[BUG] <title>RuntimeError: Expected index [1728, 4096] to be smaller than self [2048, 4096] apart from dimension 0 and to be smal ler size than src [864, 4096]

[laoShuaiGe]

是否已有关于该错误的issue或讨论？ | Is there an existing issue / discussion for this?

• [X] 我已经搜索过已有的issues和讨论 | I have searched the existing issues / discussions

该问题是否在FAQ中有解答？ | Is there an existing answer for this in FAQ?

• [X] 我已经搜索过FAQ | I have searched FAQ

当前行为 | Current Behavior

Previous issue: #345 by aoji0606 I saw the same issue remain unresolved, but was closed for some reason. So reopen it for help. Thank you in advance for any suggestions.

期望行为 | Expected Behavior

Training pipeline crashes due to the issue mentioned above.

复现方法 | Steps To Reproduce

No response

运行环境 | Environment

- OS: 20.04
- Python: 3.10
- Transformers: 4.40
- PyTorch: 2.3
- CUDA (`python -c 'import torch; print(torch.version.cuda)'`):

备注 | Anything else?

No response

[LDLINGLINGLING]

hello，check if there is a problem with your max_model_length and data construction

[davidetalon]

I am encountering the same problem, what solved the issue?