search

samlhuillier / code-llama-fine-tune-notebook

Fine-tune Code Llama to generate SQL queries from text
32 stars 6 forks source link

RuntimeError: expected mat1 and mat2 to have the same dtype, but got: c10::Half != float #8

Open A-smarneh opened 7 months ago

A-smarneh commented 7 months ago

Hi,

Thanks a lot for this notebook. I have followed the notebook and when I get to the

"trainer.train()"

I get the following run time error:

RuntimeError: expected mat1 and mat2 to have the same dtype, but got: c10::Half != float

I am working on AWS SageMaker, with ml.p3.2xlagre instance. I would appreciate any help you can provide.

Batwho commented 3 months ago

same issue, are you able to figure this out?

A-smarneh commented 3 months ago

@Batwho the problem was in the dataset. The dataset contained question and answer field and the model expected 3 question, context and answer

Batwho commented 3 months ago

got it thanks!