Geek_Huang

icon indicating an email [email protected]

icon company Nanjing University icon location SuZhou, China icon location I am a undergraduate student at Nanjing University, interested in Multi-Modality.

Results 3 comments of Geek_Huang

Question about the implementation

So why just need the last token instead of 2 or more? since there are still lot of token in the last row have not been predicted. And I think...

Question about the implementation

Thank you for you kindly response! Still one question: Code: # forward input preparation new_input_ids.append(input_ids[:, idx_in_input_ids].unsqueeze(-1)) new_position_ids.append(global_idx) local_position_ids.append(idx_in_input_ids) new_input_ids = torch.cat(new_input_ids, dim=1) new_position_ids = torch.tensor(new_position_ids, device=input_ids.device).unsqueeze_(0) local_position_ids = torch.tensor(local_position_ids, device=input_ids.device).unsqueeze_(0)...

dataloader的batch_size设置

请问您后面有尝试吗，bs不为1能正常运行吗