search

pumpikano / tf-dann

Domain-Adversarial Neural Network in Tensorflow
MIT License
628 stars 224 forks source link

Adapted feature extraction #7

Closed saadirtza closed 7 years ago

saadirtza commented 7 years ago

Hi Pumpikano,

With your help, i am able to get the prediction values, (thanks)..

but the performance is very poor. so i thought to look at the adapted features on my database. The features are all zeros. source_acc, target_acc, d_acc, dann_emb = train_and_evaluate('dann', graph, model)

in the above line 'dann_emb' is the adapted features and they are all zeros. There is no error or warning in the code.

Can you please suggest where the problem might be? Thanks alot

Best Regards Saad

saadirtza commented 7 years ago

After some iteration of DANN training, the values of batch loss is 0.61 but dloss and ploss are NAN..

Is this because i am not feeding the data properly or network configuration is not good?

Thanks

Regards

saadirtza commented 7 years ago

That was the issue with selection of initial learning paramter i.e 'lr' in the code.. Can you please suggest any tutorial on how to select this parameter for good performance?

Thanks & Regards

pumpikano commented 7 years ago

These lecture slides are a "classic" reference for intuitions and tips for setting learning rates: http://www.cs.toronto.edu/~tijmen/csc321/slides/lecture_slides_lec6.pdf