update Qwen/Qwen2.5-1.5B-Instruct to deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B, i got an error
#19
by
fengkai-llm
- opened
Error:
labels[torch.where(inputs["labels"] != self.tokenizer.pad_token_id)] = -100
[rank0]: IndexError: index 441 is out of bounds for dimension 1 with size 441
Did using the Instruct model cause this?