data-augmentation-with-llms icon indicating copy to clipboard operation
data-augmentation-with-llms copied to clipboard
verified publisher icon ServiceNow

the problem about running the prepare_dataset.py code

Open wmqnl opened this issue 3 years ago • 1 comments

Hi, when I run your code like this: python prepare_dataset.py --name clinc_oos --data_root '. /data', it occurs "RuntimeError: The size of tensor a (512) must match the size of tensor b (513) at non-singleton dimension 3" (the revelant files of the clinic_oos dataset have located at '. /data' directory). How can i solve this problem? Could you provide some advice for me? Thanks for your good work! Thank you very much!

wmqnl avatar Aug 18 '22 14:08 wmqnl

Hi @wmqnl

Sorry for the delay in response. Which mode do you see that error for? It looks like it's for GPTJ, in which case, it might be due to the constraint on the input size of the model. I can help more if you show me the full error log.

Thanks, Gaurav.

Demfier avatar Sep 15 '22 23:09 Demfier