Output for a DBN with 256-neuron RBN layers

The output of the preceding listing shows the test set's accuracy:

Reconstruction loss: 0.156712: 100%|██████████| 5/5 [00:49<00:00, 9.99s/it]Start deep belief net finetuning...Tensorboard logs dir for this run is /home/ubuntu/.yadlt/logs/run53Accuracy: 0.0868: 100%|██████████| 1/1 [00:04<00:00, 4.09s/it]Test set accuracy: 0.0868000015616

Overall accuracy and Test set accuracy is quite low. With the increase in number of iterations it improves. Let us run same sample with 20 epochs

Reconstruction loss: 0.120337: 100%|██████████| 20/20 [03:07<00:00, 8.79s/it]Start deep belief net finetuning...Tensorboard logs dir for this run is /home/ubuntu/.yadlt/logs/run80Accuracy: 0.105: 100%|██████████| ...

Get Neural Network Programming with TensorFlow now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.