ValueError: PhiForCausalLM does not support gradient checkpointing.

#12
by Tapendra - opened

Finetune on new dataset with SFTTrainer but getting ValueError: PhiForCausalLM does not support gradient checkpointing. error ?

Seems like there is a PR for a fix; since the model is on the hub nothing we can do on transformers

same issue. looking forward to solving this.

We'll add support for this model in transformers soon!

Microsoft org

Hello everyone!

This will be updated once we integrate with the Phi implementation in HF.

Best regards,
Gustavo.

gugarosa changed discussion status to closed

Sign up or log in to comment