search

albert-jin / Rce-KGQA

A novel pipeline framework for multi-hop complex KGQA task. It aims at Improving Multi-hop Embedded Knowledge Graph Question Answering by Introducing Relational Chain Reasoning
MIT License
45 stars 11 forks source link

answering-filtering中出现问题 #6

Open justyyau opened 2 years ago

justyyau commented 2 years ago

torch.tensor(tails_idxs),出现维度对应不上,tails_idxs列表的内容长短不一，这里不好直接转换成tensor,感觉作者的代码有问题

justyyau commented 2 years ago

File "E:\yaojunfei\QAandKR\Rce-KGQA-main\answer_filtering_module\dataloader.py", line 247, in _collate_fn new_heads_idx, torch.tensor(tails_idxs), max_sent_len ValueError: expected sequence of length 68 at dim 1 (got 17) 0%| | 0/1441 [01:49<?, ? batches/s]

ruomusu commented 10 months ago

文件“E:\yaojunfei\QAandKR\Rce-KGQA-main\answer_filtering_module\dataloader.py”，第 247 行，在 _collat​​e_fn new_heads_idx、tor​​ch.tensor(tails_idxs)、max_sent_len ValueError：在暗淡 1 处预期长度为 68 的序列（得到 17） ) 0%| | 0/1441 [01:49<?, ? 批次/秒] 你好，请问你运行出来了吗