Sean Robertson comments

Results 24 comments of


                                            Sean Robertson

gradvars issue

As a quick fix, you can try surrounding each with a list constructor: ``` python gradvars = list(inf_gradvars) + list(gen_gradvars) + list(rebar_optimizer.variance_gradvars) ``` `tuple()` would probably work too.

tf.nn.ctc_beam_search_decoder does not pick path with highest probability at next time step

Hello, While I can appreciate that you lot are very busy, it's been a month since I opened this issue and I'm afraid it's going to fall through the cracks....

tf.nn.ctc_beam_search_decoder does not pick path with highest probability at next time step

Hi @yjiangling, I'm fairly certain that I haven't left out any probability mass because I'm relying on the op itself to calculate the probability of all but the pruned path....

tf.nn.ctc_beam_search_decoder does not pick path with highest probability at next time step

Hi Venkat, As you can see, I made this bug report 4 years ago, and received no response from @ebrevdo. While I understand that the devs are under no obligation...