Timon0327

Hi. When I run the test.py to evaluate, the model just generates without any other token. Is there same problem?

> Thanks for your reply. But I load the 'vocab.pkl' file from the original git which shows that the 0 is 'unk'. And when I run the test.py, the model...

> Hi, because I had a bug about gather_cuda() from beam_search.py, I modified the data.dtype to float.64 to solve it. When I changed it to original version, the test was...

> Hi, > > Thanks for providing the implementation. > > I checked the paper and observe that only performance with reinforce optimization is reported. I wonder if you can...