├── MNIST_2000.csv ├── README.md ├── SUSY_100k.csv.gz ├── example_01_logistic.ipynb ├── example_02_classification_MNIST.ipynb ├── example_03_autoencoder.ipynb ├── example_04_convolution.ipynb ├── example_05_HDF5_datasets.ipynb ├── lhc_dl_tutorial.pdf └── requirements.txt /README.md: -------------------------------------------------------------------------------- 1 | # LHC Data Science Workshop 2015 2 | ## Deep Learning Tutorial 3 | 4 | [![Binder](http://mybinder.org/badge.svg)](http://mybinder.org/repo/peterjsadowski/lhc2015-dl-tutorial) 5 | 6 | # Installation 7 | pip install Theano 8 | 9 | Download Pylearn2 and put it in your python path. 10 | https://github.com/lisa-lab/pylearn2 11 | 12 | # Requirements 13 | -numpy 14 | -scipy 15 | -matplotlib 16 | -h5py 17 | 18 | -------------------------------------------------------------------------------- /SUSY_100k.csv.gz: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/peterjsadowski/lhc2015-dl-tutorial/f1a903ad209c3b912b49d01044177f98fd109ff1/SUSY_100k.csv.gz -------------------------------------------------------------------------------- /example_05_HDF5_datasets.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "markdown", 5 | "metadata": {}, 6 | "source": [ 7 | "# Summary\n", 8 | "This tutorial demonstrates how to use HDF5 files with pylearn2. By using the Pylearn2 HDF5Dataset class, we avoid loading the entire dataset into memory. \n" 9 | ] 10 | }, 11 | { 12 | "cell_type": "code", 13 | "execution_count": 2, 14 | "metadata": { 15 | "collapsed": false 16 | }, 17 | "outputs": [ 18 | { 19 | "name": "stderr", 20 | "output_type": "stream", 21 | "text": [ 22 | "Using gpu device 0: Tesla C2070 (CNMeM is disabled)\n" 23 | ] 24 | } 25 | ], 26 | "source": [ 27 | "import sys\n", 28 | "import os\n", 29 | "import pickle\n", 30 | "import numpy\n", 31 | "import theano\n", 32 | "import pylearn2\n", 33 | "import pylearn2.datasets\n", 34 | "import pylearn2.training_algorithms\n", 35 | "import pylearn2.training_algorithms.sgd\n", 36 | "import pylearn2.costs\n", 37 | "import pylearn2.models.mlp as mlp\n", 38 | "import pylearn2.train\n", 39 | "import pylearn2.termination_criteria\n", 40 | "import pylearn2.datasets.hdf5\n", 41 | "import matplotlib\n", 42 | "import matplotlib.pyplot as plt\n", 43 | "import h5py\n", 44 | "%matplotlib inline" 45 | ] 46 | }, 47 | { 48 | "cell_type": "markdown", 49 | "metadata": {}, 50 | "source": [ 51 | "# Data set\n", 52 | "Below, we first write our data set into HDF5 format, then instantiate two HDF5Dataset objects for training the model." 53 | ] 54 | }, 55 | { 56 | "cell_type": "code", 57 | "execution_count": 3, 58 | "metadata": { 59 | "collapsed": false, 60 | "scrolled": true 61 | }, 62 | "outputs": [], 63 | "source": [ 64 | "# Load sample of MNIST data.\n", 65 | "filename = './MNIST_2000.csv'\n", 66 | "X = numpy.loadtxt(filename, dtype='float32')\n", 67 | "X, labels = X[:,1:], X[:,0].astype('uint8') # First column is label (0-9)\n", 68 | "y = numpy.zeros((X.shape[0], 10), dtype='float32')\n", 69 | "for i,label in enumerate(labels):\n", 70 | " y[i, label] = 1.0\n", 71 | "\n", 72 | "# Split data into train, test.\n", 73 | "X_train, y_train = X[:1000, :], y[:1000, :]\n", 74 | "X_test, y_test = X[1000:, :], y[1000:, :]\n", 75 | "\n", 76 | "# Write data sets to HDF5 files.\n", 77 | "\n", 78 | "def write_h5file(h5filename, X, y):\n", 79 | " # Save data to HDF5\n", 80 | " with h5py.File(h5filename, 'w') as f:\n", 81 | " f.create_dataset('X', data=X)\n", 82 | " f.create_dataset('y', data=y)\n", 83 | "\n", 84 | "write_h5file('./MNIST_train.h5', X_train, y_train)\n", 85 | "write_h5file('./MNIST_test.h5', X_test, y_test)\n", 86 | "\n", 87 | "dataset_train = pylearn2.datasets.hdf5.HDF5Dataset(filename='./MNIST_train.h5', X='X', y='y')\n", 88 | "dataset_test = pylearn2.datasets.hdf5.HDF5Dataset(filename='./MNIST_test.h5', X='X', y='y')" 89 | ] 90 | }, 91 | { 92 | "cell_type": "markdown", 93 | "metadata": {}, 94 | "source": [ 95 | "# Model\n", 96 | "Model training proceeds as usual." 97 | ] 98 | }, 99 | { 100 | "cell_type": "code", 101 | "execution_count": 13, 102 | "metadata": { 103 | "collapsed": false 104 | }, 105 | "outputs": [], 106 | "source": [ 107 | "layers = []\n", 108 | "nvis = dataset_train.X.shape[1] # Number of input features.\n", 109 | "nhid = 100 # Hidden neurons per layer.\n", 110 | "\n", 111 | "# Layer 0\n", 112 | "istdev = 1.0/numpy.sqrt(nvis) # Initial weights selected from normal distribution.\n", 113 | "layer = pylearn2.models.mlp.RectifiedLinear(layer_name = 'h0', dim=nhid, istdev=istdev)\n", 114 | "layers.append(layer)\n", 115 | "\n", 116 | "# Layer 1\n", 117 | "istdev = 1.0/numpy.sqrt(nhid)\n", 118 | "layer = pylearn2.models.mlp.RectifiedLinear(layer_name = 'h1', dim=nhid, istdev=istdev)\n", 119 | "layers.append(layer)\n", 120 | "\n", 121 | "# Output layer.\n", 122 | "layer = mlp.Softmax(layer_name='y', n_classes=10, istdev=0.001)\n", 123 | "layers.append(layer)\n", 124 | "\n", 125 | "# MLP Model\n", 126 | "model = pylearn2.models.mlp.MLP(layers, nvis=nvis, seed=36)" 127 | ] 128 | }, 129 | { 130 | "cell_type": "markdown", 131 | "metadata": {}, 132 | "source": [ 133 | "# Training Algorithm" 134 | ] 135 | }, 136 | { 137 | "cell_type": "code", 138 | "execution_count": 14, 139 | "metadata": { 140 | "collapsed": false 141 | }, 142 | "outputs": [], 143 | "source": [ 144 | "# Cost/objective function.\n", 145 | "cost = pylearn2.costs.mlp.Default() # Defaults to cross-entropy loss for softmax output.\n", 146 | " \n", 147 | "# Algorithm\n", 148 | "args = {}\n", 149 | "args['cost'] = cost\n", 150 | "args['learning_rate'] = 0.1\n", 151 | "args['batch_size'] = 100\n", 152 | "args['learning_rule'] = pylearn2.training_algorithms.learning_rule.Momentum(init_momentum = 0.5)\n", 153 | "args['monitoring_dataset'] = {'train':dataset_train, 'test':dataset_test}\n", 154 | "args['termination_criterion'] = pylearn2.termination_criteria.EpochCounter(max_epochs=20)\n", 155 | "algorithm = pylearn2.training_algorithms.sgd.SGD(**args)\n", 156 | "\n", 157 | "# Train object.\n", 158 | "filename_model = './model_mnist.pkl'\n", 159 | "train = pylearn2.train.Train(dataset=dataset_train,\n", 160 | " model=model,\n", 161 | " algorithm=algorithm,\n", 162 | " #extensions=extensions,\n", 163 | " save_path=filename_model,\n", 164 | " save_freq=5)" 165 | ] 166 | }, 167 | { 168 | "cell_type": "markdown", 169 | "metadata": {}, 170 | "source": [ 171 | "# Train model" 172 | ] 173 | }, 174 | { 175 | "cell_type": "code", 176 | "execution_count": 15, 177 | "metadata": { 178 | "collapsed": false, 179 | "scrolled": true 180 | }, 181 | "outputs": [ 182 | { 183 | "name": "stdout", 184 | "output_type": "stream", 185 | "text": [ 186 | "Parameter and initial learning rate summary:\n", 187 | "\th0_W: 0.10000000149\n", 188 | "\th0_b: 0.10000000149\n", 189 | "\th1_W: 0.10000000149\n", 190 | "\th1_b: 0.10000000149\n", 191 | "\tsoftmax_b: 0.10000000149\n", 192 | "\tsoftmax_W: 0.10000000149\n", 193 | "Compiling sgd_update...\n", 194 | "Compiling sgd_update done. Time elapsed: 1.140645 seconds\n", 195 | "compiling begin_record_entry...\n", 196 | "compiling begin_record_entry done. Time elapsed: 0.429069 seconds\n", 197 | "Monitored channels: \n", 198 | "\tlearning_rate\n", 199 | "\tmomentum\n", 200 | "\ttest_h0_col_norms_max\n", 201 | "\ttest_h0_col_norms_mean\n", 202 | "\ttest_h0_col_norms_min\n", 203 | "\ttest_h0_max_x_max_u\n", 204 | "\ttest_h0_max_x_mean_u\n", 205 | "\ttest_h0_max_x_min_u\n", 206 | "\ttest_h0_mean_x_max_u\n", 207 | "\ttest_h0_mean_x_mean_u\n", 208 | "\ttest_h0_mean_x_min_u\n", 209 | "\ttest_h0_min_x_max_u\n", 210 | "\ttest_h0_min_x_mean_u\n", 211 | "\ttest_h0_min_x_min_u\n", 212 | "\ttest_h0_range_x_max_u\n", 213 | "\ttest_h0_range_x_mean_u\n", 214 | "\ttest_h0_range_x_min_u\n", 215 | "\ttest_h0_row_norms_max\n", 216 | "\ttest_h0_row_norms_mean\n", 217 | "\ttest_h0_row_norms_min\n", 218 | "\ttest_h1_col_norms_max\n", 219 | "\ttest_h1_col_norms_mean\n", 220 | "\ttest_h1_col_norms_min\n", 221 | "\ttest_h1_max_x_max_u\n", 222 | "\ttest_h1_max_x_mean_u\n", 223 | "\ttest_h1_max_x_min_u\n", 224 | "\ttest_h1_mean_x_max_u\n", 225 | "\ttest_h1_mean_x_mean_u\n", 226 | "\ttest_h1_mean_x_min_u\n", 227 | "\ttest_h1_min_x_max_u\n", 228 | "\ttest_h1_min_x_mean_u\n", 229 | "\ttest_h1_min_x_min_u\n", 230 | "\ttest_h1_range_x_max_u\n", 231 | "\ttest_h1_range_x_mean_u\n", 232 | "\ttest_h1_range_x_min_u\n", 233 | "\ttest_h1_row_norms_max\n", 234 | "\ttest_h1_row_norms_mean\n", 235 | "\ttest_h1_row_norms_min\n", 236 | "\ttest_objective\n", 237 | "\ttest_y_col_norms_max\n", 238 | "\ttest_y_col_norms_mean\n", 239 | "\ttest_y_col_norms_min\n", 240 | "\ttest_y_max_max_class\n", 241 | "\ttest_y_mean_max_class\n", 242 | "\ttest_y_min_max_class\n", 243 | "\ttest_y_misclass\n", 244 | "\ttest_y_nll\n", 245 | "\ttest_y_row_norms_max\n", 246 | "\ttest_y_row_norms_mean\n", 247 | "\ttest_y_row_norms_min\n", 248 | "\ttotal_seconds_last_epoch\n", 249 | "\ttrain_h0_col_norms_max\n", 250 | "\ttrain_h0_col_norms_mean\n", 251 | "\ttrain_h0_col_norms_min\n", 252 | "\ttrain_h0_max_x_max_u\n", 253 | "\ttrain_h0_max_x_mean_u\n", 254 | "\ttrain_h0_max_x_min_u\n", 255 | "\ttrain_h0_mean_x_max_u\n", 256 | "\ttrain_h0_mean_x_mean_u\n", 257 | "\ttrain_h0_mean_x_min_u\n", 258 | "\ttrain_h0_min_x_max_u\n", 259 | "\ttrain_h0_min_x_mean_u\n", 260 | "\ttrain_h0_min_x_min_u\n", 261 | "\ttrain_h0_range_x_max_u\n", 262 | "\ttrain_h0_range_x_mean_u\n", 263 | "\ttrain_h0_range_x_min_u\n", 264 | "\ttrain_h0_row_norms_max\n", 265 | "\ttrain_h0_row_norms_mean\n", 266 | "\ttrain_h0_row_norms_min\n", 267 | "\ttrain_h1_col_norms_max\n", 268 | "\ttrain_h1_col_norms_mean\n", 269 | "\ttrain_h1_col_norms_min\n", 270 | "\ttrain_h1_max_x_max_u\n", 271 | "\ttrain_h1_max_x_mean_u\n", 272 | "\ttrain_h1_max_x_min_u\n", 273 | "\ttrain_h1_mean_x_max_u\n", 274 | "\ttrain_h1_mean_x_mean_u\n", 275 | "\ttrain_h1_mean_x_min_u\n", 276 | "\ttrain_h1_min_x_max_u\n", 277 | "\ttrain_h1_min_x_mean_u\n", 278 | "\ttrain_h1_min_x_min_u\n", 279 | "\ttrain_h1_range_x_max_u\n", 280 | "\ttrain_h1_range_x_mean_u\n", 281 | "\ttrain_h1_range_x_min_u\n", 282 | "\ttrain_h1_row_norms_max\n", 283 | "\ttrain_h1_row_norms_mean\n", 284 | "\ttrain_h1_row_norms_min\n", 285 | "\ttrain_objective\n", 286 | "\ttrain_y_col_norms_max\n", 287 | "\ttrain_y_col_norms_mean\n", 288 | "\ttrain_y_col_norms_min\n", 289 | "\ttrain_y_max_max_class\n", 290 | "\ttrain_y_mean_max_class\n", 291 | "\ttrain_y_min_max_class\n", 292 | "\ttrain_y_misclass\n", 293 | "\ttrain_y_nll\n", 294 | "\ttrain_y_row_norms_max\n", 295 | "\ttrain_y_row_norms_mean\n", 296 | "\ttrain_y_row_norms_min\n", 297 | "\ttraining_seconds_this_epoch\n", 298 | "Compiling accum...\n", 299 | "graph size: 331\n", 300 | "graph size: 315\n", 301 | "Compiling accum done. Time elapsed: 10.521963 seconds\n", 302 | "Monitoring step:\n", 303 | "\tEpochs seen: 0\n", 304 | "\tBatches seen: 0\n", 305 | "\tExamples seen: 0\n", 306 | "\tlearning_rate: 0.099999986589\n", 307 | "\tmomentum: 0.500000059605\n", 308 | "\ttest_h0_col_norms_max: 1.06272411346\n", 309 | "\ttest_h0_col_norms_mean: 1.00180494785\n", 310 | "\ttest_h0_col_norms_min: 0.94033241272\n", 311 | "\ttest_h0_max_x_max_u: 1.44173371792\n", 312 | "\ttest_h0_max_x_mean_u: 0.644945144653\n", 313 | "\ttest_h0_max_x_min_u: 0.0393124967813\n", 314 | "\ttest_h0_mean_x_max_u: 0.649418830872\n", 315 | "\ttest_h0_mean_x_mean_u: 0.133668914437\n", 316 | "\ttest_h0_mean_x_min_u: 0.000591475865804\n", 317 | "\ttest_h0_min_x_max_u: 0.0275801625103\n", 318 | "\ttest_h0_min_x_mean_u: 0.000275801605312\n", 319 | "\ttest_h0_min_x_min_u: 0.0\n", 320 | "\ttest_h0_range_x_max_u: 1.41902625561\n", 321 | "\ttest_h0_range_x_mean_u: 0.644669353962\n", 322 | "\ttest_h0_range_x_min_u: 0.0393124967813\n", 323 | "\ttest_h0_row_norms_max: 0.43685349822\n", 324 | "\ttest_h0_row_norms_mean: 0.356941491365\n", 325 | "\ttest_h0_row_norms_min: 0.280473798513\n", 326 | "\ttest_h1_col_norms_max: 1.20874929428\n", 327 | "\ttest_h1_col_norms_mean: 1.00676298141\n", 328 | "\ttest_h1_col_norms_min: 0.806728541851\n", 329 | "\ttest_h1_max_x_max_u: 1.00942623615\n", 330 | "\ttest_h1_max_x_mean_u: 0.406806230545\n", 331 | "\ttest_h1_max_x_min_u: 0.0\n", 332 | "\ttest_h1_mean_x_max_u: 0.395200431347\n", 333 | "\ttest_h1_mean_x_mean_u: 0.0980800911784\n", 334 | "\ttest_h1_mean_x_min_u: 0.0\n", 335 | "\ttest_h1_min_x_max_u: 0.0461700670421\n", 336 | "\ttest_h1_min_x_mean_u: 0.000567106995732\n", 337 | "\ttest_h1_min_x_min_u: 0.0\n", 338 | "\ttest_h1_range_x_max_u: 0.996873319149\n", 339 | "\ttest_h1_range_x_mean_u: 0.406239151955\n", 340 | "\ttest_h1_range_x_min_u: 0.0\n", 341 | "\ttest_h1_row_norms_max: 1.16290199757\n", 342 | "\ttest_h1_row_norms_mean: 1.0068975687\n", 343 | "\ttest_h1_row_norms_min: 0.864736914635\n", 344 | "\ttest_objective: 2.30256080627\n", 345 | "\ttest_y_col_norms_max: 0.0103744743392\n", 346 | "\ttest_y_col_norms_mean: 0.009660246782\n", 347 | "\ttest_y_col_norms_min: 0.00870714895427\n", 348 | "\ttest_y_max_max_class: 0.100585304201\n", 349 | "\ttest_y_mean_max_class: 0.100267209113\n", 350 | "\ttest_y_min_max_class: 0.100082024932\n", 351 | "\ttest_y_misclass: 0.879999995232\n", 352 | "\ttest_y_nll: 2.30256080627\n", 353 | "\ttest_y_row_norms_max: 0.00512082641944\n", 354 | "\ttest_y_row_norms_mean: 0.00298687024042\n", 355 | "\ttest_y_row_norms_min: 0.0010757187847\n", 356 | "\ttotal_seconds_last_epoch: 0.0\n", 357 | "\ttrain_h0_col_norms_max: 1.06272411346\n", 358 | "\ttrain_h0_col_norms_mean: 1.00180494785\n", 359 | "\ttrain_h0_col_norms_min: 0.94033241272\n", 360 | "\ttrain_h0_max_x_max_u: 1.38562130928\n", 361 | "\ttrain_h0_max_x_mean_u: 0.635100126266\n", 362 | "\ttrain_h0_max_x_min_u: 0.0533090718091\n", 363 | "\ttrain_h0_mean_x_max_u: 0.628859221935\n", 364 | "\ttrain_h0_mean_x_mean_u: 0.131119459867\n", 365 | "\ttrain_h0_mean_x_min_u: 0.00080011296086\n", 366 | "\ttrain_h0_min_x_max_u: 0.0397322028875\n", 367 | "\ttrain_h0_min_x_mean_u: 0.000397322059143\n", 368 | "\ttrain_h0_min_x_min_u: 0.0\n", 369 | "\ttrain_h0_range_x_max_u: 1.38011538982\n", 370 | "\ttrain_h0_range_x_mean_u: 0.634702801704\n", 371 | "\ttrain_h0_range_x_min_u: 0.0533090718091\n", 372 | "\ttrain_h0_row_norms_max: 0.43685349822\n", 373 | "\ttrain_h0_row_norms_mean: 0.356941491365\n", 374 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 375 | "\ttrain_h1_col_norms_max: 1.20874929428\n", 376 | "\ttrain_h1_col_norms_mean: 1.00676298141\n", 377 | "\ttrain_h1_col_norms_min: 0.806728541851\n", 378 | "\ttrain_h1_max_x_max_u: 0.959411621094\n", 379 | "\ttrain_h1_max_x_mean_u: 0.399134337902\n", 380 | "\ttrain_h1_max_x_min_u: 0.0\n", 381 | "\ttrain_h1_mean_x_max_u: 0.402670681477\n", 382 | "\ttrain_h1_mean_x_mean_u: 0.0960004329681\n", 383 | "\ttrain_h1_mean_x_min_u: 0.0\n", 384 | "\ttrain_h1_min_x_max_u: 0.0696794986725\n", 385 | "\ttrain_h1_min_x_mean_u: 0.000998697709292\n", 386 | "\ttrain_h1_min_x_min_u: 0.0\n", 387 | "\ttrain_h1_range_x_max_u: 0.939600527287\n", 388 | "\ttrain_h1_range_x_mean_u: 0.398135662079\n", 389 | "\ttrain_h1_range_x_min_u: 0.0\n", 390 | "\ttrain_h1_row_norms_max: 1.16290199757\n", 391 | "\ttrain_h1_row_norms_mean: 1.0068975687\n", 392 | "\ttrain_h1_row_norms_min: 0.864736914635\n", 393 | "\ttrain_objective: 2.30260586739\n", 394 | "\ttrain_y_col_norms_max: 0.0103744743392\n", 395 | "\ttrain_y_col_norms_mean: 0.009660246782\n", 396 | "\ttrain_y_col_norms_min: 0.00870714895427\n", 397 | "\ttrain_y_max_max_class: 0.100583769381\n", 398 | "\ttrain_y_mean_max_class: 0.100261345506\n", 399 | "\ttrain_y_min_max_class: 0.100080445409\n", 400 | "\ttrain_y_misclass: 0.89099997282\n", 401 | "\ttrain_y_nll: 2.30260586739\n", 402 | "\ttrain_y_row_norms_max: 0.00512082641944\n", 403 | "\ttrain_y_row_norms_mean: 0.00298687024042\n", 404 | "\ttrain_y_row_norms_min: 0.0010757187847\n", 405 | "\ttraining_seconds_this_epoch: 0.0\n", 406 | "Time this epoch: 0.076504 seconds\n", 407 | "Monitoring step:\n", 408 | "\tEpochs seen: 1\n", 409 | "\tBatches seen: 10\n", 410 | "\tExamples seen: 1000\n", 411 | "\tlearning_rate: 0.099999986589\n", 412 | "\tmomentum: 0.500000059605\n", 413 | "\ttest_h0_col_norms_max: 1.0627515316\n", 414 | "\ttest_h0_col_norms_mean: 1.00198495388\n", 415 | "\ttest_h0_col_norms_min: 0.940323114395\n", 416 | "\ttest_h0_max_x_max_u: 1.4717758894\n", 417 | "\ttest_h0_max_x_mean_u: 0.66315984726\n", 418 | "\ttest_h0_max_x_min_u: 0.0384505018592\n", 419 | "\ttest_h0_mean_x_max_u: 0.657794594765\n", 420 | "\ttest_h0_mean_x_mean_u: 0.140545859933\n", 421 | "\ttest_h0_mean_x_min_u: 0.000591710035224\n", 422 | "\ttest_h0_min_x_max_u: 0.0340499989688\n", 423 | "\ttest_h0_min_x_mean_u: 0.000340500002494\n", 424 | "\ttest_h0_min_x_min_u: 0.0\n", 425 | "\ttest_h0_range_x_max_u: 1.45297896862\n", 426 | "\ttest_h0_range_x_mean_u: 0.662819385529\n", 427 | "\ttest_h0_range_x_min_u: 0.0384505018592\n", 428 | "\ttest_h0_row_norms_max: 0.43685528636\n", 429 | "\ttest_h0_row_norms_mean: 0.35700532794\n", 430 | "\ttest_h0_row_norms_min: 0.280473798513\n", 431 | "\ttest_h1_col_norms_max: 1.20875167847\n", 432 | "\ttest_h1_col_norms_mean: 1.00693702698\n", 433 | "\ttest_h1_col_norms_min: 0.807069003582\n", 434 | "\ttest_h1_max_x_max_u: 1.0623421669\n", 435 | "\ttest_h1_max_x_mean_u: 0.428363919258\n", 436 | "\ttest_h1_max_x_min_u: 0.0\n", 437 | "\ttest_h1_mean_x_max_u: 0.397820621729\n", 438 | "\ttest_h1_mean_x_mean_u: 0.104200258851\n", 439 | "\ttest_h1_mean_x_min_u: 0.0\n", 440 | "\ttest_h1_min_x_max_u: 0.0317041091621\n", 441 | "\ttest_h1_min_x_mean_u: 0.000390867382521\n", 442 | "\ttest_h1_min_x_min_u: 0.0\n", 443 | "\ttest_h1_range_x_max_u: 1.0562684536\n", 444 | "\ttest_h1_range_x_mean_u: 0.427973091602\n", 445 | "\ttest_h1_range_x_min_u: 0.0\n", 446 | "\ttest_h1_row_norms_max: 1.16293859482\n", 447 | "\ttest_h1_row_norms_mean: 1.00707542896\n", 448 | "\ttest_h1_row_norms_min: 0.864998340607\n", 449 | "\ttest_objective: 2.2671046257\n", 450 | "\ttest_y_col_norms_max: 0.11124061048\n", 451 | "\ttest_y_col_norms_mean: 0.0754992589355\n", 452 | "\ttest_y_col_norms_min: 0.053731944412\n", 453 | "\ttest_y_max_max_class: 0.115364506841\n", 454 | "\ttest_y_mean_max_class: 0.106587290764\n", 455 | "\ttest_y_min_max_class: 0.102818973362\n", 456 | "\ttest_y_misclass: 0.584999978542\n", 457 | "\ttest_y_nll: 2.2671046257\n", 458 | "\ttest_y_row_norms_max: 0.0558791719377\n", 459 | "\ttest_y_row_norms_mean: 0.0203072223812\n", 460 | "\ttest_y_row_norms_min: 0.00219676503912\n", 461 | "\ttotal_seconds_last_epoch: 0.0\n", 462 | "\ttrain_h0_col_norms_max: 1.0627515316\n", 463 | "\ttrain_h0_col_norms_mean: 1.00198495388\n", 464 | "\ttrain_h0_col_norms_min: 0.940323114395\n", 465 | "\ttrain_h0_max_x_max_u: 1.42723119259\n", 466 | "\ttrain_h0_max_x_mean_u: 0.652235984802\n", 467 | "\ttrain_h0_max_x_min_u: 0.0534820333123\n", 468 | "\ttrain_h0_mean_x_max_u: 0.638325154781\n", 469 | "\ttrain_h0_mean_x_mean_u: 0.137875661254\n", 470 | "\ttrain_h0_mean_x_min_u: 0.000787791388575\n", 471 | "\ttrain_h0_min_x_max_u: 0.0474192835391\n", 472 | "\ttrain_h0_min_x_mean_u: 0.000474192842375\n", 473 | "\ttrain_h0_min_x_min_u: 0.0\n", 474 | "\ttrain_h0_range_x_max_u: 1.42723119259\n", 475 | "\ttrain_h0_range_x_mean_u: 0.651761770248\n", 476 | "\ttrain_h0_range_x_min_u: 0.0534820333123\n", 477 | "\ttrain_h0_row_norms_max: 0.43685528636\n", 478 | "\ttrain_h0_row_norms_mean: 0.35700532794\n", 479 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 480 | "\ttrain_h1_col_norms_max: 1.20875167847\n", 481 | "\ttrain_h1_col_norms_mean: 1.00693702698\n", 482 | "\ttrain_h1_col_norms_min: 0.807069003582\n", 483 | "\ttrain_h1_max_x_max_u: 1.02196538448\n", 484 | "\ttrain_h1_max_x_mean_u: 0.419982761145\n", 485 | "\ttrain_h1_max_x_min_u: 0.0\n", 486 | "\ttrain_h1_mean_x_max_u: 0.406646072865\n", 487 | "\ttrain_h1_mean_x_mean_u: 0.102122284472\n", 488 | "\ttrain_h1_mean_x_min_u: 0.0\n", 489 | "\ttrain_h1_min_x_max_u: 0.0540733486414\n", 490 | "\ttrain_h1_min_x_mean_u: 0.000737457594369\n", 491 | "\ttrain_h1_min_x_min_u: 0.0\n", 492 | "\ttrain_h1_range_x_max_u: 1.01091790199\n", 493 | "\ttrain_h1_range_x_mean_u: 0.419245332479\n", 494 | "\ttrain_h1_range_x_min_u: 0.0\n", 495 | "\ttrain_h1_row_norms_max: 1.16293859482\n", 496 | "\ttrain_h1_row_norms_mean: 1.00707542896\n", 497 | "\ttrain_h1_row_norms_min: 0.864998340607\n", 498 | "\ttrain_objective: 2.26396894455\n", 499 | "\ttrain_y_col_norms_max: 0.11124061048\n", 500 | "\ttrain_y_col_norms_mean: 0.0754992589355\n", 501 | "\ttrain_y_col_norms_min: 0.053731944412\n", 502 | "\ttrain_y_max_max_class: 0.115357078612\n", 503 | "\ttrain_y_mean_max_class: 0.106709964573\n", 504 | "\ttrain_y_min_max_class: 0.102847002447\n", 505 | "\ttrain_y_misclass: 0.569999992847\n", 506 | "\ttrain_y_nll: 2.26396894455\n", 507 | "\ttrain_y_row_norms_max: 0.0558791719377\n", 508 | "\ttrain_y_row_norms_mean: 0.0203072223812\n", 509 | "\ttrain_y_row_norms_min: 0.00219676503912\n", 510 | "\ttraining_seconds_this_epoch: 0.0765039995313\n", 511 | "Time this epoch: 0.079490 seconds\n", 512 | "Monitoring step:\n", 513 | "\tEpochs seen: 2\n", 514 | "\tBatches seen: 20\n", 515 | "\tExamples seen: 2000\n", 516 | "\tlearning_rate: 0.099999986589\n", 517 | "\tmomentum: 0.500000059605\n", 518 | "\ttest_h0_col_norms_max: 1.06305730343\n", 519 | "\ttest_h0_col_norms_mean: 1.00336956978\n", 520 | "\ttest_h0_col_norms_min: 0.940306901932\n", 521 | "\ttest_h0_max_x_max_u: 1.83217453957\n", 522 | "\ttest_h0_max_x_mean_u: 0.77869194746\n", 523 | "\ttest_h0_max_x_min_u: 0.0376306697726\n", 524 | "\ttest_h0_mean_x_max_u: 0.814771950245\n", 525 | "\ttest_h0_mean_x_mean_u: 0.189053744078\n", 526 | "\ttest_h0_mean_x_min_u: 0.000595318269916\n", 527 | "\ttest_h0_min_x_max_u: 0.151219651103\n", 528 | "\ttest_h0_min_x_mean_u: 0.00151219638065\n", 529 | "\ttest_h0_min_x_min_u: 0.0\n", 530 | "\ttest_h0_range_x_max_u: 1.81660676003\n", 531 | "\ttest_h0_range_x_mean_u: 0.777179718018\n", 532 | "\ttest_h0_range_x_min_u: 0.0376306697726\n", 533 | "\ttest_h0_row_norms_max: 0.436813563108\n", 534 | "\ttest_h0_row_norms_mean: 0.357498139143\n", 535 | "\ttest_h0_row_norms_min: 0.280473798513\n", 536 | "\ttest_h1_col_norms_max: 1.20968019962\n", 537 | "\ttest_h1_col_norms_mean: 1.00822615623\n", 538 | "\ttest_h1_col_norms_min: 0.809271395206\n", 539 | "\ttest_h1_max_x_max_u: 1.69195461273\n", 540 | "\ttest_h1_max_x_mean_u: 0.57115894556\n", 541 | "\ttest_h1_max_x_min_u: 0.0\n", 542 | "\ttest_h1_mean_x_max_u: 0.546870112419\n", 543 | "\ttest_h1_mean_x_mean_u: 0.151245683432\n", 544 | "\ttest_h1_mean_x_min_u: 0.0\n", 545 | "\ttest_h1_min_x_max_u: 0.0247348379344\n", 546 | "\ttest_h1_min_x_mean_u: 0.000320939114317\n", 547 | "\ttest_h1_min_x_min_u: 0.0\n", 548 | "\ttest_h1_range_x_max_u: 1.69195461273\n", 549 | "\ttest_h1_range_x_mean_u: 0.570838034153\n", 550 | "\ttest_h1_range_x_min_u: 0.0\n", 551 | "\ttest_h1_row_norms_max: 1.16397964954\n", 552 | "\ttest_h1_row_norms_mean: 1.00838541985\n", 553 | "\ttest_h1_row_norms_min: 0.867279112339\n", 554 | "\ttest_objective: 2.14843034744\n", 555 | "\ttest_y_col_norms_max: 0.294056743383\n", 556 | "\ttest_y_col_norms_mean: 0.183990433812\n", 557 | "\ttest_y_col_norms_min: 0.118449114263\n", 558 | "\ttest_y_max_max_class: 0.216087907553\n", 559 | "\ttest_y_mean_max_class: 0.128736883402\n", 560 | "\ttest_y_min_max_class: 0.106923222542\n", 561 | "\ttest_y_misclass: 0.575000047684\n", 562 | "\ttest_y_nll: 2.14843034744\n", 563 | "\ttest_y_row_norms_max: 0.150932133198\n", 564 | "\ttest_y_row_norms_mean: 0.0488044582307\n", 565 | "\ttest_y_row_norms_min: 0.00257564499043\n", 566 | "\ttotal_seconds_last_epoch: 0.399715006351\n", 567 | "\ttrain_h0_col_norms_max: 1.06305730343\n", 568 | "\ttrain_h0_col_norms_mean: 1.00336956978\n", 569 | "\ttrain_h0_col_norms_min: 0.940306901932\n", 570 | "\ttrain_h0_max_x_max_u: 1.80534744263\n", 571 | "\ttrain_h0_max_x_mean_u: 0.761926949024\n", 572 | "\ttrain_h0_max_x_min_u: 0.0509638041258\n", 573 | "\ttrain_h0_mean_x_max_u: 0.787827193737\n", 574 | "\ttrain_h0_mean_x_mean_u: 0.185498431325\n", 575 | "\ttrain_h0_mean_x_min_u: 0.000702115881722\n", 576 | "\ttrain_h0_min_x_max_u: 0.136133059859\n", 577 | "\ttrain_h0_min_x_mean_u: 0.00137102941517\n", 578 | "\ttrain_h0_min_x_min_u: 0.0\n", 579 | "\ttrain_h0_range_x_max_u: 1.80470275879\n", 580 | "\ttrain_h0_range_x_mean_u: 0.760555803776\n", 581 | "\ttrain_h0_range_x_min_u: 0.0509638041258\n", 582 | "\ttrain_h0_row_norms_max: 0.436813563108\n", 583 | "\ttrain_h0_row_norms_mean: 0.357498139143\n", 584 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 585 | "\ttrain_h1_col_norms_max: 1.20968019962\n", 586 | "\ttrain_h1_col_norms_mean: 1.00822615623\n", 587 | "\ttrain_h1_col_norms_min: 0.809271395206\n", 588 | "\ttrain_h1_max_x_max_u: 1.57231628895\n", 589 | "\ttrain_h1_max_x_mean_u: 0.559122443199\n", 590 | "\ttrain_h1_max_x_min_u: 0.0\n", 591 | "\ttrain_h1_mean_x_max_u: 0.537385642529\n", 592 | "\ttrain_h1_mean_x_mean_u: 0.149182394147\n", 593 | "\ttrain_h1_mean_x_min_u: 0.0\n", 594 | "\ttrain_h1_min_x_max_u: 0.0360823534429\n", 595 | "\ttrain_h1_min_x_mean_u: 0.000581309897825\n", 596 | "\ttrain_h1_min_x_min_u: 0.0\n", 597 | "\ttrain_h1_range_x_max_u: 1.5697991848\n", 598 | "\ttrain_h1_range_x_mean_u: 0.558541119099\n", 599 | "\ttrain_h1_range_x_min_u: 0.0\n", 600 | "\ttrain_h1_row_norms_max: 1.16397964954\n", 601 | "\ttrain_h1_row_norms_mean: 1.00838541985\n", 602 | "\ttrain_h1_row_norms_min: 0.867279112339\n", 603 | "\ttrain_objective: 2.13573145866\n", 604 | "\ttrain_y_col_norms_max: 0.294056743383\n", 605 | "\ttrain_y_col_norms_mean: 0.183990433812\n", 606 | "\ttrain_y_col_norms_min: 0.118449114263\n", 607 | "\ttrain_y_max_max_class: 0.215124756098\n", 608 | "\ttrain_y_mean_max_class: 0.129534035921\n", 609 | "\ttrain_y_min_max_class: 0.106335677207\n", 610 | "\ttrain_y_misclass: 0.583000004292\n", 611 | "\ttrain_y_nll: 2.13573145866\n", 612 | "\ttrain_y_row_norms_max: 0.150932133198\n", 613 | "\ttrain_y_row_norms_mean: 0.0488044582307\n", 614 | "\ttrain_y_row_norms_min: 0.00257564499043\n", 615 | "\ttraining_seconds_this_epoch: 0.0794900134206\n", 616 | "Time this epoch: 0.076567 seconds\n", 617 | "Monitoring step:\n", 618 | "\tEpochs seen: 3\n", 619 | "\tBatches seen: 30\n", 620 | "\tExamples seen: 3000\n", 621 | "\tlearning_rate: 0.099999986589\n", 622 | "\tmomentum: 0.500000059605\n", 623 | "\ttest_h0_col_norms_max: 1.06711316109\n", 624 | "\ttest_h0_col_norms_mean: 1.00860762596\n", 625 | "\ttest_h0_col_norms_min: 0.940884590149\n", 626 | "\ttest_h0_max_x_max_u: 2.83095693588\n", 627 | "\ttest_h0_max_x_mean_u: 1.0667951107\n", 628 | "\ttest_h0_max_x_min_u: 0.0395365096629\n", 629 | "\ttest_h0_mean_x_max_u: 1.28329873085\n", 630 | "\ttest_h0_mean_x_mean_u: 0.313957154751\n", 631 | "\ttest_h0_mean_x_min_u: 0.000608601199929\n", 632 | "\ttest_h0_min_x_max_u: 0.344574779272\n", 633 | "\ttest_h0_min_x_mean_u: 0.00428696768358\n", 634 | "\ttest_h0_min_x_min_u: 0.0\n", 635 | "\ttest_h0_range_x_max_u: 2.82145404816\n", 636 | "\ttest_h0_range_x_mean_u: 1.06250810623\n", 637 | "\ttest_h0_range_x_min_u: 0.0395365096629\n", 638 | "\ttest_h0_row_norms_max: 0.436746716499\n", 639 | "\ttest_h0_row_norms_mean: 0.359357297421\n", 640 | "\ttest_h0_row_norms_min: 0.280473798513\n", 641 | "\ttest_h1_col_norms_max: 1.21659326553\n", 642 | "\ttest_h1_col_norms_mean: 1.01306569576\n", 643 | "\ttest_h1_col_norms_min: 0.816726148129\n", 644 | "\ttest_h1_max_x_max_u: 3.33260345459\n", 645 | "\ttest_h1_max_x_mean_u: 0.982680439949\n", 646 | "\ttest_h1_max_x_min_u: 0.0\n", 647 | "\ttest_h1_mean_x_max_u: 1.02735197544\n", 648 | "\ttest_h1_mean_x_mean_u: 0.287164062262\n", 649 | "\ttest_h1_mean_x_min_u: 0.0\n", 650 | "\ttest_h1_min_x_max_u: 0.0588113293052\n", 651 | "\ttest_h1_min_x_mean_u: 0.000941264093854\n", 652 | "\ttest_h1_min_x_min_u: 0.0\n", 653 | "\ttest_h1_range_x_max_u: 3.33260345459\n", 654 | "\ttest_h1_range_x_mean_u: 0.981739163399\n", 655 | "\ttest_h1_range_x_min_u: 0.0\n", 656 | "\ttest_h1_row_norms_max: 1.17268562317\n", 657 | "\ttest_h1_row_norms_mean: 1.01328611374\n", 658 | "\ttest_h1_row_norms_min: 0.875302016735\n", 659 | "\ttest_objective: 1.7220890522\n", 660 | "\ttest_y_col_norms_max: 0.57018661499\n", 661 | "\ttest_y_col_norms_mean: 0.368319630623\n", 662 | "\ttest_y_col_norms_min: 0.212889477611\n", 663 | "\ttest_y_max_max_class: 0.649778485298\n", 664 | "\ttest_y_mean_max_class: 0.22110798955\n", 665 | "\ttest_y_min_max_class: 0.120734199882\n", 666 | "\ttest_y_misclass: 0.408999979496\n", 667 | "\ttest_y_nll: 1.7220890522\n", 668 | "\ttest_y_row_norms_max: 0.283712416887\n", 669 | "\ttest_y_row_norms_mean: 0.0966741219163\n", 670 | "\ttest_y_row_norms_min: 0.00265311892144\n", 671 | "\ttotal_seconds_last_epoch: 0.39892795682\n", 672 | "\ttrain_h0_col_norms_max: 1.06711316109\n", 673 | "\ttrain_h0_col_norms_mean: 1.00860762596\n", 674 | "\ttrain_h0_col_norms_min: 0.940884590149\n", 675 | "\ttrain_h0_max_x_max_u: 2.80888795853\n", 676 | "\ttrain_h0_max_x_mean_u: 1.04323256016\n", 677 | "\ttrain_h0_max_x_min_u: 0.0423219464719\n", 678 | "\ttrain_h0_mean_x_max_u: 1.27551007271\n", 679 | "\ttrain_h0_mean_x_mean_u: 0.30756020546\n", 680 | "\ttrain_h0_mean_x_min_u: 0.000561867607757\n", 681 | "\ttrain_h0_min_x_max_u: 0.331990122795\n", 682 | "\ttrain_h0_min_x_mean_u: 0.00414318405092\n", 683 | "\ttrain_h0_min_x_min_u: 0.0\n", 684 | "\ttrain_h0_range_x_max_u: 2.80888795853\n", 685 | "\ttrain_h0_range_x_mean_u: 1.0390894413\n", 686 | "\ttrain_h0_range_x_min_u: 0.0423219464719\n", 687 | "\ttrain_h0_row_norms_max: 0.436746716499\n", 688 | "\ttrain_h0_row_norms_mean: 0.359357297421\n", 689 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 690 | "\ttrain_h1_col_norms_max: 1.21659326553\n", 691 | "\ttrain_h1_col_norms_mean: 1.01306569576\n", 692 | "\ttrain_h1_col_norms_min: 0.816726148129\n", 693 | "\ttrain_h1_max_x_max_u: 3.0735347271\n", 694 | "\ttrain_h1_max_x_mean_u: 0.962003052235\n", 695 | "\ttrain_h1_max_x_min_u: 0.0\n", 696 | "\ttrain_h1_mean_x_max_u: 0.984959363937\n", 697 | "\ttrain_h1_mean_x_mean_u: 0.283362656832\n", 698 | "\ttrain_h1_mean_x_min_u: 0.0\n", 699 | "\ttrain_h1_min_x_max_u: 0.0628297924995\n", 700 | "\ttrain_h1_min_x_mean_u: 0.000931099057198\n", 701 | "\ttrain_h1_min_x_min_u: 0.0\n", 702 | "\ttrain_h1_range_x_max_u: 3.0735347271\n", 703 | "\ttrain_h1_range_x_mean_u: 0.961071968079\n", 704 | "\ttrain_h1_range_x_min_u: 0.0\n", 705 | "\ttrain_h1_row_norms_max: 1.17268562317\n", 706 | "\ttrain_h1_row_norms_mean: 1.01328611374\n", 