New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Gradient flowing between inference_cell and generator_cell #42

Open

Rustleman opened this issue Dec 18, 2019 · 0 comments

Rustleman commented Dec 18, 2019

There seems to be an issue in the inference_rnn function where the inference_cell and generator_cell are connected together:

tf-gqn/gqn/gqn_draw.py

Lines 417 to 421 in bc84f24

    
           # estimate statistics and sample state from posterior 
        
           mu_q, sigma_q, z_q = compute_eta_and_sample_z(inf_state.lstm.h, 
        
                                                         scope="Sample_eta_q") 
        
           # input into generator RNN 
        
           gen_input = _GeneratorCellInput(representations, query_poses, z_q)

It looks like the gradient flows through z_q.

Adding the line z_q = tf.stop_gradient(z_q)
seems to improve the results when just the generator_rnn is used during testing.

The text was updated successfully, but these errors were encountered:

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment