pixel_values = torch.cat([pv for pv in all_pixel_values if pv is not None], dim=0) if any(all_pixel_values) else None
RuntimeError: Boolean value of Tensor with more than one value is ambiguous
such error exist while I am using the finetune_lora_vision script
Does your dataset has mixed modality?
I just updated the code for the issue, so you can use it. However I'm not sure this code can handle mixed modality dataset.
pixel_values = torch.cat([pv for pv in all_pixel_values if pv is not None], dim=0) if any(all_pixel_values) else None RuntimeError: Boolean value of Tensor with more than one value is ambiguous
such error exist while I am using the finetune_lora_vision script