707 | "\ttrain_h1_row_norms_min: 0.875302016735\n", 708 | "\ttrain_objective: 1.68255245686\n", 709 | "\ttrain_y_col_norms_max: 0.57018661499\n", 710 | "\ttrain_y_col_norms_mean: 0.368319630623\n", 711 | "\ttrain_y_col_norms_min: 0.212889477611\n", 712 | "\ttrain_y_max_max_class: 0.660824656487\n", 713 | "\ttrain_y_mean_max_class: 0.226250484586\n", 714 | "\ttrain_y_min_max_class: 0.118295103312\n", 715 | "\ttrain_y_misclass: 0.373000025749\n", 716 | "\ttrain_y_nll: 1.68255245686\n", 717 | "\ttrain_y_row_norms_max: 0.283712416887\n", 718 | "\ttrain_y_row_norms_mean: 0.0966741219163\n", 719 | "\ttrain_y_row_norms_min: 0.00265311892144\n", 720 | "\ttraining_seconds_this_epoch: 0.0765670016408\n", 721 | "Time this epoch: 0.076755 seconds\n", 722 | "Monitoring step:\n", 723 | "\tEpochs seen: 4\n", 724 | "\tBatches seen: 40\n", 725 | "\tExamples seen: 4000\n", 726 | "\tlearning_rate: 0.099999986589\n", 727 | "\tmomentum: 0.500000059605\n", 728 | "\ttest_h0_col_norms_max: 1.11460793018\n", 729 | "\ttest_h0_col_norms_mean: 1.01910960674\n", 730 | "\ttest_h0_col_norms_min: 0.943947970867\n", 731 | "\ttest_h0_max_x_max_u: 4.03108596802\n", 732 | "\ttest_h0_max_x_mean_u: 1.44839954376\n", 733 | "\ttest_h0_max_x_min_u: 0.0350178405643\n", 734 | "\ttest_h0_mean_x_max_u: 1.80931425095\n", 735 | "\ttest_h0_mean_x_mean_u: 0.476316422224\n", 736 | "\ttest_h0_mean_x_min_u: 0.00059692573268\n", 737 | "\ttest_h0_min_x_max_u: 0.536396503448\n", 738 | "\ttest_h0_min_x_mean_u: 0.00886026769876\n", 739 | "\ttest_h0_min_x_min_u: 0.0\n", 740 | "\ttest_h0_range_x_max_u: 4.0198469162\n", 741 | "\ttest_h0_range_x_mean_u: 1.43953919411\n", 742 | "\ttest_h0_range_x_min_u: 0.0350178405643\n", 743 | "\ttest_h0_row_norms_max: 0.443038374186\n", 744 | "\ttest_h0_row_norms_mean: 0.363063722849\n", 745 | "\ttest_h0_row_norms_min: 0.280473798513\n", 746 | "\ttest_h1_col_norms_max: 1.23125362396\n", 747 | "\ttest_h1_col_norms_mean: 1.02282571793\n", 748 | "\ttest_h1_col_norms_min: 0.831244647503\n", 749 | "\ttest_h1_max_x_max_u: 5.4942908287\n", 750 | "\ttest_h1_max_x_mean_u: 1.63247060776\n", 751 | "\ttest_h1_max_x_min_u: 0.0\n", 752 | "\ttest_h1_mean_x_max_u: 1.69427454472\n", 753 | "\ttest_h1_mean_x_mean_u: 0.516746878624\n", 754 | "\ttest_h1_mean_x_min_u: 0.0\n", 755 | "\ttest_h1_min_x_max_u: 0.0896493941545\n", 756 | "\ttest_h1_min_x_mean_u: 0.00180161022581\n", 757 | "\ttest_h1_min_x_min_u: 0.0\n", 758 | "\ttest_h1_range_x_max_u: 5.4942908287\n", 759 | "\ttest_h1_range_x_mean_u: 1.63066899776\n", 760 | "\ttest_h1_range_x_min_u: 0.0\n", 761 | "\ttest_h1_row_norms_max: 1.19547796249\n", 762 | "\ttest_h1_row_norms_mean: 1.02314937115\n", 763 | "\ttest_h1_row_norms_min: 0.889712035656\n", 764 | "\ttest_objective: 1.15405845642\n", 765 | "\ttest_y_col_norms_max: 0.821692407131\n", 766 | "\ttest_y_col_norms_mean: 0.59273904562\n", 767 | "\ttest_y_col_norms_min: 0.367552161217\n", 768 | "\ttest_y_max_max_class: 0.93451744318\n", 769 | "\ttest_y_mean_max_class: 0.406836867332\n", 770 | "\ttest_y_min_max_class: 0.156486883759\n", 771 | "\ttest_y_misclass: 0.288999974728\n", 772 | "\ttest_y_nll: 1.15405845642\n", 773 | "\ttest_y_row_norms_max: 0.399145454168\n", 774 | "\ttest_y_row_norms_mean: 0.153999328613\n", 775 | "\ttest_y_row_norms_min: 0.00266336160712\n", 776 | "\ttotal_seconds_last_epoch: 0.391640037298\n", 777 | "\ttrain_h0_col_norms_max: 1.11460793018\n", 778 | "\ttrain_h0_col_norms_mean: 1.01910960674\n", 779 | "\ttrain_h0_col_norms_min: 0.943947970867\n", 780 | "\ttrain_h0_max_x_max_u: 3.86370563507\n", 781 | "\ttrain_h0_max_x_mean_u: 1.41748452187\n", 782 | "\ttrain_h0_max_x_min_u: 0.0320980213583\n", 783 | "\ttrain_h0_mean_x_max_u: 1.78839421272\n", 784 | "\ttrain_h0_mean_x_mean_u: 0.465445041656\n", 785 | "\ttrain_h0_mean_x_min_u: 0.000428176659625\n", 786 | "\ttrain_h0_min_x_max_u: 0.533168613911\n", 787 | "\ttrain_h0_min_x_mean_u: 0.00889937113971\n", 788 | "\ttrain_h0_min_x_min_u: 0.0\n", 789 | "\ttrain_h0_range_x_max_u: 3.86141324043\n", 790 | "\ttrain_h0_range_x_mean_u: 1.40858519077\n", 791 | "\ttrain_h0_range_x_min_u: 0.0320980213583\n", 792 | "\ttrain_h0_row_norms_max: 0.443038374186\n", 793 | "\ttrain_h0_row_norms_mean: 0.363063722849\n", 794 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 795 | "\ttrain_h1_col_norms_max: 1.23125362396\n", 796 | "\ttrain_h1_col_norms_mean: 1.02282571793\n", 797 | "\ttrain_h1_col_norms_min: 0.831244647503\n", 798 | "\ttrain_h1_max_x_max_u: 5.08467340469\n", 799 | "\ttrain_h1_max_x_mean_u: 1.5936088562\n", 800 | "\ttrain_h1_max_x_min_u: 0.0\n", 801 | "\ttrain_h1_mean_x_max_u: 1.67192387581\n", 802 | "\ttrain_h1_mean_x_mean_u: 0.508422672749\n", 803 | "\ttrain_h1_mean_x_min_u: 0.0\n", 804 | "\ttrain_h1_min_x_max_u: 0.101231276989\n", 805 | "\ttrain_h1_min_x_mean_u: 0.0018501750892\n", 806 | "\ttrain_h1_min_x_min_u: 0.0\n", 807 | "\ttrain_h1_range_x_max_u: 5.08467340469\n", 808 | "\ttrain_h1_range_x_mean_u: 1.59175896645\n", 809 | "\ttrain_h1_range_x_min_u: 0.0\n", 810 | "\ttrain_h1_row_norms_max: 1.19547796249\n", 811 | "\ttrain_h1_row_norms_mean: 1.02314937115\n", 812 | "\ttrain_h1_row_norms_min: 0.889712035656\n", 813 | "\ttrain_objective: 1.06958007812\n", 814 | "\ttrain_y_col_norms_max: 0.821692407131\n", 815 | "\ttrain_y_col_norms_mean: 0.59273904562\n", 816 | "\ttrain_y_col_norms_min: 0.367552161217\n", 817 | "\ttrain_y_max_max_class: 0.939358353615\n", 818 | "\ttrain_y_mean_max_class: 0.427326142788\n", 819 | "\ttrain_y_min_max_class: 0.152194455266\n", 820 | "\ttrain_y_misclass: 0.230999976397\n", 821 | "\ttrain_y_nll: 1.06958007812\n", 822 | "\ttrain_y_row_norms_max: 0.399145454168\n", 823 | "\ttrain_y_row_norms_mean: 0.153999328613\n", 824 | "\ttrain_y_row_norms_min: 0.00266336160712\n", 825 | "\ttraining_seconds_this_epoch: 0.0767549946904\n", 826 | "Time this epoch: 0.076340 seconds\n", 827 | "Monitoring step:\n", 828 | "\tEpochs seen: 5\n", 829 | "\tBatches seen: 50\n", 830 | "\tExamples seen: 5000\n", 831 | "\tlearning_rate: 0.099999986589\n", 832 | "\tmomentum: 0.500000059605\n", 833 | "\ttest_h0_col_norms_max: 1.14932739735\n", 834 | "\ttest_h0_col_norms_mean: 1.02921593189\n", 835 | "\ttest_h0_col_norms_min: 0.947646379471\n", 836 | "\ttest_h0_max_x_max_u: 4.65499973297\n", 837 | "\ttest_h0_max_x_mean_u: 1.69929552078\n", 838 | "\ttest_h0_max_x_min_u: 0.0342949405313\n", 839 | "\ttest_h0_mean_x_max_u: 2.08451151848\n", 840 | "\ttest_h0_mean_x_mean_u: 0.575312197208\n", 841 | "\ttest_h0_mean_x_min_u: 0.000570055155549\n", 842 | "\ttest_h0_min_x_max_u: 0.607917308807\n", 843 | "\ttest_h0_min_x_mean_u: 0.0106261633337\n", 844 | "\ttest_h0_min_x_min_u: 0.0\n", 845 | "\ttest_h0_range_x_max_u: 4.64167976379\n", 846 | "\ttest_h0_range_x_mean_u: 1.68866920471\n", 847 | "\ttest_h0_range_x_min_u: 0.0342949405313\n", 848 | "\ttest_h0_row_norms_max: 0.45808044076\n", 849 | "\ttest_h0_row_norms_mean: 0.366596549749\n", 850 | "\ttest_h0_row_norms_min: 0.280473798513\n", 851 | "\ttest_h1_col_norms_max: 1.24947822094\n", 852 | "\ttest_h1_col_norms_mean: 1.0322893858\n", 853 | "\ttest_h1_col_norms_min: 0.845647394657\n", 854 | "\ttest_h1_max_x_max_u: 6.84989500046\n", 855 | "\ttest_h1_max_x_mean_u: 2.11669826508\n", 856 | "\ttest_h1_max_x_min_u: 0.0\n", 857 | "\ttest_h1_mean_x_max_u: 2.13412213326\n", 858 | "\ttest_h1_mean_x_mean_u: 0.684898376465\n", 859 | "\ttest_h1_mean_x_min_u: 0.0\n", 860 | "\ttest_h1_min_x_max_u: 0.142428070307\n", 861 | "\ttest_h1_min_x_mean_u: 0.00252998713404\n", 862 | "\ttest_h1_min_x_min_u: 0.0\n", 863 | "\ttest_h1_range_x_max_u: 6.84989500046\n", 864 | "\ttest_h1_range_x_mean_u: 2.11416840553\n", 865 | "\ttest_h1_range_x_min_u: 0.0\n", 866 | "\ttest_h1_row_norms_max: 1.21548783779\n", 867 | "\ttest_h1_row_norms_mean: 1.03275668621\n", 868 | "\ttest_h1_row_norms_min: 0.889980256557\n", 869 | "\ttest_objective: 0.834180235863\n", 870 | "\ttest_y_col_norms_max: 1.00634145737\n", 871 | "\ttest_y_col_norms_mean: 0.751444160938\n", 872 | "\ttest_y_col_norms_min: 0.522492587566\n", 873 | "\ttest_y_max_max_class: 0.980666279793\n", 874 | "\ttest_y_mean_max_class: 0.581911981106\n", 875 | "\ttest_y_min_max_class: 0.200448006392\n", 876 | "\ttest_y_misclass: 0.237000003457\n", 877 | "\ttest_y_nll: 0.834180235863\n", 878 | "\ttest_y_row_norms_max: 0.486268281937\n", 879 | "\ttest_y_row_norms_mean: 0.195263013244\n", 880 | "\ttest_y_row_norms_min: 0.00266337138601\n", 881 | "\ttotal_seconds_last_epoch: 0.391252964735\n", 882 | "\ttrain_h0_col_norms_max: 1.14932739735\n", 883 | "\ttrain_h0_col_norms_mean: 1.02921593189\n", 884 | "\ttrain_h0_col_norms_min: 0.947646379471\n", 885 | "\ttrain_h0_max_x_max_u: 4.44496536255\n", 886 | "\ttrain_h0_max_x_mean_u: 1.66401827335\n", 887 | "\ttrain_h0_max_x_min_u: 0.0278791002929\n", 888 | "\ttrain_h0_mean_x_max_u: 2.05176591873\n", 889 | "\ttrain_h0_mean_x_mean_u: 0.561449885368\n", 890 | "\ttrain_h0_mean_x_min_u: 0.00037754495861\n", 891 | "\ttrain_h0_min_x_max_u: 0.61878490448\n", 892 | "\ttrain_h0_min_x_mean_u: 0.0100459652022\n", 893 | "\ttrain_h0_min_x_min_u: 0.0\n", 894 | "\ttrain_h0_range_x_max_u: 4.43668365479\n", 895 | "\ttrain_h0_range_x_mean_u: 1.65397238731\n", 896 | "\ttrain_h0_range_x_min_u: 0.0278791002929\n", 897 | "\ttrain_h0_row_norms_max: 0.45808044076\n", 898 | "\ttrain_h0_row_norms_mean: 0.366596549749\n", 899 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 900 | "\ttrain_h1_col_norms_max: 1.24947822094\n", 901 | "\ttrain_h1_col_norms_mean: 1.0322893858\n", 902 | "\ttrain_h1_col_norms_min: 0.845647394657\n", 903 | "\ttrain_h1_max_x_max_u: 6.43478679657\n", 904 | "\ttrain_h1_max_x_mean_u: 2.06896901131\n", 905 | "\ttrain_h1_max_x_min_u: 0.0\n", 906 | "\ttrain_h1_mean_x_max_u: 2.11113262177\n", 907 | "\ttrain_h1_mean_x_mean_u: 0.672600626945\n", 908 | "\ttrain_h1_mean_x_min_u: 0.0\n", 909 | "\ttrain_h1_min_x_max_u: 0.181353420019\n", 910 | "\ttrain_h1_min_x_mean_u: 0.00286162574776\n", 911 | "\ttrain_h1_min_x_min_u: 0.0\n", 912 | "\ttrain_h1_range_x_max_u: 6.43478679657\n", 913 | "\ttrain_h1_range_x_mean_u: 2.06610751152\n", 914 | "\ttrain_h1_range_x_min_u: 0.0\n", 915 | "\ttrain_h1_row_norms_max: 1.21548783779\n", 916 | "\ttrain_h1_row_norms_mean: 1.03275668621\n", 917 | "\ttrain_h1_row_norms_min: 0.889980256557\n", 918 | "\ttrain_objective: 0.727973520756\n", 919 | "\ttrain_y_col_norms_max: 1.00634145737\n", 920 | "\ttrain_y_col_norms_mean: 0.751444160938\n", 921 | "\ttrain_y_col_norms_min: 0.522492587566\n", 922 | "\ttrain_y_max_max_class: 0.978774309158\n", 923 | "\ttrain_y_mean_max_class: 0.610808432102\n", 924 | "\ttrain_y_min_max_class: 0.19889087975\n", 925 | "\ttrain_y_misclass: 0.197999984026\n", 926 | "\ttrain_y_nll: 0.727973520756\n", 927 | "\ttrain_y_row_norms_max: 0.486268281937\n", 928 | "\ttrain_y_row_norms_mean: 0.195263013244\n", 929 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 930 | "\ttraining_seconds_this_epoch: 0.0763399973512\n", 931 | "Saving to ./model_mnist.pkl...\n", 932 | "Saving to ./model_mnist.pkl done. Time elapsed: 0.280740 seconds\n", 933 | "Time this epoch: 0.076995 seconds\n", 934 | "Monitoring step:\n", 935 | "\tEpochs seen: 6\n", 936 | "\tBatches seen: 60\n", 937 | "\tExamples seen: 6000\n", 938 | "\tlearning_rate: 0.099999986589\n", 939 | "\tmomentum: 0.500000059605\n", 940 | "\ttest_h0_col_norms_max: 1.16920340061\n", 941 | "\ttest_h0_col_norms_mean: 1.03661429882\n", 942 | "\ttest_h0_col_norms_min: 0.950977563858\n", 943 | "\ttest_h0_max_x_max_u: 5.02766466141\n", 944 | "\ttest_h0_max_x_mean_u: 1.83139204979\n", 945 | "\ttest_h0_max_x_min_u: 0.0397325791419\n", 946 | "\ttest_h0_mean_x_max_u: 2.27318048477\n", 947 | "\ttest_h0_mean_x_mean_u: 0.623455286026\n", 948 | "\ttest_h0_mean_x_min_u: 0.000664046092425\n", 949 | "\ttest_h0_min_x_max_u: 0.62577098608\n", 950 | "\ttest_h0_min_x_mean_u: 0.0116202309728\n", 951 | "\ttest_h0_min_x_min_u: 0.0\n", 952 | "\ttest_h0_range_x_max_u: 5.01301288605\n", 953 | "\ttest_h0_range_x_mean_u: 1.81977188587\n", 954 | "\ttest_h0_range_x_min_u: 0.0397325791419\n", 955 | "\ttest_h0_row_norms_max: 0.472008615732\n", 956 | "\ttest_h0_row_norms_mean: 0.369161456823\n", 957 | "\ttest_h0_row_norms_min: 0.280473798513\n", 958 | "\ttest_h1_col_norms_max: 1.26524698734\n", 959 | "\ttest_h1_col_norms_mean: 1.03916871548\n", 960 | "\ttest_h1_col_norms_min: 0.853746473789\n", 961 | "\ttest_h1_max_x_max_u: 7.66941070557\n", 962 | "\ttest_h1_max_x_mean_u: 2.3909034729\n", 963 | "\ttest_h1_max_x_min_u: 0.0\n", 964 | "\ttest_h1_mean_x_max_u: 2.43283319473\n", 965 | "\ttest_h1_mean_x_mean_u: 0.77819788456\n", 966 | "\ttest_h1_mean_x_min_u: 0.0\n", 967 | "\ttest_h1_min_x_max_u: 0.0850924924016\n", 968 | "\ttest_h1_min_x_mean_u: 0.00154214771464\n", 969 | "\ttest_h1_min_x_min_u: 0.0\n", 970 | "\ttest_h1_range_x_max_u: 7.66941070557\n", 971 | "\ttest_h1_range_x_mean_u: 2.38936161995\n", 972 | "\ttest_h1_range_x_min_u: 0.0\n", 973 | "\ttest_h1_row_norms_max: 1.22704899311\n", 974 | "\ttest_h1_row_norms_mean: 1.03975486755\n", 975 | "\ttest_h1_row_norms_min: 0.890263557434\n", 976 | "\ttest_objective: 0.714465558529\n", 977 | "\ttest_y_col_norms_max: 1.12016665936\n", 978 | "\ttest_y_col_norms_mean: 0.851332485676\n", 979 | "\ttest_y_col_norms_min: 0.62745475769\n", 980 | "\ttest_y_max_max_class: 0.991564691067\n", 981 | "\ttest_y_mean_max_class: 0.673109650612\n", 982 | "\ttest_y_min_max_class: 0.247116580606\n", 983 | "\ttest_y_misclass: 0.203999981284\n", 984 | "\ttest_y_nll: 0.714465558529\n", 985 | "\ttest_y_row_norms_max: 0.559674739838\n", 986 | "\ttest_y_row_norms_mean: 0.22137606144\n", 987 | "\ttest_y_row_norms_min: 0.00266337138601\n", 988 | "\ttotal_seconds_last_epoch: 0.674911916256\n", 989 | "\ttrain_h0_col_norms_max: 1.16920340061\n", 990 | "\ttrain_h0_col_norms_mean: 1.03661429882\n", 991 | "\ttrain_h0_col_norms_min: 0.950977563858\n", 992 | "\ttrain_h0_max_x_max_u: 4.77302360535\n", 993 | "\ttrain_h0_max_x_mean_u: 1.79474043846\n", 994 | "\ttrain_h0_max_x_min_u: 0.0351286865771\n", 995 | "\ttrain_h0_mean_x_max_u: 2.22935128212\n", 996 | "\ttrain_h0_mean_x_mean_u: 0.608126759529\n", 997 | "\ttrain_h0_mean_x_min_u: 0.000505105941556\n", 998 | "\ttrain_h0_min_x_max_u: 0.64672935009\n", 999 | "\ttrain_h0_min_x_mean_u: 0.0102401375771\n", 1000 | "\ttrain_h0_min_x_min_u: 0.0\n", 1001 | "\ttrain_h0_range_x_max_u: 4.76895236969\n", 1002 | "\ttrain_h0_range_x_mean_u: 1.78450036049\n", 1003 | "\ttrain_h0_range_x_min_u: 0.0351286865771\n", 1004 | "\ttrain_h0_row_norms_max: 0.472008615732\n", 1005 | "\ttrain_h0_row_norms_mean: 0.369161456823\n", 1006 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 1007 | "\ttrain_h1_col_norms_max: 1.26524698734\n", 1008 | "\ttrain_h1_col_norms_mean: 1.03916871548\n", 1009 | "\ttrain_h1_col_norms_min: 0.853746473789\n", 1010 | "\ttrain_h1_max_x_max_u: 7.28800201416\n", 1011 | "\ttrain_h1_max_x_mean_u: 2.3407292366\n", 1012 | "\ttrain_h1_max_x_min_u: 0.0\n", 1013 | "\ttrain_h1_mean_x_max_u: 2.40500187874\n", 1014 | "\ttrain_h1_mean_x_mean_u: 0.763260304928\n", 1015 | "\ttrain_h1_mean_x_min_u: 0.0\n", 1016 | "\ttrain_h1_min_x_max_u: 0.0953525155783\n", 1017 | "\ttrain_h1_min_x_mean_u: 0.00115710240789\n", 1018 | "\ttrain_h1_min_x_min_u: 0.0\n", 1019 | "\ttrain_h1_range_x_max_u: 7.28651809692\n", 1020 | "\ttrain_h1_range_x_mean_u: 2.33957195282\n", 1021 | "\ttrain_h1_range_x_min_u: 0.0\n", 1022 | "\ttrain_h1_row_norms_max: 1.22704899311\n", 1023 | "\ttrain_h1_row_norms_mean: 1.03975486755\n", 1024 | "\ttrain_h1_row_norms_min: 0.890263557434\n", 1025 | "\ttrain_objective: 0.577620565891\n", 1026 | "\ttrain_y_col_norms_max: 1.12016665936\n", 1027 | "\ttrain_y_col_norms_mean: 0.851332485676\n", 1028 | "\ttrain_y_col_norms_min: 0.62745475769\n", 1029 | "\ttrain_y_max_max_class: 0.989890098572\n", 1030 | "\ttrain_y_mean_max_class: 0.705089151859\n", 1031 | "\ttrain_y_min_max_class: 0.240849375725\n", 1032 | "\ttrain_y_misclass: 0.157000005245\n", 1033 | "\ttrain_y_nll: 0.577620565891\n", 1034 | "\ttrain_y_row_norms_max: 0.559674739838\n", 1035 | "\ttrain_y_row_norms_mean: 0.22137606144\n", 1036 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 1037 | "\ttraining_seconds_this_epoch: 0.0769950002432\n", 1038 | "Time this epoch: 0.077085 seconds\n", 1039 | "Monitoring step:\n", 1040 | "\tEpochs seen: 7\n", 1041 | "\tBatches seen: 70\n", 1042 | "\tExamples seen: 7000\n", 1043 | "\tlearning_rate: 0.099999986589\n", 1044 | "\tmomentum: 0.500000059605\n", 1045 | "\ttest_h0_col_norms_max: 1.17077636719\n", 1046 | "\ttest_h0_col_norms_mean: 1.042345047\n", 1047 | "\ttest_h0_col_norms_min: 0.954309046268\n", 1048 | "\ttest_h0_max_x_max_u: 4.94107437134\n", 1049 | "\ttest_h0_max_x_mean_u: 1.90625333786\n", 1050 | "\ttest_h0_max_x_min_u: 0.0462045930326\n", 1051 | "\ttest_h0_mean_x_max_u: 2.15502882004\n", 1052 | "\ttest_h0_mean_x_mean_u: 0.647934436798\n", 1053 | "\ttest_h0_mean_x_min_u: 0.00081706547644\n", 1054 | "\ttest_h0_min_x_max_u: 0.615577101707\n", 1055 | "\ttest_h0_min_x_mean_u: 0.0103767225519\n", 1056 | "\ttest_h0_min_x_min_u: 0.0\n", 1057 | "\ttest_h0_range_x_max_u: 4.93711090088\n", 1058 | "\ttest_h0_range_x_mean_u: 1.89587628841\n", 1059 | "\ttest_h0_range_x_min_u: 0.0462045930326\n", 1060 | "\ttest_h0_row_norms_max: 0.47823575139\n", 1061 | "\ttest_h0_row_norms_mean: 0.371127814054\n", 1062 | "\ttest_h0_row_norms_min: 0.280473798513\n", 1063 | "\ttest_h1_col_norms_max: 1.27824509144\n", 1064 | "\ttest_h1_col_norms_mean: 1.04432523251\n", 1065 | "\ttest_h1_col_norms_min: 0.861365139484\n", 1066 | "\ttest_h1_max_x_max_u: 7.76571750641\n", 1067 | "\ttest_h1_max_x_mean_u: 2.5423271656\n", 1068 | "\ttest_h1_max_x_min_u: 0.0\n", 1069 | "\ttest_h1_mean_x_max_u: 2.43446969986\n", 1070 | "\ttest_h1_mean_x_mean_u: 0.814803242683\n", 1071 | "\ttest_h1_mean_x_min_u: 0.0\n", 1072 | "\ttest_h1_min_x_max_u: 0.05322836712\n", 1073 | "\ttest_h1_min_x_mean_u: 0.000682621437591\n", 1074 | "\ttest_h1_min_x_min_u: 0.0\n", 1075 | "\ttest_h1_range_x_max_u: 7.76571750641\n", 1076 | "\ttest_h1_range_x_mean_u: 2.54164457321\n", 1077 | "\ttest_h1_range_x_min_u: 0.0\n", 1078 | "\ttest_h1_row_norms_max: 1.23856174946\n", 1079 | "\ttest_h1_row_norms_mean: 1.04502367973\n", 1080 | "\ttest_h1_row_norms_min: 0.890461206436\n", 1081 | "\ttest_objective: 0.613446354866\n", 1082 | "\ttest_y_col_norms_max: 1.17661118507\n", 1083 | "\ttest_y_col_norms_mean: 0.922674536705\n", 1084 | "\ttest_y_col_norms_min: 0.691035449505\n", 1085 | "\ttest_y_max_max_class: 0.997032165527\n", 1086 | "\ttest_y_mean_max_class: 0.724932789803\n", 1087 | "\ttest_y_min_max_class: 0.256317406893\n", 1088 | "\ttest_y_misclass: 0.18599998951\n", 1089 | "\ttest_y_nll: 0.613446354866\n", 1090 | "\ttest_y_row_norms_max: 0.606767475605\n", 1091 | "\ttest_y_row_norms_mean: 0.240113765001\n", 1092 | "\ttest_y_row_norms_min: 0.00266337138601\n", 1093 | "\ttotal_seconds_last_epoch: 0.390940010548\n", 1094 | "\ttrain_h0_col_norms_max: 1.17077636719\n", 1095 | "\ttrain_h0_col_norms_mean: 1.042345047\n", 1096 | "\ttrain_h0_col_norms_min: 0.954309046268\n", 1097 | "\ttrain_h0_max_x_max_u: 4.79293251038\n", 1098 | "\ttrain_h0_max_x_mean_u: 1.86645793915\n", 1099 | "\ttrain_h0_max_x_min_u: 0.0428738966584\n", 1100 | "\ttrain_h0_mean_x_max_u: 2.10996413231\n", 1101 | "\ttrain_h0_mean_x_mean_u: 0.631803154945\n", 1102 | "\ttrain_h0_mean_x_min_u: 0.000632651848719\n", 1103 | "\ttrain_h0_min_x_max_u: 0.635077476501\n", 1104 | "\ttrain_h0_min_x_mean_u: 0.00887346547097\n", 1105 | "\ttrain_h0_min_x_min_u: 0.0\n", 1106 | "\ttrain_h0_range_x_max_u: 4.78711128235\n", 1107 | "\ttrain_h0_range_x_mean_u: 1.85758447647\n", 1108 | "\ttrain_h0_range_x_min_u: 0.0428738966584\n", 1109 | "\ttrain_h0_row_norms_max: 0.47823575139\n", 1110 | "\ttrain_h0_row_norms_mean: 0.371127814054\n", 1111 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 1112 | "\ttrain_h1_col_norms_max: 1.27824509144\n", 1113 | "\ttrain_h1_col_norms_mean: 1.04432523251\n", 1114 | "\ttrain_h1_col_norms_min: 0.861365139484\n", 1115 | "\ttrain_h1_max_x_max_u: 7.52039861679\n", 1116 | "\ttrain_h1_max_x_mean_u: 2.48978161812\n", 1117 | "\ttrain_h1_max_x_min_u: 0.0\n", 1118 | "\ttrain_h1_mean_x_max_u: 2.4125893116\n", 1119 | "\ttrain_h1_mean_x_mean_u: 0.799419045448\n", 1120 | "\ttrain_h1_mean_x_min_u: 0.0\n", 1121 | "\ttrain_h1_min_x_max_u: 0.0693615972996\n", 1122 | "\ttrain_h1_min_x_mean_u: 0.000934896816034\n", 1123 | "\ttrain_h1_min_x_min_u: 0.0\n", 1124 | "\ttrain_h1_range_x_max_u: 7.52039861679\n", 1125 | "\ttrain_h1_range_x_mean_u: 2.48884677887\n", 1126 | "\ttrain_h1_range_x_min_u: 0.0\n", 1127 | "\ttrain_h1_row_norms_max: 1.23856174946\n", 1128 | "\ttrain_h1_row_norms_mean: 1.04502367973\n", 1129 | "\ttrain_h1_row_norms_min: 0.890461206436\n", 1130 | "\ttrain_objective: 0.468716025352\n", 1131 | "\ttrain_y_col_norms_max: 1.17661118507\n", 1132 | "\ttrain_y_col_norms_mean: 0.922674536705\n", 1133 | "\ttrain_y_col_norms_min: 0.691035449505\n", 1134 | "\ttrain_y_max_max_class: 0.996845006943\n", 1135 | "\ttrain_y_mean_max_class: 0.761720895767\n", 1136 | "\ttrain_y_min_max_class: 0.252530395985\n", 1137 | "\ttrain_y_misclass: 0.130999997258\n", 1138 | "\ttrain_y_nll: 0.468716025352\n", 1139 | "\ttrain_y_row_norms_max: 0.606767475605\n", 1140 | "\ttrain_y_row_norms_mean: 0.240113765001\n", 1141 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 1142 | "\ttraining_seconds_this_epoch: 0.0770850032568\n", 1143 | "Time this epoch: 0.076852 seconds\n", 1144 | "Monitoring step:\n", 1145 | "\tEpochs seen: 8\n", 1146 | "\tBatches seen: 80\n", 1147 | "\tExamples seen: 8000\n", 1148 | "\tlearning_rate: 0.099999986589\n", 1149 | "\tmomentum: 0.500000059605\n", 1150 | "\ttest_h0_col_norms_max: 1.1883585453\n", 1151 | "\ttest_h0_col_norms_mean: 1.04716253281\n", 1152 | "\ttest_h0_col_norms_min: 0.95645314455\n", 1153 | "\ttest_h0_max_x_max_u: 5.18693971634\n", 1154 | "\ttest_h0_max_x_mean_u: 1.95527303219\n", 1155 | "\ttest_h0_max_x_min_u: 0.0574900060892\n", 1156 | "\ttest_h0_mean_x_max_u: 2.30134701729\n", 1157 | "\ttest_h0_mean_x_mean_u: 0.663498818874\n", 1158 | "\ttest_h0_mean_x_min_u: 0.00122210860718\n", 1159 | "\ttest_h0_min_x_max_u: 0.60200715065\n", 1160 | "\ttest_h0_min_x_mean_u: 0.0111785950139\n", 1161 | "\ttest_h0_min_x_min_u: 0.0\n", 1162 | "\ttest_h0_range_x_max_u: 5.17663192749\n", 1163 | "\ttest_h0_range_x_mean_u: 1.94409441948\n", 1164 | "\ttest_h0_range_x_min_u: 0.0574900060892\n", 1165 | "\ttest_h0_row_norms_max: 0.488236308098\n", 1166 | "\ttest_h0_row_norms_mean: 0.372778385878\n", 1167 | "\ttest_h0_row_norms_min: 0.280473798513\n", 1168 | "\ttest_h1_col_norms_max: 1.28770697117\n", 1169 | "\ttest_h1_col_norms_mean: 1.04868757725\n", 1170 | "\ttest_h1_col_norms_min: 0.866229593754\n", 1171 | "\ttest_h1_max_x_max_u: 8.15259933472\n", 1172 | "\ttest_h1_max_x_mean_u: 2.66557407379\n", 1173 | "\ttest_h1_max_x_min_u: 0.0\n", 1174 | "\ttest_h1_mean_x_max_u: 2.57336997986\n", 1175 | "\ttest_h1_mean_x_mean_u: 0.853820800781\n", 1176 | "\ttest_h1_mean_x_min_u: 0.0\n", 1177 | "\ttest_h1_min_x_max_u: 0.0631824657321\n", 1178 | "\ttest_h1_min_x_mean_u: 0.000758003152441\n", 1179 | "\ttest_h1_min_x_min_u: 0.0\n", 1180 | "\ttest_h1_range_x_max_u: 8.15259933472\n", 1181 | "\ttest_h1_range_x_mean_u: 2.66481590271\n", 1182 | "\ttest_h1_range_x_min_u: 0.0\n", 1183 | "\ttest_h1_row_norms_max: 1.24253559113\n", 1184 | "\ttest_h1_row_norms_mean: 1.04946482182\n", 1185 | "\ttest_h1_row_norms_min: 0.890857934952\n", 1186 | "\ttest_objective: 0.549885928631\n", 1187 | "\ttest_y_col_norms_max: 1.22743856907\n", 1188 | "\ttest_y_col_norms_mean: 0.978164553642\n", 1189 | "\ttest_y_col_norms_min: 0.755949676037\n", 1190 | "\ttest_y_max_max_class: 0.997386872768\n", 1191 | "\ttest_y_mean_max_class: 0.763035237789\n", 1192 | "\ttest_y_min_max_class: 0.26855763793\n", 1193 | "\ttest_y_misclass: 0.154999986291\n", 1194 | "\ttest_y_nll: 0.549885928631\n", 1195 | "\ttest_y_row_norms_max: 0.64081287384\n", 1196 | "\ttest_y_row_norms_mean: 0.254772484303\n", 1197 | "\ttest_y_row_norms_min: 0.00266337138601\n", 1198 | "\ttotal_seconds_last_epoch: 0.392641991377\n", 1199 | "\ttrain_h0_col_norms_max: 1.1883585453\n", 1200 | "\ttrain_h0_col_norms_mean: 1.04716253281\n", 1201 | "\ttrain_h0_col_norms_min: 0.95645314455\n", 1202 | "\ttrain_h0_max_x_max_u: 5.00512981415\n", 1203 | "\ttrain_h0_max_x_mean_u: 1.91368103027\n", 1204 | "\ttrain_h0_max_x_min_u: 0.0560524612665\n", 1205 | "\ttrain_h0_mean_x_max_u: 2.24890422821\n", 1206 | "\ttrain_h0_mean_x_mean_u: 0.646934449673\n", 1207 | "\ttrain_h0_mean_x_min_u: 0.000873918936122\n", 1208 | "\ttrain_h0_min_x_max_u: 0.620710134506\n", 1209 | "\ttrain_h0_min_x_mean_u: 0.00883941538632\n", 1210 | "\ttrain_h0_min_x_min_u: 0.0\n", 1211 | "\ttrain_h0_range_x_max_u: 4.9994802475\n", 1212 | "\ttrain_h0_range_x_mean_u: 1.90484189987\n", 1213 | "\ttrain_h0_range_x_min_u: 0.0560524612665\n", 1214 | "\ttrain_h0_row_norms_max: 0.488236308098\n", 1215 | "\ttrain_h0_row_norms_mean: 0.372778385878\n", 1216 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 1217 | "\ttrain_h1_col_norms_max: 1.28770697117\n", 1218 | "\ttrain_h1_col_norms_mean: 1.04868757725\n", 1219 | "\ttrain_h1_col_norms_min: 0.866229593754\n", 1220 | "\ttrain_h1_max_x_max_u: 7.88111543655\n", 1221 | "\ttrain_h1_max_x_mean_u: 2.61039280891\n", 1222 | "\ttrain_h1_max_x_min_u: 0.0\n", 1223 | "\ttrain_h1_mean_x_max_u: 2.55444598198\n", 1224 | "\ttrain_h1_mean_x_mean_u: 0.837123394012\n", 1225 | "\ttrain_h1_mean_x_min_u: 0.0\n", 1226 | "\ttrain_h1_min_x_max_u: 0.06248851493\n", 1227 | "\ttrain_h1_min_x_mean_u: 0.000725128513295\n", 1228 | "\ttrain_h1_min_x_min_u: 0.0\n", 1229 | "\ttrain_h1_range_x_max_u: 7.88111543655\n", 1230 | "\ttrain_h1_range_x_mean_u: 2.60966730118\n", 1231 | "\ttrain_h1_range_x_min_u: 0.0\n", 1232 | "\ttrain_h1_row_norms_max: 1.24253559113\n", 1233 | "\ttrain_h1_row_norms_mean: 1.04946482182\n", 1234 | "\ttrain_h1_row_norms_min: 0.890857934952\n", 1235 | "\ttrain_objective: 0.382768630981\n", 1236 | "\ttrain_y_col_norms_max: 1.22743856907\n", 1237 | "\ttrain_y_col_norms_mean: 0.978164553642\n", 1238 | "\ttrain_y_col_norms_min: 0.755949676037\n", 1239 | "\ttrain_y_max_max_class: 0.996851742268\n", 1240 | "\ttrain_y_mean_max_class: 0.800312995911\n", 1241 | "\ttrain_y_min_max_class: 0.287817567587\n", 1242 | "\ttrain_y_misclass: 0.10000000149\n", 1243 | "\ttrain_y_nll: 0.382768630981\n", 1244 | "\ttrain_y_row_norms_max: 0.64081287384\n", 1245 | "\ttrain_y_row_norms_mean: 0.254772484303\n", 1246 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 1247 | "\ttraining_seconds_this_epoch: 0.0768520012498\n", 1248 | "Time this epoch: 0.076838 seconds\n", 1249 | "Monitoring step:\n", 1250 | "\tEpochs seen: 9\n", 1251 | "\tBatches seen: 90\n", 1252 | "\tExamples seen: 9000\n", 1253 | "\tlearning_rate: 0.099999986589\n", 1254 | "\tmomentum: 0.500000059605\n", 1255 | "\ttest_h0_col_norms_max: 1.21042382717\n", 1256 | "\ttest_h0_col_norms_mean: 1.05139183998\n", 1257 | "\ttest_h0_col_norms_min: 0.959612369537\n", 1258 | "\ttest_h0_max_x_max_u: 5.23728370667\n", 1259 | "\ttest_h0_max_x_mean_u: 1.98849892616\n", 1260 | "\ttest_h0_max_x_min_u: 0.0686023682356\n", 1261 | "\ttest_h0_mean_x_max_u: 2.33452033997\n", 1262 | "\ttest_h0_mean_x_mean_u: 0.673880696297\n", 1263 | "\ttest_h0_mean_x_min_u: 0.00141649576835\n", 1264 | "\ttest_h0_min_x_max_u: 0.598745048046\n", 1265 | "\ttest_h0_min_x_mean_u: 0.0114188473672\n", 1266 | "\ttest_h0_min_x_min_u: 0.0\n", 1267 | "\ttest_h0_range_x_max_u: 5.22485589981\n", 1268 | "\ttest_h0_range_x_mean_u: 1.97707986832\n", 1269 | "\ttest_h0_range_x_min_u: 0.0686023682356\n", 1270 | "\ttest_h0_row_norms_max: 0.495743811131\n", 1271 | "\ttest_h0_row_norms_mean: 0.374224215746\n", 1272 | "\ttest_h0_row_norms_min: 0.280473798513\n", 1273 | "\ttest_h1_col_norms_max: 1.29437077045\n", 1274 | "\ttest_h1_col_norms_mean: 1.05243492126\n", 1275 | "\ttest_h1_col_norms_min: 0.870068728924\n", 1276 | "\ttest_h1_max_x_max_u: 8.14818382263\n", 1277 | "\ttest_h1_max_x_mean_u: 2.75054955482\n", 1278 | "\ttest_h1_max_x_min_u: 0.0\n", 1279 | "\ttest_h1_mean_x_max_u: 2.54476070404\n", 1280 | "\ttest_h1_mean_x_mean_u: 0.879209578037\n", 1281 | "\ttest_h1_mean_x_min_u: 0.0\n", 1282 | "\ttest_h1_min_x_max_u: 0.0774512737989\n", 1283 | "\ttest_h1_min_x_mean_u: 0.00121434987523\n", 1284 | "\ttest_h1_min_x_min_u: 0.0\n", 1285 | "\ttest_h1_range_x_max_u: 8.14818382263\n", 1286 | "\ttest_h1_range_x_mean_u: 2.749335289\n", 1287 | "\ttest_h1_range_x_min_u: 0.0\n", 1288 | "\ttest_h1_row_norms_max: 1.24782478809\n", 1289 | "\ttest_h1_row_norms_mean: 1.05328190327\n", 1290 | "\ttest_h1_row_norms_min: 0.891310214996\n", 1291 | "\ttest_objective: 0.562562406063\n", 1292 | "\ttest_y_col_norms_max: 1.26394033432\n", 1293 | "\ttest_y_col_norms_mean: 1.02366733551\n", 1294 | "\ttest_y_col_norms_min: 0.801639020443\n", 1295 | "\ttest_y_max_max_class: 0.996918797493\n", 1296 | "\ttest_y_mean_max_class: 0.782383561134\n", 1297 | "\ttest_y_min_max_class: 0.272926717997\n", 1298 | "\ttest_y_misclass: 0.159999996424\n", 1299 | "\ttest_y_nll: 0.562562406063\n", 1300 | "\ttest_y_row_norms_max: 0.66804343462\n", 1301 | "\ttest_y_row_norms_mean: 0.266814947128\n", 1302 | "\ttest_y_row_norms_min: 0.00266337138601\n", 1303 | "\ttotal_seconds_last_epoch: 0.391522973776\n", 1304 | "\ttrain_h0_col_norms_max: 1.21042382717\n", 1305 | "\ttrain_h0_col_norms_mean: 1.05139183998\n", 1306 | "\ttrain_h0_col_norms_min: 0.959612369537\n", 1307 | "\ttrain_h0_max_x_max_u: 5.04359769821\n", 1308 | "\ttrain_h0_max_x_mean_u: 1.94545042515\n", 1309 | "\ttrain_h0_max_x_min_u: 0.0644600838423\n", 1310 | "\ttrain_h0_mean_x_max_u: 2.27801418304\n", 1311 | "\ttrain_h0_mean_x_mean_u: 0.656861424446\n", 1312 | "\ttrain_h0_mean_x_min_u: 0.0010711546056\n", 1313 | "\ttrain_h0_min_x_max_u: 0.619264960289\n", 1314 | "\ttrain_h0_min_x_mean_u: 0.00865776557475\n", 1315 | "\ttrain_h0_min_x_min_u: 0.0\n", 1316 | "\ttrain_h0_range_x_max_u: 5.03672027588\n", 1317 | "\ttrain_h0_range_x_mean_u: 1.93679261208\n", 1318 | "\ttrain_h0_range_x_min_u: 0.0644600838423\n", 1319 | "\ttrain_h0_row_norms_max: 0.495743811131\n", 1320 | "\ttrain_h0_row_norms_mean: 0.374224215746\n", 1321 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 1322 | "\ttrain_h1_col_norms_max: 1.29437077045\n", 1323 | "\ttrain_h1_col_norms_mean: 1.05243492126\n", 1324 | "\ttrain_h1_col_norms_min: 0.870068728924\n", 1325 | "\ttrain_h1_max_x_max_u: 7.94437932968\n", 1326 | "\ttrain_h1_max_x_mean_u: 2.69649624825\n", 1327 | "\ttrain_h1_max_x_min_u: 