results for a small DBN experiment

python code/DBN_small.py
Downloading data from http://www.iro.umontreal.ca/~lisa/deep/data/mnist/mnist.pkl.gz
... loading data
... building the model
... getting the pretraining functions
... pre-training the model
Pre-training layer 0, epoch 0, cost  -98.605633351
Pre-training layer 0, epoch 1, cost  -83.821740962
Pre-training layer 0, epoch 2, cost  -80.7250660333
Pre-training layer 0, epoch 3, cost  -79.0545566378
Pre-training layer 0, epoch 4, cost  -77.9373883434
Pre-training layer 0, epoch 5, cost  -77.0672617796
Pre-training layer 0, epoch 6, cost  -76.4264764766
Pre-training layer 0, epoch 7, cost  -75.8230576646
Pre-training layer 0, epoch 8, cost  -75.3795806083
Pre-training layer 0, epoch 9, cost  -74.9426026512
Pre-training layer 1, epoch 0, cost  -259.323603643
Pre-training layer 1, epoch 1, cost  -234.892072755
Pre-training layer 1, epoch 2, cost  -229.839405554
Pre-training layer 1, epoch 3, cost  -227.188812758
Pre-training layer 1, epoch 4, cost  -225.417284897
Pre-training layer 1, epoch 5, cost  -224.139577706
Pre-training layer 1, epoch 6, cost  -223.164434731
Pre-training layer 1, epoch 7, cost  -222.394942526
Pre-training layer 1, epoch 8, cost  -221.768370618
Pre-training layer 1, epoch 9, cost  -221.278308113
Pre-training layer 2, epoch 0, cost  -76.0649870035
Pre-training layer 2, epoch 1, cost  -64.5806778821
Pre-training layer 2, epoch 2, cost  -62.436519382
Pre-training layer 2, epoch 3, cost  -61.3510303461
Pre-training layer 2, epoch 4, cost  -60.6772809506
Pre-training layer 2, epoch 5, cost  -60.2360054935
Pre-training layer 2, epoch 6, cost  -59.810636797
Pre-training layer 2, epoch 7, cost  -59.5407355314
Pre-training layer 2, epoch 8, cost  -59.3057561615
Pre-training layer 2, epoch 9, cost  -59.0920642013
The pretraining code for file DBN_small.py ran for 79.22m
... getting the finetuning functions
... finetunning the model
epoch 1, minibatch 5000/5000, validation error 3.820000 %
     epoch 1, minibatch 5000/5000, test error of best model 4.390000 %
epoch 2, minibatch 5000/5000, validation error 3.070000 %
     epoch 2, minibatch 5000/5000, test error of best model 3.510000 %
epoch 3, minibatch 5000/5000, validation error 2.710000 %
     epoch 3, minibatch 5000/5000, test error of best model 3.010000 %
epoch 4, minibatch 5000/5000, validation error 2.460000 %
     epoch 4, minibatch 5000/5000, test error of best model 2.640000 %
epoch 5, minibatch 5000/5000, validation error 2.200000 %
     epoch 5, minibatch 5000/5000, test error of best model 2.450000 %
epoch 6, minibatch 5000/5000, validation error 2.130000 %
     epoch 6, minibatch 5000/5000, test error of best model 2.230000 %
epoch 7, minibatch 5000/5000, validation error 2.050000 %
     epoch 7, minibatch 5000/5000, test error of best model 2.120000 %
epoch 8, minibatch 5000/5000, validation error 1.980000 %
     epoch 8, minibatch 5000/5000, test error of best model 2.060000 %
epoch 9, minibatch 5000/5000, validation error 2.000000 %
epoch 10, minibatch 5000/5000, validation error 1.980000 %
     epoch 10, minibatch 5000/5000, test error of best model 1.890000 %
Optimization complete with best validation score of 1.980000 %,with test performance 1.890000 %
The fine tuning code for file DBN_small.py ran for 32.26m

Note: A full DBN experiment will give result like:

Optimization complete with best validation score of 1.280000 %,with test performance 1.290000 %
The fine tuning code for file DBN.py ran for 648.83m