\n",
    "plt.plot(sls.mse_train,'bo-', label=\"Training data (similarity = 1)\")\n",
    "plt.plot(sls.mse_test,'go-', label=\"Testing data (similarity = 1)\")\n",
    "plt.legend()\n",
    "\n",
    "plt.ylabel('MSE')\n",
    "plt.xlabel('Epoches Times')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4YAAAHcCAYAAACK+bS1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n9/98Im8z3/g7tgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f14349bac90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "matplotlib.rcParams['figure.figsize'] = (15.0, 7.5)\n",
    "plt.plot(sls.tops[1],'ro-', label = \"TOP-1 @ Trianing data (similarity = 1)\")\n",
    "plt.plot(sls.tops[5],'bo-', label = \"TOP-5 @ Trianing data (similarity = 1)\" )\n",
    "plt.plot(sls.tops[10],'go-', label = \"TOP-10 @ Trianing data (similarity = 1)\")\n",
    "plt.plot(sls.tops_test[1],'ro--', label = \"TOP-1 @ Test data (similarity = 1)\")\n",
    "plt.plot(sls.tops_test[5],'bo--', label = \"TOP-5 @ Test data (similarity = 1)\")\n",
    "plt.plot(sls.tops_test[10],'go--', label = \"TOP-10 @ Test data (similarity = 1)\")\n",
    "plt.ylabel('Amount of documents belongs to top[1,5,10]')\n",
    "plt.xlabel('Epoches')\n",
    "plt.title('Tendancy of accuracy during training')\n",
    "plt.legend(loc = 2)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "sentences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [Root]",
   "language": "python",
   "name": "Python [Root]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}