0.0\n", 1328 | "\ttrain_h1_mean_x_max_u: 2.53754758835\n", 1329 | "\ttrain_h1_mean_x_mean_u: 0.861332058907\n", 1330 | "\ttrain_h1_mean_x_min_u: 0.0\n", 1331 | "\ttrain_h1_min_x_max_u: 0.0527574419975\n", 1332 | "\ttrain_h1_min_x_mean_u: 0.000555102771614\n", 1333 | "\ttrain_h1_min_x_min_u: 0.0\n", 1334 | "\ttrain_h1_range_x_max_u: 7.94437932968\n", 1335 | "\ttrain_h1_range_x_mean_u: 2.69594097137\n", 1336 | "\ttrain_h1_range_x_min_u: 0.0\n", 1337 | "\ttrain_h1_row_norms_max: 1.24782478809\n", 1338 | "\ttrain_h1_row_norms_mean: 1.05328190327\n", 1339 | "\ttrain_h1_row_norms_min: 0.891310214996\n", 1340 | "\ttrain_objective: 0.351353615522\n", 1341 | "\ttrain_y_col_norms_max: 1.26394033432\n", 1342 | "\ttrain_y_col_norms_mean: 1.02366733551\n", 1343 | "\ttrain_y_col_norms_min: 0.801639020443\n", 1344 | "\ttrain_y_max_max_class: 0.998119235039\n", 1345 | "\ttrain_y_mean_max_class: 0.819598913193\n", 1346 | "\ttrain_y_min_max_class: 0.311893731356\n", 1347 | "\ttrain_y_misclass: 0.0949999913573\n", 1348 | "\ttrain_y_nll: 0.351353615522\n", 1349 | "\ttrain_y_row_norms_max: 0.66804343462\n", 1350 | "\ttrain_y_row_norms_mean: 0.266814947128\n", 1351 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 1352 | "\ttraining_seconds_this_epoch: 0.0768379941583\n", 1353 | "Time this epoch: 0.076683 seconds\n", 1354 | "Monitoring step:\n", 1355 | "\tEpochs seen: 10\n", 1356 | "\tBatches seen: 100\n", 1357 | "\tExamples seen: 10000\n", 1358 | "\tlearning_rate: 0.099999986589\n", 1359 | "\tmomentum: 0.500000059605\n", 1360 | "\ttest_h0_col_norms_max: 1.23557794094\n", 1361 | "\ttest_h0_col_norms_mean: 1.05545055866\n", 1362 | "\ttest_h0_col_norms_min: 0.962563753128\n", 1363 | "\ttest_h0_max_x_max_u: 5.10529375076\n", 1364 | "\ttest_h0_max_x_mean_u: 2.01641106606\n", 1365 | "\ttest_h0_max_x_min_u: 0.0810308307409\n", 1366 | "\ttest_h0_mean_x_max_u: 2.22507834435\n", 1367 | "\ttest_h0_mean_x_mean_u: 0.680157065392\n", 1368 | "\ttest_h0_mean_x_min_u: 0.00166573619936\n", 1369 | "\ttest_h0_min_x_max_u: 0.598347783089\n", 1370 | "\ttest_h0_min_x_mean_u: 0.0095747243613\n", 1371 | "\ttest_h0_min_x_min_u: 0.0\n", 1372 | "\ttest_h0_range_x_max_u: 5.09927988052\n", 1373 | "\ttest_h0_range_x_mean_u: 2.00683641434\n", 1374 | "\ttest_h0_range_x_min_u: 0.0810308307409\n", 1375 | "\ttest_h0_row_norms_max: 0.503167450428\n", 1376 | "\ttest_h0_row_norms_mean: 0.375609278679\n", 1377 | "\ttest_h0_row_norms_min: 0.280473798513\n", 1378 | "\ttest_h1_col_norms_max: 1.29984366894\n", 1379 | "\ttest_h1_col_norms_mean: 1.05600309372\n", 1380 | "\ttest_h1_col_norms_min: 0.874703466892\n", 1381 | "\ttest_h1_max_x_max_u: 8.38683700562\n", 1382 | "\ttest_h1_max_x_mean_u: 2.80211305618\n", 1383 | "\ttest_h1_max_x_min_u: 0.0\n", 1384 | "\ttest_h1_mean_x_max_u: 2.66854643822\n", 1385 | "\ttest_h1_mean_x_mean_u: 0.881503105164\n", 1386 | "\ttest_h1_mean_x_min_u: 0.0\n", 1387 | "\ttest_h1_min_x_max_u: 0.0255277194083\n", 1388 | "\ttest_h1_min_x_mean_u: 0.000376855314244\n", 1389 | "\ttest_h1_min_x_min_u: 0.0\n", 1390 | "\ttest_h1_range_x_max_u: 8.38683700562\n", 1391 | "\ttest_h1_range_x_mean_u: 2.80173611641\n", 1392 | "\ttest_h1_range_x_min_u: 0.0\n", 1393 | "\ttest_h1_row_norms_max: 1.25574398041\n", 1394 | "\ttest_h1_row_norms_mean: 1.0569344759\n", 1395 | "\ttest_h1_row_norms_min: 0.891548454762\n", 1396 | "\ttest_objective: 0.502181708813\n", 1397 | "\ttest_y_col_norms_max: 1.29418241978\n", 1398 | "\ttest_y_col_norms_mean: 1.06550705433\n", 1399 | "\ttest_y_col_norms_min: 0.834318339825\n", 1400 | "\ttest_y_max_max_class: 0.998865067959\n", 1401 | "\ttest_y_mean_max_class: 0.801096379757\n", 1402 | "\ttest_y_min_max_class: 0.282263696194\n", 1403 | "\ttest_y_misclass: 0.156000003219\n", 1404 | "\ttest_y_nll: 0.502181708813\n", 1405 | "\ttest_y_row_norms_max: 0.70500010252\n", 1406 | "\ttest_y_row_norms_mean: 0.277966141701\n", 1407 | "\ttest_y_row_norms_min: 0.00266337138601\n", 1408 | "\ttotal_seconds_last_epoch: 0.391250967979\n", 1409 | "\ttrain_h0_col_norms_max: 1.23557794094\n", 1410 | "\ttrain_h0_col_norms_mean: 1.05545055866\n", 1411 | "\ttrain_h0_col_norms_min: 0.962563753128\n", 1412 | "\ttrain_h0_max_x_max_u: 4.98928642273\n", 1413 | "\ttrain_h0_max_x_mean_u: 1.97139668465\n", 1414 | "\ttrain_h0_max_x_min_u: 0.0727752968669\n", 1415 | "\ttrain_h0_mean_x_max_u: 2.17031216621\n", 1416 | "\ttrain_h0_mean_x_mean_u: 0.662902116776\n", 1417 | "\ttrain_h0_mean_x_min_u: 0.00133575883228\n", 1418 | "\ttrain_h0_min_x_max_u: 0.613275527954\n", 1419 | "\ttrain_h0_min_x_mean_u: 0.00743194669485\n", 1420 | "\ttrain_h0_min_x_min_u: 0.0\n", 1421 | "\ttrain_h0_range_x_max_u: 4.98928642273\n", 1422 | "\ttrain_h0_range_x_mean_u: 1.96396493912\n", 1423 | "\ttrain_h0_range_x_min_u: 0.0727752968669\n", 1424 | "\ttrain_h0_row_norms_max: 0.503167450428\n", 1425 | "\ttrain_h0_row_norms_mean: 0.375609278679\n", 1426 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 1427 | "\ttrain_h1_col_norms_max: 1.29984366894\n", 1428 | "\ttrain_h1_col_norms_mean: 1.05600309372\n", 1429 | "\ttrain_h1_col_norms_min: 0.874703466892\n", 1430 | "\ttrain_h1_max_x_max_u: 8.23232650757\n", 1431 | "\ttrain_h1_max_x_mean_u: 2.74831795692\n", 1432 | "\ttrain_h1_max_x_min_u: 0.0\n", 1433 | "\ttrain_h1_mean_x_max_u: 2.60985064507\n", 1434 | "\ttrain_h1_mean_x_mean_u: 0.864189982414\n", 1435 | "\ttrain_h1_mean_x_min_u: 0.0\n", 1436 | "\ttrain_h1_min_x_max_u: 0.0363659411669\n", 1437 | "\ttrain_h1_min_x_mean_u: 0.000363659433788\n", 1438 | "\ttrain_h1_min_x_min_u: 0.0\n", 1439 | "\ttrain_h1_range_x_max_u: 8.23232650757\n", 1440 | "\ttrain_h1_range_x_mean_u: 2.74795436859\n", 1441 | "\ttrain_h1_range_x_min_u: 0.0\n", 1442 | "\ttrain_h1_row_norms_max: 1.25574398041\n", 1443 | "\ttrain_h1_row_norms_mean: 1.0569344759\n", 1444 | "\ttrain_h1_row_norms_min: 0.891548454762\n", 1445 | "\ttrain_objective: 0.301149189472\n", 1446 | "\ttrain_y_col_norms_max: 1.29418241978\n", 1447 | "\ttrain_y_col_norms_mean: 1.06550705433\n", 1448 | "\ttrain_y_col_norms_min: 0.834318339825\n", 1449 | "\ttrain_y_max_max_class: 0.998896121979\n", 1450 | "\ttrain_y_mean_max_class: 0.843701899052\n", 1451 | "\ttrain_y_min_max_class: 0.332697063684\n", 1452 | "\ttrain_y_misclass: 0.082999996841\n", 1453 | "\ttrain_y_nll: 0.301149189472\n", 1454 | "\ttrain_y_row_norms_max: 0.70500010252\n", 1455 | "\ttrain_y_row_norms_mean: 0.277966141701\n", 1456 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 1457 | "\ttraining_seconds_this_epoch: 0.0766830071807\n", 1458 | "Saving to ./model_mnist.pkl...\n", 1459 | "Saving to ./model_mnist.pkl done. Time elapsed: 0.079974 seconds\n", 1460 | "Time this epoch: 0.076912 seconds\n", 1461 | "Monitoring step:\n", 1462 | "\tEpochs seen: 11\n", 1463 | "\tBatches seen: 110\n", 1464 | "\tExamples seen: 11000\n", 1465 | "\tlearning_rate: 0.099999986589\n", 1466 | "\tmomentum: 0.500000059605\n", 1467 | "\ttest_h0_col_norms_max: 1.25335729122\n", 1468 | "\ttest_h0_col_norms_mean: 1.05905938148\n", 1469 | "\ttest_h0_col_norms_min: 0.964543700218\n", 1470 | "\ttest_h0_max_x_max_u: 5.2987985611\n", 1471 | "\ttest_h0_max_x_mean_u: 2.04572296143\n", 1472 | "\ttest_h0_max_x_min_u: 0.0866454318166\n", 1473 | "\ttest_h0_mean_x_max_u: 2.35758185387\n", 1474 | "\ttest_h0_mean_x_mean_u: 0.690596044064\n", 1475 | "\ttest_h0_mean_x_min_u: 0.00182079733349\n", 1476 | "\ttest_h0_min_x_max_u: 0.590359926224\n", 1477 | "\ttest_h0_min_x_mean_u: 0.0109091578051\n", 1478 | "\ttest_h0_min_x_min_u: 0.0\n", 1479 | "\ttest_h0_range_x_max_u: 5.28407907486\n", 1480 | "\ttest_h0_range_x_mean_u: 2.0348136425\n", 1481 | "\ttest_h0_range_x_min_u: 0.0866454318166\n", 1482 | "\ttest_h0_row_norms_max: 0.511548280716\n", 1483 | "\ttest_h0_row_norms_mean: 0.37684443593\n", 1484 | "\ttest_h0_row_norms_min: 0.280473798513\n", 1485 | "\ttest_h1_col_norms_max: 1.30880248547\n", 1486 | "\ttest_h1_col_norms_mean: 1.05919194221\n", 1487 | "\ttest_h1_col_norms_min: 0.879486143589\n", 1488 | "\ttest_h1_max_x_max_u: 8.47961425781\n", 1489 | "\ttest_h1_max_x_mean_u: 2.88864159584\n", 1490 | "\ttest_h1_max_x_min_u: 0.0\n", 1491 | "\ttest_h1_mean_x_max_u: 2.64655375481\n", 1492 | "\ttest_h1_mean_x_mean_u: 0.918162703514\n", 1493 | "\ttest_h1_mean_x_min_u: 0.0\n", 1494 | "\ttest_h1_min_x_max_u: 0.0780160352588\n", 1495 | "\ttest_h1_min_x_mean_u: 0.00123767822515\n", 1496 | "\ttest_h1_min_x_min_u: 0.0\n", 1497 | "\ttest_h1_range_x_max_u: 8.47961425781\n", 1498 | "\ttest_h1_range_x_mean_u: 2.887403965\n", 1499 | "\ttest_h1_range_x_min_u: 0.0\n", 1500 | "\ttest_h1_row_norms_max: 1.27046346664\n", 1501 | "\ttest_h1_row_norms_mean: 1.06017994881\n", 1502 | "\ttest_h1_row_norms_min: 0.89188170433\n", 1503 | "\ttest_objective: 0.508433103561\n", 1504 | "\ttest_y_col_norms_max: 1.32279527187\n", 1505 | "\ttest_y_col_norms_mean: 1.10169160366\n", 1506 | "\ttest_y_col_norms_min: 0.879823625088\n", 1507 | "\ttest_y_max_max_class: 0.999233484268\n", 1508 | "\ttest_y_mean_max_class: 0.820999860764\n", 1509 | "\ttest_y_min_max_class: 0.310878247023\n", 1510 | "\ttest_y_misclass: 0.144000008702\n", 1511 | "\ttest_y_nll: 0.508433103561\n", 1512 | "\ttest_y_row_norms_max: 0.716839492321\n", 1513 | "\ttest_y_row_norms_mean: 0.287569135427\n", 1514 | "\ttest_y_row_norms_min: 0.00266337138601\n", 1515 | "\ttotal_seconds_last_epoch: 0.498600929976\n", 1516 | "\ttrain_h0_col_norms_max: 1.25335729122\n", 1517 | "\ttrain_h0_col_norms_mean: 1.05905938148\n", 1518 | "\ttrain_h0_col_norms_min: 0.964543700218\n", 1519 | "\ttrain_h0_max_x_max_u: 5.14989566803\n", 1520 | "\ttrain_h0_max_x_mean_u: 1.99691367149\n", 1521 | "\ttrain_h0_max_x_min_u: 0.0791974589229\n", 1522 | "\ttrain_h0_mean_x_max_u: 2.29658269882\n", 1523 | "\ttrain_h0_mean_x_mean_u: 0.673060178757\n", 1524 | "\ttrain_h0_mean_x_min_u: 0.00153974723071\n", 1525 | "\ttrain_h0_min_x_max_u: 0.606921255589\n", 1526 | "\ttrain_h0_min_x_mean_u: 0.00873211864382\n", 1527 | "\ttrain_h0_min_x_min_u: 0.0\n", 1528 | "\ttrain_h0_range_x_max_u: 5.1348195076\n", 1529 | "\ttrain_h0_range_x_mean_u: 1.98818135262\n", 1530 | "\ttrain_h0_range_x_min_u: 0.0791974589229\n", 1531 | "\ttrain_h0_row_norms_max: 0.511548280716\n", 1532 | "\ttrain_h0_row_norms_mean: 0.37684443593\n", 1533 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 1534 | "\ttrain_h1_col_norms_max: 1.30880248547\n", 1535 | "\ttrain_h1_col_norms_mean: 1.05919194221\n", 1536 | "\ttrain_h1_col_norms_min: 0.879486143589\n", 1537 | "\ttrain_h1_max_x_max_u: 8.30659675598\n", 1538 | "\ttrain_h1_max_x_mean_u: 2.82923531532\n", 1539 | "\ttrain_h1_max_x_min_u: 0.0\n", 1540 | "\ttrain_h1_mean_x_max_u: 2.64465999603\n", 1541 | "\ttrain_h1_mean_x_mean_u: 0.899735271931\n", 1542 | "\ttrain_h1_mean_x_min_u: 0.0\n", 1543 | "\ttrain_h1_min_x_max_u: 0.0357568822801\n", 1544 | "\ttrain_h1_min_x_mean_u: 0.000409733387642\n", 1545 | "\ttrain_h1_min_x_min_u: 0.0\n", 1546 | "\ttrain_h1_range_x_max_u: 8.30659675598\n", 1547 | "\ttrain_h1_range_x_mean_u: 2.8288257122\n", 1548 | "\ttrain_h1_range_x_min_u: 0.0\n", 1549 | "\ttrain_h1_row_norms_max: 1.27046346664\n", 1550 | "\ttrain_h1_row_norms_mean: 1.06017994881\n", 1551 | "\ttrain_h1_row_norms_min: 0.89188170433\n", 1552 | "\ttrain_objective: 0.276619881392\n", 1553 | "\ttrain_y_col_norms_max: 1.32279527187\n", 1554 | "\ttrain_y_col_norms_mean: 1.10169160366\n", 1555 | "\ttrain_y_col_norms_min: 0.879823625088\n", 1556 | "\ttrain_y_max_max_class: 0.99935489893\n", 1557 | "\ttrain_y_mean_max_class: 0.856062114239\n", 1558 | "\ttrain_y_min_max_class: 0.321078747511\n", 1559 | "\ttrain_y_misclass: 0.0810000002384\n", 1560 | "\ttrain_y_nll: 0.276619881392\n", 1561 | "\ttrain_y_row_norms_max: 0.716839492321\n", 1562 | "\ttrain_y_row_norms_mean: 0.287569135427\n", 1563 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 1564 | "\ttraining_seconds_this_epoch: 0.0769120007753\n", 1565 | "Time this epoch: 0.076811 seconds\n", 1566 | "Monitoring step:\n", 1567 | "\tEpochs seen: 12\n", 1568 | "\tBatches seen: 120\n", 1569 | "\tExamples seen: 12000\n", 1570 | "\tlearning_rate: 0.099999986589\n", 1571 | "\tmomentum: 0.500000059605\n", 1572 | "\ttest_h0_col_norms_max: 1.27268207073\n", 1573 | "\ttest_h0_col_norms_mean: 1.06261181831\n", 1574 | "\ttest_h0_col_norms_min: 0.967070996761\n", 1575 | "\ttest_h0_max_x_max_u: 5.22094202042\n", 1576 | "\ttest_h0_max_x_mean_u: 2.05405211449\n", 1577 | "\ttest_h0_max_x_min_u: 0.0909740328789\n", 1578 | "\ttest_h0_mean_x_max_u: 2.30208063126\n", 1579 | "\ttest_h0_mean_x_mean_u: 0.687728643417\n", 1580 | "\ttest_h0_mean_x_min_u: 0.00190567364916\n", 1581 | "\ttest_h0_min_x_max_u: 0.563181042671\n", 1582 | "\ttest_h0_min_x_mean_u: 0.00934776104987\n", 1583 | "\ttest_h0_min_x_min_u: 0.0\n", 1584 | "\ttest_h0_range_x_max_u: 5.20952367783\n", 1585 | "\ttest_h0_range_x_mean_u: 2.04470443726\n", 1586 | "\ttest_h0_range_x_min_u: 0.0909740328789\n", 1587 | "\ttest_h0_row_norms_max: 0.520189642906\n", 1588 | "\ttest_h0_row_norms_mean: 0.378056198359\n", 1589 | "\ttest_h0_row_norms_min: 0.280473798513\n", 1590 | "\ttest_h1_col_norms_max: 1.31339359283\n", 1591 | "\ttest_h1_col_norms_mean: 1.06223833561\n", 1592 | "\ttest_h1_col_norms_min: 0.882938563824\n", 1593 | "\ttest_h1_max_x_max_u: 8.51983261108\n", 1594 | "\ttest_h1_max_x_mean_u: 2.90860939026\n", 1595 | "\ttest_h1_max_x_min_u: 0.0\n", 1596 | "\ttest_h1_mean_x_max_u: 2.66145586967\n", 1597 | "\ttest_h1_mean_x_mean_u: 0.909021735191\n", 1598 | "\ttest_h1_mean_x_min_u: 0.0\n", 1599 | "\ttest_h1_min_x_max_u: 0.0449122749269\n", 1600 | "\ttest_h1_min_x_mean_u: 0.000772155588493\n", 1601 | "\ttest_h1_min_x_min_u: 0.0\n", 1602 | "\ttest_h1_range_x_max_u: 8.51983261108\n", 1603 | "\ttest_h1_range_x_mean_u: 2.90783715248\n", 1604 | "\ttest_h1_range_x_min_u: 0.0\n", 1605 | "\ttest_h1_row_norms_max: 1.28586041927\n", 1606 | "\ttest_h1_row_norms_mean: 1.06328785419\n", 1607 | "\ttest_h1_row_norms_min: 0.892130732536\n", 1608 | "\ttest_objective: 0.495049893856\n", 1609 | "\ttest_y_col_norms_max: 1.34686243534\n", 1610 | "\ttest_y_col_norms_mean: 1.13473248482\n", 1611 | "\ttest_y_col_norms_min: 0.917872130871\n", 1612 | "\ttest_y_max_max_class: 0.99938750267\n", 1613 | "\ttest_y_mean_max_class: 0.836464464664\n", 1614 | "\ttest_y_min_max_class: 0.343189746141\n", 1615 | "\ttest_y_misclass: 0.142999991775\n", 1616 | "\ttest_y_nll: 0.495049893856\n", 1617 | "\ttest_y_row_norms_max: 0.733798205853\n", 1618 | "\ttest_y_row_norms_mean: 0.296515792608\n", 1619 | "\ttest_y_row_norms_min: 0.00266337138601\n", 1620 | "\ttotal_seconds_last_epoch: 0.392366021872\n", 1621 | "\ttrain_h0_col_norms_max: 1.27268207073\n", 1622 | "\ttrain_h0_col_norms_mean: 1.06261181831\n", 1623 | "\ttrain_h0_col_norms_min: 0.967070996761\n", 1624 | "\ttrain_h0_max_x_max_u: 5.10215473175\n", 1625 | "\ttrain_h0_max_x_mean_u: 2.00527834892\n", 1626 | "\ttrain_h0_max_x_min_u: 0.0821754857898\n", 1627 | "\ttrain_h0_mean_x_max_u: 2.24179244041\n", 1628 | "\ttrain_h0_mean_x_mean_u: 0.670412421227\n", 1629 | "\ttrain_h0_mean_x_min_u: 0.00172259681858\n", 1630 | "\ttrain_h0_min_x_max_u: 0.58440387249\n", 1631 | "\ttrain_h0_min_x_mean_u: 0.00785345956683\n", 1632 | "\ttrain_h0_min_x_min_u: 0.0\n", 1633 | "\ttrain_h0_range_x_max_u: 5.09986066818\n", 1634 | "\ttrain_h0_range_x_mean_u: 1.99742484093\n", 