Please let us know which model this issue is about (specify the top-level directory)
textsum
when decode mode and set --beam_size=1, happen "ValueError: Cannot feed value of shape (512, 0) for Tensor u'seq2se1/encoder0/BiRNN_FW/cond_39/Merge_1:0', while has shape '(1, 512)' "
I don't know where the “512” has set and why happened this error.
Addition, when decode mode and set beam_size more than 1, it spent long time(about 15 hours) and I have no much time to wait.Is it normal?
We recommend posting on Stack Overflow for community support. If you're sure there's a bug in the model, please open a new issue with the specific commands you're using so we can reproduce the error.
Please let us know which model this issue is about (specify the top-level directory)
textsum
when decode mode and set --beam_size=1, happen "ValueError: Cannot feed value of shape (512, 0) for Tensor u'seq2se1/encoder0/BiRNN_FW/cond_39/Merge_1:0', while has shape '(1, 512)' " I don't know where the “512” has set and why happened this error. Addition, when decode mode and set beam_size more than 1, it spent long time(about 15 hours) and I have no much time to wait.Is it normal?