search

ayumiymk / aster.pytorch

ASTER in Pytorch
MIT License
663 stars 169 forks source link

ValueError: too many values to unpack (expected 3) #13

Closed zhijieL387 closed 4 years ago

zhijieL387 commented 5 years ago

~/aster.pytorch/lib/models/model_builder.py in forward(self, input_dict) 89 return_dict['losses']['loss_rec'] = loss_rec 90 else: ---> 91 rec_pred, rec_pred_scores = self.decoder.beam_search(encoder_feats, global_args.beam_width, self.eos) 92 recpred = self.decoder([encoder_feats, rec_targets, rec_lengths]) 93 loss_rec = self.rec_crit(recpred, rec_targets, rec_lengths)

~/aster.pytorch/lib/models/attention_recognition_head.py in beam_search(self, x, beam_width, eos) 74 75 # https://github.com/IBM/pytorch-seq2seq/blob/fede87655ddce6c94b38886089e05321dc9802af/seq2seq/models/TopKDecoder.py ---> 76 batch_size, l, d = x.size() 77 # inflated_encoder_feats = _inflate(encoder_feats, beam_width, 0) # ABC --> AABBCC -/-> ABCABC 78 inflated_encoder_feats = x.unsqueeze(1).permute((1,0,2,3)).repeat((beam_width,1,1,1)).permute((1,0,2,3)).contiguous().view(-1, l, d)

ValueError: too many values to unpack (expected 3)

ayumiymk commented 5 years ago

I have not met such a problem but someone else has. I will reopen the related issue. Please refer to #6 to look for help.

luoqingyu commented 4 years ago

Do you solve this problem?

luoqingyu commented 4 years ago

I have find the reason， parameter in config.py is different with scripts\stn_att_rec.sh. Do not run main.py directly, run stn_att_rec.sh as readme show。