1635 | "\ttrain_h0_range_x_min_u: 0.0821754857898\n", 1636 | "\ttrain_h0_row_norms_max: 0.520189642906\n", 1637 | "\ttrain_h0_row_norms_mean: 0.378056198359\n", 1638 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 1639 | "\ttrain_h1_col_norms_max: 1.31339359283\n", 1640 | "\ttrain_h1_col_norms_mean: 1.06223833561\n", 1641 | "\ttrain_h1_col_norms_min: 0.882938563824\n", 1642 | "\ttrain_h1_max_x_max_u: 8.35809230804\n", 1643 | "\ttrain_h1_max_x_mean_u: 2.85300898552\n", 1644 | "\ttrain_h1_max_x_min_u: 0.0\n", 1645 | "\ttrain_h1_mean_x_max_u: 2.65430593491\n", 1646 | "\ttrain_h1_mean_x_mean_u: 0.891693353653\n", 1647 | "\ttrain_h1_mean_x_min_u: 0.0\n", 1648 | "\ttrain_h1_min_x_max_u: 0.0136786298826\n", 1649 | "\ttrain_h1_min_x_mean_u: 0.000136786286021\n", 1650 | "\ttrain_h1_min_x_min_u: 0.0\n", 1651 | "\ttrain_h1_range_x_max_u: 8.35809230804\n", 1652 | "\ttrain_h1_range_x_mean_u: 2.85287261009\n", 1653 | "\ttrain_h1_range_x_min_u: 0.0\n", 1654 | "\ttrain_h1_row_norms_max: 1.28586041927\n", 1655 | "\ttrain_h1_row_norms_mean: 1.06328785419\n", 1656 | "\ttrain_h1_row_norms_min: 0.892130732536\n", 1657 | "\ttrain_objective: 0.244498074055\n", 1658 | "\ttrain_y_col_norms_max: 1.34686243534\n", 1659 | "\ttrain_y_col_norms_mean: 1.13473248482\n", 1660 | "\ttrain_y_col_norms_min: 0.917872130871\n", 1661 | "\ttrain_y_max_max_class: 0.999567329884\n", 1662 | "\ttrain_y_mean_max_class: 0.871302306652\n", 1663 | "\ttrain_y_min_max_class: 0.316225677729\n", 1664 | "\ttrain_y_misclass: 0.0689999982715\n", 1665 | "\ttrain_y_nll: 0.244498074055\n", 1666 | "\ttrain_y_row_norms_max: 0.733798205853\n", 1667 | "\ttrain_y_row_norms_mean: 0.296515792608\n", 1668 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 1669 | "\ttraining_seconds_this_epoch: 0.0768110081553\n", 1670 | "Time this epoch: 0.076614 seconds\n", 1671 | "Monitoring step:\n", 1672 | "\tEpochs seen: 13\n", 1673 | "\tBatches seen: 130\n", 1674 | "\tExamples seen: 13000\n", 1675 | "\tlearning_rate: 0.099999986589\n", 1676 | "\tmomentum: 0.500000059605\n", 1677 | "\ttest_h0_col_norms_max: 1.29060137272\n", 1678 | "\ttest_h0_col_norms_mean: 1.06600356102\n", 1679 | "\ttest_h0_col_norms_min: 0.968188047409\n", 1680 | "\ttest_h0_max_x_max_u: 5.27886724472\n", 1681 | "\ttest_h0_max_x_mean_u: 2.07054448128\n", 1682 | "\ttest_h0_max_x_min_u: 0.0944659039378\n", 1683 | "\ttest_h0_mean_x_max_u: 2.34954357147\n", 1684 | "\ttest_h0_mean_x_mean_u: 0.690488278866\n", 1685 | "\ttest_h0_mean_x_min_u: 0.00203144224361\n", 1686 | "\ttest_h0_min_x_max_u: 0.548946678638\n", 1687 | "\ttest_h0_min_x_mean_u: 0.00964189134538\n", 1688 | "\ttest_h0_min_x_min_u: 0.0\n", 1689 | "\ttest_h0_range_x_max_u: 5.26520204544\n", 1690 | "\ttest_h0_range_x_mean_u: 2.06090283394\n", 1691 | "\ttest_h0_range_x_min_u: 0.0944659039378\n", 1692 | "\ttest_h0_row_norms_max: 0.523216068745\n", 1693 | "\ttest_h0_row_norms_mean: 0.379221647978\n", 1694 | "\ttest_h0_row_norms_min: 0.280473798513\n", 1695 | "\ttest_h1_col_norms_max: 1.31975185871\n", 1696 | "\ttest_h1_col_norms_mean: 1.06518316269\n", 1697 | "\ttest_h1_col_norms_min: 0.886828660965\n", 1698 | "\ttest_h1_max_x_max_u: 8.68297100067\n", 1699 | "\ttest_h1_max_x_mean_u: 2.95573759079\n", 1700 | "\ttest_h1_max_x_min_u: 0.0\n", 1701 | "\ttest_h1_mean_x_max_u: 2.69091677666\n", 1702 | "\ttest_h1_mean_x_mean_u: 0.921531140804\n", 1703 | "\ttest_h1_mean_x_min_u: 0.0\n", 1704 | "\ttest_h1_min_x_max_u: 0.0551998056471\n", 1705 | "\ttest_h1_min_x_mean_u: 0.000812256301288\n", 1706 | "\ttest_h1_min_x_min_u: 0.0\n", 1707 | "\ttest_h1_range_x_max_u: 8.68297100067\n", 1708 | "\ttest_h1_range_x_mean_u: 2.95492553711\n", 1709 | "\ttest_h1_range_x_min_u: 0.0\n", 1710 | "\ttest_h1_row_norms_max: 1.30019950867\n", 1711 | "\ttest_h1_row_norms_mean: 1.06629896164\n", 1712 | "\ttest_h1_row_norms_min: 0.892613649368\n", 1713 | "\ttest_objective: 0.495440065861\n", 1714 | "\ttest_y_col_norms_max: 1.37123155594\n", 1715 | "\ttest_y_col_norms_mean: 1.16500031948\n", 1716 | "\ttest_y_col_norms_min: 0.951216876507\n", 1717 | "\ttest_y_max_max_class: 0.999775230885\n", 1718 | "\ttest_y_mean_max_class: 0.846490442753\n", 1719 | "\ttest_y_min_max_class: 0.328825891018\n", 1720 | "\ttest_y_misclass: 0.134000003338\n", 1721 | "\ttest_y_nll: 0.495440065861\n", 1722 | "\ttest_y_row_norms_max: 0.753279983997\n", 1723 | "\ttest_y_row_norms_mean: 0.30464592576\n", 1724 | "\ttest_y_row_norms_min: 0.00266337138601\n", 1725 | "\ttotal_seconds_last_epoch: 0.391244977713\n", 1726 | "\ttrain_h0_col_norms_max: 1.29060137272\n", 1727 | "\ttrain_h0_col_norms_mean: 1.06600356102\n", 1728 | "\ttrain_h0_col_norms_min: 0.968188047409\n", 1729 | "\ttrain_h0_max_x_max_u: 5.13583707809\n", 1730 | "\ttrain_h0_max_x_mean_u: 2.02164840698\n", 1731 | "\ttrain_h0_max_x_min_u: 0.0830406472087\n", 1732 | "\ttrain_h0_mean_x_max_u: 2.28486585617\n", 1733 | "\ttrain_h0_mean_x_mean_u: 0.673171520233\n", 1734 | "\ttrain_h0_mean_x_min_u: 0.00176332960837\n", 1735 | "\ttrain_h0_min_x_max_u: 0.570852339268\n", 1736 | "\ttrain_h0_min_x_mean_u: 0.00793320592493\n", 1737 | "\ttrain_h0_min_x_min_u: 0.0\n", 1738 | "\ttrain_h0_range_x_max_u: 5.12853765488\n", 1739 | "\ttrain_h0_range_x_mean_u: 2.01371502876\n", 1740 | "\ttrain_h0_range_x_min_u: 0.0830406472087\n", 1741 | "\ttrain_h0_row_norms_max: 0.523216068745\n", 1742 | "\ttrain_h0_row_norms_mean: 0.379221647978\n", 1743 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 1744 | "\ttrain_h1_col_norms_max: 1.31975185871\n", 1745 | "\ttrain_h1_col_norms_mean: 1.06518316269\n", 1746 | "\ttrain_h1_col_norms_min: 0.886828660965\n", 1747 | "\ttrain_h1_max_x_max_u: 8.50638008118\n", 1748 | "\ttrain_h1_max_x_mean_u: 2.89869999886\n", 1749 | "\ttrain_h1_max_x_min_u: 0.0\n", 1750 | "\ttrain_h1_mean_x_max_u: 2.69055533409\n", 1751 | "\ttrain_h1_mean_x_mean_u: 0.903810441494\n", 1752 | "\ttrain_h1_mean_x_min_u: 0.0\n", 1753 | "\ttrain_h1_min_x_max_u: 0.00346307968721\n", 1754 | "\ttrain_h1_min_x_mean_u: 3.46307970176e-05\n", 1755 | "\ttrain_h1_min_x_min_u: 0.0\n", 1756 | "\ttrain_h1_range_x_max_u: 8.50638008118\n", 1757 | "\ttrain_h1_range_x_mean_u: 2.89866566658\n", 1758 | "\ttrain_h1_range_x_min_u: 0.0\n", 1759 | "\ttrain_h1_row_norms_max: 1.30019950867\n", 1760 | "\ttrain_h1_row_norms_mean: 1.06629896164\n", 1761 | "\ttrain_h1_row_norms_min: 0.892613649368\n", 1762 | "\ttrain_objective: 0.225919157267\n", 1763 | "\ttrain_y_col_norms_max: 1.37123155594\n", 1764 | "\ttrain_y_col_norms_mean: 1.16500031948\n", 1765 | "\ttrain_y_col_norms_min: 0.951216876507\n", 1766 | "\ttrain_y_max_max_class: 0.99980866909\n", 1767 | "\ttrain_y_mean_max_class: 0.880507230759\n", 1768 | "\ttrain_y_min_max_class: 0.333441525698\n", 1769 | "\ttrain_y_misclass: 0.0619999989867\n", 1770 | "\ttrain_y_nll: 0.225919157267\n", 1771 | "\ttrain_y_row_norms_max: 0.753279983997\n", 1772 | "\ttrain_y_row_norms_mean: 0.30464592576\n", 1773 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 1774 | "\ttraining_seconds_this_epoch: 0.0766139999032\n", 1775 | "Time this epoch: 0.076868 seconds\n", 1776 | "Monitoring step:\n", 1777 | "\tEpochs seen: 14\n", 1778 | "\tBatches seen: 140\n", 1779 | "\tExamples seen: 14000\n", 1780 | "\tlearning_rate: 0.099999986589\n", 1781 | "\tmomentum: 0.500000059605\n", 1782 | "\ttest_h0_col_norms_max: 1.30594861507\n", 1783 | "\ttest_h0_col_norms_mean: 1.0690485239\n", 1784 | "\ttest_h0_col_norms_min: 0.968669712543\n", 1785 | "\ttest_h0_max_x_max_u: 5.29942798615\n", 1786 | "\ttest_h0_max_x_mean_u: 2.08129835129\n", 1787 | "\ttest_h0_max_x_min_u: 0.100918442011\n", 1788 | "\ttest_h0_mean_x_max_u: 2.35635018349\n", 1789 | "\ttest_h0_mean_x_mean_u: 0.694277226925\n", 1790 | "\ttest_h0_mean_x_min_u: 0.00219478178769\n", 1791 | "\ttest_h0_min_x_max_u: 0.569123864174\n", 1792 | "\ttest_h0_min_x_mean_u: 0.00975045282394\n", 1793 | "\ttest_h0_min_x_min_u: 0.0\n", 1794 | "\ttest_h0_range_x_max_u: 5.28551721573\n", 1795 | "\ttest_h0_range_x_mean_u: 2.07154798508\n", 1796 | "\ttest_h0_range_x_min_u: 0.100918442011\n", 1797 | "\ttest_h0_row_norms_max: 0.532616436481\n", 1798 | "\ttest_h0_row_norms_mean: 0.38025650382\n", 1799 | "\ttest_h0_row_norms_min: 0.280473798513\n", 1800 | "\ttest_h1_col_norms_max: 1.32189166546\n", 1801 | "\ttest_h1_col_norms_mean: 1.06786954403\n", 1802 | "\ttest_h1_col_norms_min: 0.890783011913\n", 1803 | "\ttest_h1_max_x_max_u: 9.04478549957\n", 1804 | "\ttest_h1_max_x_mean_u: 2.99396109581\n", 1805 | "\ttest_h1_max_x_min_u: 0.0\n", 1806 | "\ttest_h1_mean_x_max_u: 2.88102555275\n", 1807 | "\ttest_h1_mean_x_mean_u: 0.93541187048\n", 1808 | "\ttest_h1_mean_x_min_u: 0.0\n", 1809 | "\ttest_h1_min_x_max_u: 0.0271273870021\n", 1810 | "\ttest_h1_min_x_mean_u: 0.000445573910838\n", 1811 | "\ttest_h1_min_x_min_u: 0.0\n", 1812 | "\ttest_h1_range_x_max_u: 9.04478549957\n", 1813 | "\ttest_h1_range_x_mean_u: 2.99351525307\n", 1814 | "\ttest_h1_range_x_min_u: 0.0\n", 1815 | "\ttest_h1_row_norms_max: 1.3142465353\n", 1816 | "\ttest_h1_row_norms_mean: 1.06904518604\n", 1817 | "\ttest_h1_row_norms_min: 0.893028438091\n", 1818 | "\ttest_objective: 0.468659281731\n", 1819 | "\ttest_y_col_norms_max: 1.39471578598\n", 1820 | "\ttest_y_col_norms_mean: 1.1923186779\n", 1821 | "\ttest_y_col_norms_min: 0.982104003429\n", 1822 | "\ttest_y_max_max_class: 0.999705731869\n", 1823 | "\ttest_y_mean_max_class: 0.848345577717\n", 1824 | "\ttest_y_min_max_class: 0.302234917879\n", 1825 | "\ttest_y_misclass: 0.148000001907\n", 1826 | "\ttest_y_nll: 0.468659281731\n", 1827 | "\ttest_y_row_norms_max: 0.778974711895\n", 1828 | "\ttest_y_row_norms_mean: 0.311919778585\n", 1829 | "\ttest_y_row_norms_min: 0.00266337138601\n", 1830 | "\ttotal_seconds_last_epoch: 0.391975015402\n", 1831 | "\ttrain_h0_col_norms_max: 1.30594861507\n", 1832 | "\ttrain_h0_col_norms_mean: 1.0690485239\n", 1833 | "\ttrain_h0_col_norms_min: 0.968669712543\n", 1834 | "\ttrain_h0_max_x_max_u: 5.16828250885\n", 1835 | "\ttrain_h0_max_x_mean_u: 2.03235030174\n", 1836 | "\ttrain_h0_max_x_min_u: 0.0902600586414\n", 1837 | "\ttrain_h0_mean_x_max_u: 2.29059481621\n", 1838 | "\ttrain_h0_mean_x_mean_u: 0.676722168922\n", 1839 | "\ttrain_h0_mean_x_min_u: 0.00194805779029\n", 1840 | "\ttrain_h0_min_x_max_u: 0.591661036015\n", 1841 | "\ttrain_h0_min_x_mean_u: 0.00771585060284\n", 1842 | "\ttrain_h0_min_x_min_u: 0.0\n", 1843 | "\ttrain_h0_range_x_max_u: 5.15138053894\n", 1844 | "\ttrain_h0_range_x_mean_u: 2.02463459969\n", 1845 | "\ttrain_h0_range_x_min_u: 0.0902600586414\n", 1846 | "\ttrain_h0_row_norms_max: 0.532616436481\n", 1847 | "\ttrain_h0_row_norms_mean: 0.38025650382\n", 1848 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 1849 | "\ttrain_h1_col_norms_max: 1.32189166546\n", 1850 | "\ttrain_h1_col_norms_mean: 1.06786954403\n", 1851 | "\ttrain_h1_col_norms_min: 0.890783011913\n", 1852 | "\ttrain_h1_max_x_max_u: 8.83148670197\n", 1853 | "\ttrain_h1_max_x_mean_u: 2.93978428841\n", 1854 | "\ttrain_h1_max_x_min_u: 0.0\n", 1855 | "\ttrain_h1_mean_x_max_u: 2.83285927773\n", 1856 | "\ttrain_h1_mean_x_mean_u: 0.916635155678\n", 1857 | "\ttrain_h1_mean_x_min_u: 0.0\n", 1858 | "\ttrain_h1_min_x_max_u: 0.0171617604792\n", 1859 | "\ttrain_h1_min_x_mean_u: 0.00017161759024\n", 1860 | "\ttrain_h1_min_x_min_u: 0.0\n", 1861 | "\ttrain_h1_range_x_max_u: 8.83148670197\n", 1862 | "\ttrain_h1_range_x_mean_u: 2.93961262703\n", 1863 | "\ttrain_h1_range_x_min_u: 0.0\n", 1864 | "\ttrain_h1_row_norms_max: 1.3142465353\n", 1865 | "\ttrain_h1_row_norms_mean: 1.06904518604\n", 1866 | "\ttrain_h1_row_norms_min: 0.893028438091\n", 1867 | "\ttrain_objective: 0.19838373363\n", 1868 | "\ttrain_y_col_norms_max: 1.39471578598\n", 1869 | "\ttrain_y_col_norms_mean: 1.1923186779\n", 1870 | "\ttrain_y_col_norms_min: 0.982104003429\n", 1871 | "\ttrain_y_max_max_class: 0.999663114548\n", 1872 | "\ttrain_y_mean_max_class: 0.883097231388\n", 1873 | "\ttrain_y_min_max_class: 0.340608209372\n", 1874 | "\ttrain_y_misclass: 0.0439999997616\n", 1875 | "\ttrain_y_nll: 0.19838373363\n", 1876 | "\ttrain_y_row_norms_max: 0.778974711895\n", 1877 | "\ttrain_y_row_norms_mean: 0.311919778585\n", 1878 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 1879 | "\ttraining_seconds_this_epoch: 0.0768680050969\n", 1880 | "Time this epoch: 0.077220 seconds\n", 1881 | "Monitoring step:\n", 1882 | "\tEpochs seen: 15\n", 1883 | "\tBatches seen: 150\n", 1884 | "\tExamples seen: 15000\n", 1885 | "\tlearning_rate: 0.099999986589\n", 1886 | "\tmomentum: 0.500000059605\n", 1887 | "\ttest_h0_col_norms_max: 1.32223975658\n", 1888 | "\ttest_h0_col_norms_mean: 1.07213902473\n", 1889 | "\ttest_h0_col_norms_min: 0.969218611717\n", 1890 | "\ttest_h0_max_x_max_u: 5.26269626617\n", 1891 | "\ttest_h0_max_x_mean_u: 2.10021162033\n", 1892 | "\ttest_h0_max_x_min_u: 0.10486997664\n", 1893 | "\ttest_h0_mean_x_max_u: 2.34699392319\n", 1894 | "\ttest_h0_mean_x_mean_u: 0.701870262623\n", 1895 | "\ttest_h0_mean_x_min_u: 0.00228205672465\n", 1896 | "\ttest_h0_min_x_max_u: 0.569497942924\n", 1897 | "\ttest_h0_min_x_mean_u: 0.00974835269153\n", 1898 | "\ttest_h0_min_x_min_u: 0.0\n", 1899 | "\ttest_h0_range_x_max_u: 5.24736881256\n", 1900 | "\ttest_h0_range_x_mean_u: 2.09046316147\n", 1901 | "\ttest_h0_range_x_min_u: 0.10486997664\n", 1902 | "\ttest_h0_row_norms_max: 0.538556814194\n", 1903 | "\ttest_h0_row_norms_mean: 0.38131120801\n", 1904 | "\ttest_h0_row_norms_min: 0.280473798513\n", 1905 | "\ttest_h1_col_norms_max: 1.33036005497\n", 1906 | "\ttest_h1_col_norms_mean: 1.07063043118\n", 1907 | "\ttest_h1_col_norms_min: 0.89462095499\n", 1908 | "\ttest_h1_max_x_max_u: 9.04273509979\n", 1909 | "\ttest_h1_max_x_mean_u: 3.05497527122\n", 1910 | "\ttest_h1_max_x_min_u: 0.0\n", 1911 | "\ttest_h1_mean_x_max_u: 2.86822247505\n", 1912 | "\ttest_h1_mean_x_mean_u: 0.956351459026\n", 1913 | "\ttest_h1_mean_x_min_u: 0.0\n", 1914 | "\ttest_h1_min_x_max_u: 0.0496255382895\n", 1915 | "\ttest_h1_min_x_mean_u: 0.000700458767824\n", 1916 | "\ttest_h1_min_x_min_u: 0.0\n", 1917 | "\ttest_h1_range_x_max_u: 9.04273509979\n", 1918 | "\ttest_h1_range_x_mean_u: 3.05427479744\n", 1919 | "\ttest_h1_range_x_min_u: 0.0\n", 1920 | "\ttest_h1_row_norms_max: 1.32876288891\n", 1921 | "\ttest_h1_row_norms_mean: 1.07187759876\n", 1922 | "\ttest_h1_row_norms_min: 0.89346575737\n", 1923 | "\ttest_objective: 0.462522119284\n", 1924 | "\ttest_y_col_norms_max: 1.41865599155\n", 1925 | "\ttest_y_col_norms_mean: 1.21935606003\n", 1926 | "\ttest_y_col_norms_min: 0.999803483486\n", 1927 | "\ttest_y_max_max_class: 0.999775528908\n", 1928 | "\ttest_y_mean_max_class: 0.857300519943\n", 1929 | "\ttest_y_min_max_class: 0.311234474182\n", 1930 | "\ttest_y_misclass: 0.137999996543\n", 1931 | "\ttest_y_nll: 0.462522119284\n", 1932 | "\ttest_y_row_norms_max: 0.791741073132\n", 1933 | "\ttest_y_row_norms_mean: 0.319242209196\n", 1934 | "\ttest_y_row_norms_min: 0.00266337138601\n", 1935 | "\ttotal_seconds_last_epoch: 0.390672057867\n", 1936 | "\ttrain_h0_col_norms_max: 1.32223975658\n", 1937 | "\ttrain_h0_col_norms_mean: 1.07213902473\n", 1938 | "\ttrain_h0_col_norms_min: 0.969218611717\n", 1939 | "\ttrain_h0_max_x_max_u: 5.12161302567\n", 1940 | "\ttrain_h0_max_x_mean_u: 2.04921555519\n", 1941 | "\ttrain_h0_max_x_min_u: 0.0975362285972\n", 1942 | "\ttrain_h0_mean_x_max_u: 2.27929067612\n", 1943 | "\ttrain_h0_mean_x_mean_u: 0.684022068977\n", 1944 | "\ttrain_h0_mean_x_min_u: 0.00203897664323\n", 1945 | "\ttrain_h0_min_x_max_u: 0.59866553545\n", 1946 | "\ttrain_h0_min_x_mean_u: 0.00747454492375\n", 1947 | "\ttrain_h0_min_x_min_u: 0.0\n", 1948 | "\ttrain_h0_range_x_max_u: 5.10862159729\n", 1949 | "\ttrain_h0_range_x_mean_u: 2.04174089432\n", 1950 | "\ttrain_h0_range_x_min_u: 0.0975362285972\n", 1951 | "\ttrain_h0_row_norms_max: 0.538556814194\n", 1952 | "\ttrain_h0_row_norms_mean: 0.38131120801\n", 1953 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 1954 | "\ttrain_h1_col_norms_max: 1.33036005497\n", 1955 | "\ttrain_h1_col_norms_mean: 1.07063043118\n", 1956 | "\ttrain_h1_col_norms_min: 0.89462095499\n", 1957 | "\ttrain_h1_max_x_max_u: 8.84126472473\n", 1958 | "\ttrain_h1_max_x_mean_u: 2.99682807922\n", 1959 | "\ttrain_h1_max_x_min_u: 0.0\n", 1960 | "\ttrain_h1_mean_x_max_u: 2.81727814674\n", 1961 | "\ttrain_h1_mean_x_mean_u: 0.937330007553\n", 1962 | "\ttrain_h1_mean_x_min_u: 0.0\n", 1963 | "\ttrain_h1_min_x_max_u: 0.0128950383514\n", 1964 | "\ttrain_h1_min_x_mean_u: 0.00012895038526\n", 1965 | "\ttrain_h1_min_x_min_u: 0.0\n", 1966 | "\ttrain_h1_range_x_max_u: 8.84126472473\n", 1967 | "\ttrain_h1_range_x_mean_u: 2.99669933319\n", 1968 | "\ttrain_h1_range_x_min_u: 0.0\n", 1969 | "\ttrain_h1_row_norms_max: 1.32876288891\n", 1970 | "\ttrain_h1_row_norms_mean: 1.07187759876\n", 1971 | "\ttrain_h1_row_norms_min: 0.89346575737\n", 1972 | "\ttrain_objective: 0.183309301734\n", 1973 | "\ttrain_y_col_norms_max: 1.41865599155\n", 1974 | "\ttrain_y_col_norms_mean: 1.21935606003\n", 1975 | "\ttrain_y_col_norms_min: 0.999803483486\n", 1976 | "\ttrain_y_max_max_class: 0.99985152483\n", 1977 | "\ttrain_y_mean_max_class: 0.891006588936\n", 1978 | "\ttrain_y_min_max_class: 0.353921562433\n", 1979 | "\ttrain_y_misclass: 0.0530000030994\n", 1980 | "\ttrain_y_nll: 0.183309301734\n", 1981 | "\ttrain_y_row_norms_max: 0.791741073132\n", 1982 | "\ttrain_y_row_norms_mean: 0.319242209196\n", 1983 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 1984 | "\ttraining_seconds_this_epoch: 0.0772200003266\n", 1985 | "Saving to ./model_mnist.pkl...