tgc1997 / RMN

IJCAI2020: Learning to Discretely Compose Reasoning Module Networks for Video Captioning
79 stars 12 forks source link

Would I ask one question? #26

Open maybekssss opened 2 years ago

maybekssss commented 2 years ago

File "/RMN-master/models/allennlp_beamsearch.py", line 257, in search state_tensor.reshape(batch_size, self.beam_size, *last_dims) RuntimeError: gather_out_cuda(): Expected dtype int64 for index