list index out of range

[ShenRuiLi]

02/21/2022 15:34:22 - INFO - main - Running training 02/21/2022 15:34:22 - INFO - main - Num post-response pairs = 27550 02/21/2022 15:34:22 - INFO - main - Batch size = 8 02/21/2022 15:34:22 - INFO - main - Num steps = 10331 Epoch: 0%| | 0/3 [00:00<?, ?it/s] Traceback (most recent call last): File "C:/Users/25687/Desktop/NLP_Paper/KdConv-master/benchmark/bertret/run_BERTRetrieval.py", line 384, in main() File "C:/Users/25687/Desktop/NLP_Paper/KdConv-master/benchmark/bertret/run_BERTRetrieval.py", line 255, in main data = dataManager.get_next_batch(key='train') File "D:\Anaconda\envs\KdConv-master\lib\site-packages\cotk\dataloader\dataloader.py", line 195, in get_next_batch res = self.get_batch(key, index) File "C:\Users\25687\Desktop\NLP_Paper\KdConv-master\benchmark\bertret\myCoTK\dataloader\bert_dataloader.py", line 163, in get_batch resp_distractors_bert = self.data[key]['resp_distractors_bert'][idx] IndexError: list index out of range train set restart, 3443 batches and 6 left 在windows上跑bertret代码时出现了这个问题，一直说索引超出界限，请问怎么解决呢？

[chujiezheng]

可以先打印一下len(self.data[key]['resp_distractors_bert'])和idx看看

[ShenRuiLi]

[chujiezheng]

idx理应不会超。你改了其他地方的代码了吗？

[ShenRuiLi]

其他地方的代码没有改动。我看idx的值都很大，是一些文件的原因导致的吗？

[chujiezheng]

我这儿没有同样的环境，不太好复现。可以把distractor文件删掉，重新跑一下看看

[chujiezheng]

python版本如果很高（比如3.8、3.9），也可以替换成3.5或3.6试试

[ShenRuiLi]

好的，谢谢您

[ShenRuiLi]

不好意思，是我改了一些代码导致的问题，抱歉打扰你了