\n", 1986 | "Saving to ./model_mnist.pkl done. Time elapsed: 0.089112 seconds\n", 1987 | "Time this epoch: 0.077308 seconds\n", 1988 | "Monitoring step:\n", 1989 | "\tEpochs seen: 16\n", 1990 | "\tBatches seen: 160\n", 1991 | "\tExamples seen: 16000\n", 1992 | "\tlearning_rate: 0.099999986589\n", 1993 | "\tmomentum: 0.500000059605\n", 1994 | "\ttest_h0_col_norms_max: 1.33124780655\n", 1995 | "\ttest_h0_col_norms_mean: 1.07505905628\n", 1996 | "\ttest_h0_col_norms_min: 0.969755113125\n", 1997 | "\ttest_h0_max_x_max_u: 5.20827150345\n", 1998 | "\ttest_h0_max_x_mean_u: 2.10036230087\n", 1999 | "\ttest_h0_max_x_min_u: 0.109824202955\n", 2000 | "\ttest_h0_mean_x_max_u: 2.30157351494\n", 2001 | "\ttest_h0_mean_x_mean_u: 0.697842955589\n", 2002 | "\ttest_h0_mean_x_min_u: 0.00241158530116\n", 2003 | "\ttest_h0_min_x_max_u: 0.558968901634\n", 2004 | "\ttest_h0_min_x_mean_u: 0.00927729532123\n", 2005 | "\ttest_h0_min_x_min_u: 0.0\n", 2006 | "\ttest_h0_range_x_max_u: 5.19507265091\n", 2007 | "\ttest_h0_range_x_mean_u: 2.09108543396\n", 2008 | "\ttest_h0_range_x_min_u: 0.109824202955\n", 2009 | "\ttest_h0_row_norms_max: 0.545352041721\n", 2010 | "\ttest_h0_row_norms_mean: 0.382304221392\n", 2011 | "\ttest_h0_row_norms_min: 0.280473798513\n", 2012 | "\ttest_h1_col_norms_max: 1.33779060841\n", 2013 | "\ttest_h1_col_norms_mean: 1.07314157486\n", 2014 | "\ttest_h1_col_norms_min: 0.896570086479\n", 2015 | "\ttest_h1_max_x_max_u: 8.98954582214\n", 2016 | "\ttest_h1_max_x_mean_u: 3.05989265442\n", 2017 | "\ttest_h1_max_x_min_u: 0.0\n", 2018 | "\ttest_h1_mean_x_max_u: 2.80987167358\n", 2019 | "\ttest_h1_mean_x_mean_u: 0.948491275311\n", 2020 | "\ttest_h1_mean_x_min_u: 0.0\n", 2021 | "\ttest_h1_min_x_max_u: 0.0724629089236\n", 2022 | "\ttest_h1_min_x_mean_u: 0.000807438394986\n", 2023 | "\ttest_h1_min_x_min_u: 0.0\n", 2024 | "\ttest_h1_range_x_max_u: 8.98954582214\n", 2025 | "\ttest_h1_range_x_mean_u: 3.05908513069\n", 2026 | "\ttest_h1_range_x_min_u: 0.0\n", 2027 | "\ttest_h1_row_norms_max: 1.33591425419\n", 2028 | "\ttest_h1_row_norms_mean: 1.0744549036\n", 2029 | "\ttest_h1_row_norms_min: 0.893976449966\n", 2030 | "\ttest_objective: 0.476188957691\n", 2031 | "\ttest_y_col_norms_max: 1.43042373657\n", 2032 | "\ttest_y_col_norms_mean: 1.24442100525\n", 2033 | "\ttest_y_col_norms_min: 1.03444623947\n", 2034 | "\ttest_y_max_max_class: 0.999842941761\n", 2035 | "\ttest_y_mean_max_class: 0.863466084003\n", 2036 | "\ttest_y_min_max_class: 0.35065561533\n", 2037 | "\ttest_y_misclass: 0.135000020266\n", 2038 | "\ttest_y_nll: 0.476188957691\n", 2039 | "\ttest_y_row_norms_max: 0.807280480862\n", 2040 | "\ttest_y_row_norms_mean: 0.326072484255\n", 2041 | "\ttest_y_row_norms_min: 0.00266337138601\n", 2042 | "\ttotal_seconds_last_epoch: 0.48313704133\n", 2043 | "\ttrain_h0_col_norms_max: 1.33124780655\n", 2044 | "\ttrain_h0_col_norms_mean: 1.07505905628\n", 2045 | "\ttrain_h0_col_norms_min: 0.969755113125\n", 2046 | "\ttrain_h0_max_x_max_u: 5.06391477585\n", 2047 | "\ttrain_h0_max_x_mean_u: 2.04905104637\n", 2048 | "\ttrain_h0_max_x_min_u: 0.101039871573\n", 2049 | "\ttrain_h0_mean_x_max_u: 2.2342300415\n", 2050 | "\ttrain_h0_mean_x_mean_u: 0.680156111717\n", 2051 | "\ttrain_h0_mean_x_min_u: 0.00213845772669\n", 2052 | "\ttrain_h0_min_x_max_u: 0.584780037403\n", 2053 | "\ttrain_h0_min_x_mean_u: 0.00739581463858\n", 2054 | "\ttrain_h0_min_x_min_u: 0.0\n", 2055 | "\ttrain_h0_range_x_max_u: 5.06391477585\n", 2056 | "\ttrain_h0_range_x_mean_u: 2.04165554047\n", 2057 | "\ttrain_h0_range_x_min_u: 0.101039871573\n", 2058 | "\ttrain_h0_row_norms_max: 0.545352041721\n", 2059 | "\ttrain_h0_row_norms_mean: 0.382304221392\n", 2060 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 2061 | "\ttrain_h1_col_norms_max: 1.33779060841\n", 2062 | "\ttrain_h1_col_norms_mean: 1.07314157486\n", 2063 | "\ttrain_h1_col_norms_min: 0.896570086479\n", 2064 | "\ttrain_h1_max_x_max_u: 8.85223007202\n", 2065 | "\ttrain_h1_max_x_mean_u: 3.00178360939\n", 2066 | "\ttrain_h1_max_x_min_u: 0.0\n", 2067 | "\ttrain_h1_mean_x_max_u: 2.79580116272\n", 2068 | "\ttrain_h1_mean_x_mean_u: 0.92992657423\n", 2069 | "\ttrain_h1_mean_x_min_u: 0.0\n", 2070 | "\ttrain_h1_min_x_max_u: 0.0\n", 2071 | "\ttrain_h1_min_x_mean_u: 0.0\n", 2072 | "\ttrain_h1_min_x_min_u: 0.0\n", 2073 | "\ttrain_h1_range_x_max_u: 8.85223007202\n", 2074 | "\ttrain_h1_range_x_mean_u: 3.00178360939\n", 2075 | "\ttrain_h1_range_x_min_u: 0.0\n", 2076 | "\ttrain_h1_row_norms_max: 1.33591425419\n", 2077 | "\ttrain_h1_row_norms_mean: 1.0744549036\n", 2078 | "\ttrain_h1_row_norms_min: 0.893976449966\n", 2079 | "\ttrain_objective: 0.155664160848\n", 2080 | "\ttrain_y_col_norms_max: 1.43042373657\n", 2081 | "\ttrain_y_col_norms_mean: 1.24442100525\n", 2082 | "\ttrain_y_col_norms_min: 1.03444623947\n", 2083 | "\ttrain_y_max_max_class: 0.999903380871\n", 2084 | "\ttrain_y_mean_max_class: 0.905024886131\n", 2085 | "\ttrain_y_min_max_class: 0.349106669426\n", 2086 | "\ttrain_y_misclass: 0.0359999984503\n", 2087 | "\ttrain_y_nll: 0.155664160848\n", 2088 | "\ttrain_y_row_norms_max: 0.807280480862\n", 2089 | "\ttrain_y_row_norms_mean: 0.326072484255\n", 2090 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 2091 | "\ttraining_seconds_this_epoch: 0.0773079991341\n", 2092 | "Time this epoch: 0.076765 seconds\n", 2093 | "Monitoring step:\n", 2094 | "\tEpochs seen: 17\n", 2095 | "\tBatches seen: 170\n", 2096 | "\tExamples seen: 17000\n", 2097 | "\tlearning_rate: 0.099999986589\n", 2098 | "\tmomentum: 0.500000059605\n", 2099 | "\ttest_h0_col_norms_max: 1.34282577038\n", 2100 | "\ttest_h0_col_norms_mean: 1.07798469067\n", 2101 | "\ttest_h0_col_norms_min: 0.970460116863\n", 2102 | "\ttest_h0_max_x_max_u: 5.36118173599\n", 2103 | "\ttest_h0_max_x_mean_u: 2.11505937576\n", 2104 | "\ttest_h0_max_x_min_u: 0.112354725599\n", 2105 | "\ttest_h0_mean_x_max_u: 2.41735649109\n", 2106 | "\ttest_h0_mean_x_mean_u: 0.703463435173\n", 2107 | "\ttest_h0_mean_x_min_u: 0.00249769957736\n", 2108 | "\ttest_h0_min_x_max_u: 0.551767468452\n", 2109 | "\ttest_h0_min_x_mean_u: 0.0101659270003\n", 2110 | "\ttest_h0_min_x_min_u: 0.0\n", 2111 | "\ttest_h0_range_x_max_u: 5.34299945831\n", 2112 | "\ttest_h0_range_x_mean_u: 2.10489344597\n", 2113 | "\ttest_h0_range_x_min_u: 0.112354725599\n", 2114 | "\ttest_h0_row_norms_max: 0.551928222179\n", 2115 | "\ttest_h0_row_norms_mean: 0.383294433355\n", 2116 | "\ttest_h0_row_norms_min: 0.280473798513\n", 2117 | "\ttest_h1_col_norms_max: 1.3447688818\n", 2118 | "\ttest_h1_col_norms_mean: 1.07580840588\n", 2119 | "\ttest_h1_col_norms_min: 0.896570086479\n", 2120 | "\ttest_h1_max_x_max_u: 8.9066696167\n", 2121 | "\ttest_h1_max_x_mean_u: 3.11536884308\n", 2122 | "\ttest_h1_max_x_min_u: 0.0\n", 2123 | "\ttest_h1_mean_x_max_u: 2.82497262955\n", 2124 | "\ttest_h1_mean_x_mean_u: 0.970781683922\n", 2125 | "\ttest_h1_mean_x_min_u: 0.0\n", 2126 | "\ttest_h1_min_x_max_u: 0.0630861967802\n", 2127 | "\ttest_h1_min_x_mean_u: 0.000853632576764\n", 2128 | "\ttest_h1_min_x_min_u: 0.0\n", 2129 | "\ttest_h1_range_x_max_u: 8.9066696167\n", 2130 | "\ttest_h1_range_x_mean_u: 3.11451530457\n", 2131 | "\ttest_h1_range_x_min_u: 0.0\n", 2132 | "\ttest_h1_row_norms_max: 1.34632813931\n", 2133 | "\ttest_h1_row_norms_mean: 1.0771780014\n", 2134 | "\ttest_h1_row_norms_min: 0.89445245266\n", 2135 | "\ttest_objective: 0.499997317791\n", 2136 | "\ttest_y_col_norms_max: 1.45787668228\n", 2137 | "\ttest_y_col_norms_mean: 1.26962137222\n", 2138 | "\ttest_y_col_norms_min: 1.06204032898\n", 2139 | "\ttest_y_max_max_class: 0.999864578247\n", 2140 | "\ttest_y_mean_max_class: 0.87062984705\n", 2141 | "\ttest_y_min_max_class: 0.332253485918\n", 2142 | "\ttest_y_misclass: 0.129999995232\n", 2143 | "\ttest_y_nll: 0.499997317791\n", 2144 | "\ttest_y_row_norms_max: 0.820121824741\n", 2145 | "\ttest_y_row_norms_mean: 0.332938224077\n", 2146 | "\ttest_y_row_norms_min: 0.00266337138601\n", 2147 | "\ttotal_seconds_last_epoch: 0.392895966768\n", 2148 | "\ttrain_h0_col_norms_max: 1.34282577038\n", 2149 | "\ttrain_h0_col_norms_mean: 1.07798469067\n", 2150 | "\ttrain_h0_col_norms_min: 0.970460116863\n", 2151 | "\ttrain_h0_max_x_max_u: 5.11973667145\n", 2152 | "\ttrain_h0_max_x_mean_u: 2.06346940994\n", 2153 | "\ttrain_h0_max_x_min_u: 0.103111185133\n", 2154 | "\ttrain_h0_mean_x_max_u: 2.34504413605\n", 2155 | "\ttrain_h0_mean_x_mean_u: 0.685505449772\n", 2156 | "\ttrain_h0_mean_x_min_u: 0.00216499948874\n", 2157 | "\ttrain_h0_min_x_max_u: 0.580263853073\n", 2158 | "\ttrain_h0_min_x_mean_u: 0.00772378966212\n", 2159 | "\ttrain_h0_min_x_min_u: 0.0\n", 2160 | "\ttrain_h0_range_x_max_u: 5.11168718338\n", 2161 | "\ttrain_h0_range_x_mean_u: 2.05574584007\n", 2162 | "\ttrain_h0_range_x_min_u: 0.103111185133\n", 2163 | "\ttrain_h0_row_norms_max: 0.551928222179\n", 2164 | "\ttrain_h0_row_norms_mean: 0.383294433355\n", 2165 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 2166 | "\ttrain_h1_col_norms_max: 1.3447688818\n", 2167 | "\ttrain_h1_col_norms_mean: 1.07580840588\n", 2168 | "\ttrain_h1_col_norms_min: 0.896570086479\n", 2169 | "\ttrain_h1_max_x_max_u: 8.84166717529\n", 2170 | "\ttrain_h1_max_x_mean_u: 3.05626320839\n", 2171 | "\ttrain_h1_max_x_min_u: 0.0\n", 2172 | "\ttrain_h1_mean_x_max_u: 2.78574609756\n", 2173 | "\ttrain_h1_mean_x_mean_u: 0.951011180878\n", 2174 | "\ttrain_h1_mean_x_min_u: 0.0\n", 2175 | "\ttrain_h1_min_x_max_u: 0.0\n", 2176 | "\ttrain_h1_min_x_mean_u: 0.0\n", 2177 | "\ttrain_h1_min_x_min_u: 0.0\n", 2178 | "\ttrain_h1_range_x_max_u: 8.84166717529\n", 2179 | "\ttrain_h1_range_x_mean_u: 3.05626320839\n", 2180 | "\ttrain_h1_range_x_min_u: 0.0\n", 2181 | "\ttrain_h1_row_norms_max: 1.34632813931\n", 2182 | "\ttrain_h1_row_norms_mean: 1.0771780014\n", 2183 | "\ttrain_h1_row_norms_min: 0.89445245266\n", 2184 | "\ttrain_objective: 0.147527873516\n", 2185 | "\ttrain_y_col_norms_max: 1.45787668228\n", 2186 | "\ttrain_y_col_norms_mean: 1.26962137222\n", 2187 | "\ttrain_y_col_norms_min: 1.06204032898\n", 2188 | "\ttrain_y_max_max_class: 0.999917984009\n", 2189 | "\ttrain_y_mean_max_class: 0.908993124962\n", 2190 | "\ttrain_y_min_max_class: 0.362698823214\n", 2191 | "\ttrain_y_misclass: 0.0330000035465\n", 2192 | "\ttrain_y_nll: 0.147527873516\n", 2193 | "\ttrain_y_row_norms_max: 0.820121824741\n", 2194 | "\ttrain_y_row_norms_mean: 0.332938224077\n", 2195 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 2196 | "\ttraining_seconds_this_epoch: 0.0767650008202\n", 2197 | "Time this epoch: 0.076701 seconds\n", 2198 | "Monitoring step:\n", 2199 | "\tEpochs seen: 18\n", 2200 | "\tBatches seen: 180\n", 2201 | "\tExamples seen: 18000\n", 2202 | "\tlearning_rate: 0.099999986589\n", 2203 | "\tmomentum: 0.500000059605\n", 2204 | "\ttest_h0_col_norms_max: 1.36036837101\n", 2205 | "\ttest_h0_col_norms_mean: 1.08088362217\n", 2206 | "\ttest_h0_col_norms_min: 0.971190273762\n", 2207 | "\ttest_h0_max_x_max_u: 5.37962865829\n", 2208 | "\ttest_h0_max_x_mean_u: 2.12070894241\n", 2209 | "\ttest_h0_max_x_min_u: 0.119706019759\n", 2210 | "\ttest_h0_mean_x_max_u: 2.42943692207\n", 2211 | "\ttest_h0_mean_x_mean_u: 0.701944708824\n", 2212 | "\ttest_h0_mean_x_min_u: 0.00280549121089\n", 2213 | "\ttest_h0_min_x_max_u: 0.554049432278\n", 2214 | "\ttest_h0_min_x_mean_u: 0.0102141033858\n", 2215 | "\ttest_h0_min_x_min_u: 0.0\n", 2216 | "\ttest_h0_range_x_max_u: 5.36142396927\n", 2217 | "\ttest_h0_range_x_mean_u: 2.11049461365\n", 2218 | "\ttest_h0_range_x_min_u: 0.119706019759\n", 2219 | "\ttest_h0_row_norms_max: 0.557518601418\n", 2220 | "\ttest_h0_row_norms_mean: 0.384280800819\n", 2221 | "\ttest_h0_row_norms_min: 0.280473798513\n", 2222 | "\ttest_h1_col_norms_max: 1.3490973711\n", 2223 | "\ttest_h1_col_norms_mean: 1.07836294174\n", 2224 | "\ttest_h1_col_norms_min: 0.896570086479\n", 2225 | "\ttest_h1_max_x_max_u: 9.4879989624\n", 2226 | "\ttest_h1_max_x_mean_u: 3.14028596878\n", 2227 | "\ttest_h1_max_x_min_u: 0.0\n", 2228 | "\ttest_h1_mean_x_max_u: 3.05321621895\n", 2229 | "\ttest_h1_mean_x_mean_u: 0.976984739304\n", 2230 | "\ttest_h1_mean_x_min_u: 0.0\n", 2231 | "\ttest_h1_min_x_max_u: 0.0363551490009\n", 2232 | "\ttest_h1_min_x_mean_u: 0.00043936088332\n", 2233 | "\ttest_h1_min_x_min_u: 0.0\n", 2234 | "\ttest_h1_range_x_max_u: 9.4879989624\n", 2235 | "\ttest_h1_range_x_mean_u: 3.1398460865\n", 2236 | "\ttest_h1_range_x_min_u: 0.0\n", 2237 | "\ttest_h1_row_norms_max: 1.36258947849\n", 2238 | "\ttest_h1_row_norms_mean: 1.07979047298\n", 2239 | "\ttest_h1_row_norms_min: 0.8949213624\n", 2240 | "\ttest_objective: 0.493919700384\n", 2241 | "\ttest_y_col_norms_max: 1.47577524185\n", 2242 | "\ttest_y_col_norms_mean: 1.29343950748\n", 2243 | "\ttest_y_col_norms_min: 1.08780992031\n", 2244 | "\ttest_y_max_max_class: 0.99991106987\n", 2245 | "\ttest_y_mean_max_class: 0.870437264442\n", 2246 | "\ttest_y_min_max_class: 0.340471714735\n", 2247 | "\ttest_y_misclass: 0.148000001907\n", 2248 | "\ttest_y_nll: 0.493919700384\n", 2249 | "\ttest_y_row_norms_max: 0.838829219341\n", 2250 | "\ttest_y_row_norms_mean: 0.339229732752\n", 