Kim-Dongjun/Soft-Truncation

Issues

• FFHQ 256

  #5 opened 2 years ago by georgestein
  1

• What is the NLL results with uniform dequantization?

  #4 opened 2 years ago by zhengkw18
  1

• Why `likelihood_residual_fn_lossless_compression` doesn't minus p_entropy as `likelihood_residual_fn`?

  #2 opened 2 years ago by baofff
  4

• When will the pretrained model be released?

  #1 opened 2 years ago by baofff
  2

• Issue on computing log-likelihood (bpd)

  #3 opened 2 years ago by Kim-Dongjun
  1