2251 | "\ttest_y_row_norms_min: 0.00266337138601\n", 2252 | "\ttotal_seconds_last_epoch: 0.392343997955\n", 2253 | "\ttrain_h0_col_norms_max: 1.36036837101\n", 2254 | "\ttrain_h0_col_norms_mean: 1.08088362217\n", 2255 | "\ttrain_h0_col_norms_min: 0.971190273762\n", 2256 | "\ttrain_h0_max_x_max_u: 5.19341230392\n", 2257 | "\ttrain_h0_max_x_mean_u: 2.06951165199\n", 2258 | "\ttrain_h0_max_x_min_u: 0.111943788826\n", 2259 | "\ttrain_h0_mean_x_max_u: 2.35668778419\n", 2260 | "\ttrain_h0_mean_x_mean_u: 0.68404597044\n", 2261 | "\ttrain_h0_mean_x_min_u: 0.00236209016293\n", 2262 | "\ttrain_h0_min_x_max_u: 0.585445165634\n", 2263 | "\ttrain_h0_min_x_mean_u: 0.00744519708678\n", 2264 | "\ttrain_h0_min_x_min_u: 0.0\n", 2265 | "\ttrain_h0_range_x_max_u: 5.16996622086\n", 2266 | "\ttrain_h0_range_x_mean_u: 2.0620663166\n", 2267 | "\ttrain_h0_range_x_min_u: 0.111943788826\n", 2268 | "\ttrain_h0_row_norms_max: 0.557518601418\n", 2269 | "\ttrain_h0_row_norms_mean: 0.384280800819\n", 2270 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 2271 | "\ttrain_h1_col_norms_max: 1.3490973711\n", 2272 | "\ttrain_h1_col_norms_mean: 1.07836294174\n", 2273 | "\ttrain_h1_col_norms_min: 0.896570086479\n", 2274 | "\ttrain_h1_max_x_max_u: 9.24939441681\n", 2275 | "\ttrain_h1_max_x_mean_u: 3.07910609245\n", 2276 | "\ttrain_h1_max_x_min_u: 0.0\n", 2277 | "\ttrain_h1_mean_x_max_u: 2.98392343521\n", 2278 | "\ttrain_h1_mean_x_mean_u: 0.956997871399\n", 2279 | "\ttrain_h1_mean_x_min_u: 0.0\n", 2280 | "\ttrain_h1_min_x_max_u: 0.0178324989974\n", 2281 | "\ttrain_h1_min_x_mean_u: 0.000178324989974\n", 2282 | "\ttrain_h1_min_x_min_u: 0.0\n", 2283 | "\ttrain_h1_range_x_max_u: 9.24939441681\n", 2284 | "\ttrain_h1_range_x_mean_u: 3.07892799377\n", 2285 | "\ttrain_h1_range_x_min_u: 0.0\n", 2286 | "\ttrain_h1_row_norms_max: 1.36258947849\n", 2287 | "\ttrain_h1_row_norms_mean: 1.07979047298\n", 2288 | "\ttrain_h1_row_norms_min: 0.8949213624\n", 2289 | "\ttrain_objective: 0.140654325485\n", 2290 | "\ttrain_y_col_norms_max: 1.47577524185\n", 2291 | "\ttrain_y_col_norms_mean: 1.29343950748\n", 2292 | "\ttrain_y_col_norms_min: 1.08780992031\n", 2293 | "\ttrain_y_max_max_class: 0.999949634075\n", 2294 | "\ttrain_y_mean_max_class: 0.912002623081\n", 2295 | "\ttrain_y_min_max_class: 0.359933674335\n", 2296 | "\ttrain_y_misclass: 0.0299999993294\n", 2297 | "\ttrain_y_nll: 0.140654325485\n", 2298 | "\ttrain_y_row_norms_max: 0.838829219341\n", 2299 | "\ttrain_y_row_norms_mean: 0.339229732752\n", 2300 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 2301 | "\ttraining_seconds_this_epoch: 0.0767010003328\n", 2302 | "Time this epoch: 0.076711 seconds\n", 2303 | "Monitoring step:\n", 2304 | "\tEpochs seen: 19\n", 2305 | "\tBatches seen: 190\n", 2306 | "\tExamples seen: 19000\n", 2307 | "\tlearning_rate: 0.099999986589\n", 2308 | "\tmomentum: 0.500000059605\n", 2309 | "\ttest_h0_col_norms_max: 1.37143599987\n", 2310 | "\ttest_h0_col_norms_mean: 1.08365881443\n", 2311 | "\ttest_h0_col_norms_min: 0.971715211868\n", 2312 | "\ttest_h0_max_x_max_u: 5.36840486526\n", 2313 | "\ttest_h0_max_x_mean_u: 2.12656331062\n", 2314 | "\ttest_h0_max_x_min_u: 0.123854331672\n", 2315 | "\ttest_h0_mean_x_max_u: 2.40991473198\n", 2316 | "\ttest_h0_mean_x_mean_u: 0.702306210995\n", 2317 | "\ttest_h0_mean_x_min_u: 0.00299808918498\n", 2318 | "\ttest_h0_min_x_max_u: 0.534740746021\n", 2319 | "\ttest_h0_min_x_mean_u: 0.00993497017771\n", 2320 | "\ttest_h0_min_x_min_u: 0.0\n", 2321 | "\ttest_h0_range_x_max_u: 5.35095453262\n", 2322 | "\ttest_h0_range_x_mean_u: 2.11662817001\n", 2323 | "\ttest_h0_range_x_min_u: 0.123854331672\n", 2324 | "\ttest_h0_row_norms_max: 0.564710319042\n", 2325 | "\ttest_h0_row_norms_mean: 0.385214805603\n", 2326 | "\ttest_h0_row_norms_min: 0.280473798513\n", 2327 | "\ttest_h1_col_norms_max: 1.36141693592\n", 2328 | "\ttest_h1_col_norms_mean: 1.0807813406\n", 2329 | "\ttest_h1_col_norms_min: 0.896570086479\n", 2330 | "\ttest_h1_max_x_max_u: 9.21926689148\n", 2331 | "\ttest_h1_max_x_mean_u: 3.15925455093\n", 2332 | "\ttest_h1_max_x_min_u: 0.0\n", 2333 | "\ttest_h1_mean_x_max_u: 2.91600131989\n", 2334 | "\ttest_h1_mean_x_mean_u: 0.975942790508\n", 2335 | "\ttest_h1_mean_x_min_u: 0.0\n", 2336 | "\ttest_h1_min_x_max_u: 0.04676309973\n", 2337 | "\ttest_h1_min_x_mean_u: 0.00052209867863\n", 2338 | "\ttest_h1_min_x_min_u: 0.0\n", 2339 | "\ttest_h1_range_x_max_u: 9.21926689148\n", 2340 | "\ttest_h1_range_x_mean_u: 3.15873241425\n", 2341 | "\ttest_h1_range_x_min_u: 0.0\n", 2342 | "\ttest_h1_row_norms_max: 1.37135279179\n", 2343 | "\ttest_h1_row_norms_mean: 1.08229625225\n", 2344 | "\ttest_h1_row_norms_min: 0.895482897758\n", 2345 | "\ttest_objective: 0.460972487926\n", 2346 | "\ttest_y_col_norms_max: 1.49516808987\n", 2347 | "\ttest_y_col_norms_mean: 1.31553649902\n", 2348 | "\ttest_y_col_norms_min: 1.11032485962\n", 2349 | "\ttest_y_max_max_class: 0.999933898449\n", 2350 | "\ttest_y_mean_max_class: 0.881941497326\n", 2351 | "\ttest_y_min_max_class: 0.352427333593\n", 2352 | "\ttest_y_misclass: 0.123000003397\n", 2353 | "\ttest_y_nll: 0.460972487926\n", 2354 | "\ttest_y_row_norms_max: 0.848902404308\n", 2355 | "\ttest_y_row_norms_mean: 0.34513476491\n", 2356 | "\ttest_y_row_norms_min: 0.00266337138601\n", 2357 | "\ttotal_seconds_last_epoch: 0.3914090693\n", 2358 | "\ttrain_h0_col_norms_max: 1.37143599987\n", 2359 | "\ttrain_h0_col_norms_mean: 1.08365881443\n", 2360 | "\ttrain_h0_col_norms_min: 0.971715211868\n", 2361 | "\ttrain_h0_max_x_max_u: 5.1641201973\n", 2362 | "\ttrain_h0_max_x_mean_u: 2.07583594322\n", 2363 | "\ttrain_h0_max_x_min_u: 0.116156749427\n", 2364 | "\ttrain_h0_mean_x_max_u: 2.33537507057\n", 2365 | "\ttrain_h0_mean_x_mean_u: 0.684384226799\n", 2366 | "\ttrain_h0_mean_x_min_u: 0.00248867250048\n", 2367 | "\ttrain_h0_min_x_max_u: 0.56894415617\n", 2368 | "\ttrain_h0_min_x_mean_u: 0.00741918431595\n", 2369 | "\ttrain_h0_min_x_min_u: 0.0\n", 2370 | "\ttrain_h0_range_x_max_u: 5.15645742416\n", 2371 | "\ttrain_h0_range_x_mean_u: 2.06841659546\n", 2372 | "\ttrain_h0_range_x_min_u: 0.116156749427\n", 2373 | "\ttrain_h0_row_norms_max: 0.564710319042\n", 2374 | "\ttrain_h0_row_norms_mean: 0.385214805603\n", 2375 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 2376 | "\ttrain_h1_col_norms_max: 1.36141693592\n", 2377 | "\ttrain_h1_col_norms_mean: 1.0807813406\n", 2378 | "\ttrain_h1_col_norms_min: 0.896570086479\n", 2379 | "\ttrain_h1_max_x_max_u: 9.14778137207\n", 2380 | "\ttrain_h1_max_x_mean_u: 3.10251712799\n", 2381 | "\ttrain_h1_max_x_min_u: 0.0\n", 2382 | "\ttrain_h1_mean_x_max_u: 2.88491630554\n", 2383 | "\ttrain_h1_mean_x_mean_u: 0.956087648869\n", 2384 | "\ttrain_h1_mean_x_min_u: 0.0\n", 2385 | "\ttrain_h1_min_x_max_u: 0.00415239948779\n", 2386 | "\ttrain_h1_min_x_mean_u: 4.15239956055e-05\n", 2387 | "\ttrain_h1_min_x_min_u: 0.0\n", 2388 | "\ttrain_h1_range_x_max_u: 9.14778137207\n", 2389 | "\ttrain_h1_range_x_mean_u: 3.10247564316\n", 2390 | "\ttrain_h1_range_x_min_u: 0.0\n", 2391 | "\ttrain_h1_row_norms_max: 1.37135279179\n", 2392 | "\ttrain_h1_row_norms_mean: 1.08229625225\n", 2393 | "\ttrain_h1_row_norms_min: 0.895482897758\n", 2394 | "\ttrain_objective: 0.11272007972\n", 2395 | "\ttrain_y_col_norms_max: 1.49516808987\n", 2396 | "\ttrain_y_col_norms_mean: 1.31553649902\n", 2397 | "\ttrain_y_col_norms_min: 1.11032485962\n", 2398 | "\ttrain_y_max_max_class: 0.99991953373\n", 2399 | "\ttrain_y_mean_max_class: 0.922433435917\n", 2400 | "\ttrain_y_min_max_class: 0.395698159933\n", 2401 | "\ttrain_y_misclass: 0.0210000015795\n", 2402 | "\ttrain_y_nll: 0.11272007972\n", 2403 | "\ttrain_y_row_norms_max: 0.848902404308\n", 2404 | "\ttrain_y_row_norms_mean: 0.34513476491\n", 2405 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 2406 | "\ttraining_seconds_this_epoch: 0.076710999012\n", 2407 | "Time this epoch: 0.077043 seconds\n", 2408 | "Monitoring step:\n", 2409 | "\tEpochs seen: 20\n", 2410 | "\tBatches seen: 200\n", 2411 | "\tExamples seen: 20000\n", 2412 | "\tlearning_rate: 0.099999986589\n", 2413 | "\tmomentum: 0.500000059605\n", 2414 | "\ttest_h0_col_norms_max: 1.38282668591\n", 2415 | "\ttest_h0_col_norms_mean: 1.08640444279\n", 2416 | "\ttest_h0_col_norms_min: 0.971933662891\n", 2417 | "\ttest_h0_max_x_max_u: 5.3245806694\n", 2418 | "\ttest_h0_max_x_mean_u: 2.14106059074\n", 2419 | "\ttest_h0_max_x_min_u: 0.127138972282\n", 2420 | "\ttest_h0_mean_x_max_u: 2.38005948067\n", 2421 | "\ttest_h0_mean_x_mean_u: 0.708713293076\n", 2422 | "\ttest_h0_mean_x_min_u: 0.00312190386467\n", 2423 | "\ttest_h0_min_x_max_u: 0.554714679718\n", 2424 | "\ttest_h0_min_x_mean_u: 0.00989470258355\n", 2425 | "\ttest_h0_min_x_min_u: 0.0\n", 2426 | "\ttest_h0_range_x_max_u: 5.30910110474\n", 2427 | "\ttest_h0_range_x_mean_u: 2.13116574287\n", 2428 | "\ttest_h0_range_x_min_u: 0.127138972282\n", 2429 | "\ttest_h0_row_norms_max: 0.56689542532\n", 2430 | "\ttest_h0_row_norms_mean: 0.386148780584\n", 2431 | "\ttest_h0_row_norms_min: 0.280473798513\n", 2432 | "\ttest_h1_col_norms_max: 1.36655056477\n", 2433 | "\ttest_h1_col_norms_mean: 1.08320844173\n", 2434 | "\ttest_h1_col_norms_min: 0.896570086479\n", 2435 | "\ttest_h1_max_x_max_u: 9.48558425903\n", 2436 | "\ttest_h1_max_x_mean_u: 3.20883893967\n", 2437 | "\ttest_h1_max_x_min_u: 0.0\n", 2438 | "\ttest_h1_mean_x_max_u: 3.01143312454\n", 2439 | "\ttest_h1_mean_x_mean_u: 0.997267007828\n", 2440 | "\ttest_h1_mean_x_min_u: 0.0\n", 2441 | "\ttest_h1_min_x_max_u: 0.0436610169709\n", 2442 | "\ttest_h1_min_x_mean_u: 0.000580249004997\n", 2443 | "\ttest_h1_min_x_min_u: 0.0\n", 2444 | "\ttest_h1_range_x_max_u: 9.48558425903\n", 2445 | "\ttest_h1_range_x_mean_u: 3.20825862885\n", 2446 | "\ttest_h1_range_x_min_u: 0.0\n", 2447 | "\ttest_h1_row_norms_max: 1.38067495823\n", 2448 | "\ttest_h1_row_norms_mean: 1.08479678631\n", 2449 | "\ttest_h1_row_norms_min: 0.895965099335\n", 2450 | "\ttest_objective: 0.472773820162\n", 2451 | "\ttest_y_col_norms_max: 1.52235996723\n", 2452 | "\ttest_y_col_norms_mean: 1.33737313747\n", 2453 | "\ttest_y_col_norms_min: 1.1340098381\n", 2454 | "\ttest_y_max_max_class: 0.999922096729\n", 2455 | "\ttest_y_mean_max_class: 0.884098291397\n", 2456 | "\ttest_y_min_max_class: 0.361968785524\n", 2457 | "\ttest_y_misclass: 0.126000002027\n", 2458 | "\ttest_y_nll: 0.472773820162\n", 2459 | "\ttest_y_row_norms_max: 0.863828599453\n", 2460 | "\ttest_y_row_norms_mean: 0.350933045149\n", 2461 | "\ttest_y_row_norms_min: 0.00266337138601\n", 2462 | "\ttotal_seconds_last_epoch: 0.393661975861\n", 2463 | "\ttrain_h0_col_norms_max: 1.38282668591\n", 2464 | "\ttrain_h0_col_norms_mean: 1.08640444279\n", 2465 | "\ttrain_h0_col_norms_min: 0.971933662891\n", 2466 | "\ttrain_h0_max_x_max_u: 5.14531040192\n", 2467 | "\ttrain_h0_max_x_mean_u: 2.08976364136\n", 2468 | "\ttrain_h0_max_x_min_u: 0.121576607227\n", 2469 | "\ttrain_h0_mean_x_max_u: 2.3057539463\n", 2470 | "\ttrain_h0_mean_x_mean_u: 0.690565824509\n", 2471 | "\ttrain_h0_mean_x_min_u: 0.00260553648695\n", 2472 | "\ttrain_h0_min_x_max_u: 0.593227982521\n", 2473 | "\ttrain_h0_min_x_mean_u: 0.00747096119449\n", 2474 | "\ttrain_h0_min_x_min_u: 0.0\n", 2475 | "\ttrain_h0_range_x_max_u: 5.14017391205\n", 2476 | "\ttrain_h0_range_x_mean_u: 2.08229255676\n", 2477 | "\ttrain_h0_range_x_min_u: 0.121576607227\n", 2478 | "\ttrain_h0_row_norms_max: 0.56689542532\n", 2479 | "\ttrain_h0_row_norms_mean: 0.386148780584\n", 2480 | "\ttrain_h0_row_norms_min: 0.280473798513\n", 2481 | "\ttrain_h1_col_norms_max: 1.36655056477\n", 2482 | "\ttrain_h1_col_norms_mean: 1.08320844173\n", 2483 | "\ttrain_h1_col_norms_min: 0.896570086479\n", 2484 | "\ttrain_h1_max_x_max_u: 9.36334991455\n", 2485 | "\ttrain_h1_max_x_mean_u: 3.14843082428\n", 2486 | "\ttrain_h1_max_x_min_u: 0.0\n", 2487 | "\ttrain_h1_mean_x_max_u: 2.98397874832\n", 2488 | "\ttrain_h1_mean_x_mean_u: 0.977112054825\n", 2489 | "\ttrain_h1_mean_x_min_u: 0.0\n", 2490 | "\ttrain_h1_min_x_max_u: 0.00650532450527\n", 2491 | "\ttrain_h1_min_x_mean_u: 6.50532456348e-05\n", 2492 | "\ttrain_h1_min_x_min_u: 0.0\n", 2493 | "\ttrain_h1_range_x_max_u: 9.36334991455\n", 2494 | "\ttrain_h1_range_x_mean_u: 3.14836573601\n", 2495 | "\ttrain_h1_range_x_min_u: 0.0\n", 2496 | "\ttrain_h1_row_norms_max: 1.38067495823\n", 2497 | "\ttrain_h1_row_norms_mean: 1.08479678631\n", 2498 | "\ttrain_h1_row_norms_min: 0.895965099335\n", 2499 | "\ttrain_objective: 0.098839096725\n", 2500 | "\ttrain_y_col_norms_max: 1.52235996723\n", 2501 | "\ttrain_y_col_norms_mean: 1.33737313747\n", 2502 | "\ttrain_y_col_norms_min: 1.1340098381\n", 2503 | "\ttrain_y_max_max_class: 0.999954998493\n", 2504 | "\ttrain_y_mean_max_class: 0.927945494652\n", 2505 | "\ttrain_y_min_max_class: 0.41922056675\n", 2506 | "\ttrain_y_misclass: 0.01600000076\n", 2507 | "\ttrain_y_nll: 0.098839096725\n", 2508 | "\ttrain_y_row_norms_max: 0.863828599453\n", 2509 | "\ttrain_y_row_norms_mean: 0.350933045149\n", 2510 | "\ttrain_y_row_norms_min: 0.00266337138601\n", 2511 | "\ttraining_seconds_this_epoch: 0.0770429968834\n", 2512 | "Saving to ./model_mnist.pkl...\n", 2513 | "Saving to ./model_mnist.pkl done. Time elapsed: 0.099798 seconds\n", 2514 | "Saving to ./model_mnist.pkl...\n", 2515 | "Saving to ./model_mnist.pkl done. Time elapsed: 0.095648 seconds\n" 2516 | ] 2517 | } 2518 | ], 2519 | "source": [ 2520 | "# Run training loop.\n", 2521 | "train.main_loop()" 2522 | ] 2523 | }, 2524 | { 2525 | "cell_type": "code", 2526 | "execution_count": 16, 2527 | "metadata": { 2528 | "collapsed": false 2529 | }, 2530 | "outputs": [ 2531 | { 2532 | "name": "stdout", 2533 | "output_type": "stream", 2534 | "text": [ 2535 | "Train objective = 0.016000\n", 2536 | "Valid objective = 0.126000\n" 2537 | ] 2538 | } 2539 | ], 2540 | "source": [ 2541 | "# Load saved model.\n", 2542 | "model = pickle.load(open(filename_model, 'r'))\n", 2543 | "\n", 2544 | "# Print objective function after training.\n", 2545 | "channels = model.monitor.channels\n", 2546 | "print 'Train objective = %f' % channels['train_y_misclass'].val_record[-1]\n", 2547 | "print 'Valid objective = %f' % channels['test_y_misclass'].val_record[-1]" 2548 | ] 2549 | }, 2550 | { 2551 | "cell_type": "code", 2552 | "execution_count": null, 2553 | "metadata": { 2554 | "collapsed": true 2555 | }, 2556 | "outputs": [], 2557 | "source": [] 2558 | } 2559 | ], 2560 | "metadata": { 2561 | "kernelspec": { 2562 | "display_name": "Python 2", 2563 | "language": "python", 2564 | "name": "python2" 2565 | }, 2566 | "language_info": { 2567 | "codemirror_mode": { 2568 | "name": "ipython", 2569 | "version": 2 2570 | }, 2571 | "file_extension": ".py", 2572 | "mimetype": "text/x-python", 2573 | "name": "python", 2574 | "nbconvert_exporter": "python", 2575 | "pygments_lexer": "ipython2", 2576 | "version": "2.7.8" 2577 | } 2578 | }, 2579 | "nbformat": 4, 2580 | "nbformat_minor": 0 2581 | } 2582 | -------------------------------------------------------------------------------- /lhc_dl_tutorial.pdf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/peterjsadowski/lhc2015-dl-tutorial/f1a903ad209c3b912b49d01044177f98fd109ff1/lhc_dl_tutorial.pdf -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- 1 | Theano 2 | 3 | --------------------------------------------------------------------------------