Sultan

Hi Folks, Please add ELECTRA to the list of the supported models. ELECTRA is an efficient model that could match ALBERT's performance using less resources.

Hi, Any plan to implement RAG? RAG already had TF code with hugging face library and it actually uses PyTorch lightning. RAG achieve its best performance with LARGE BART and...

I think there is a bug in src/tevatron/driver/jax_train.py this line : https://github.com/texttron/tevatron/blob/0e939457444f78284ab0471da74a0c74bc76a833/src/tevatron/driver/jax_train.py#L147C43-L147C56 The issue is caused by defining the max_length to 32, assuming all queries will not exceed this length,...

converting Gemma maxtext compatible checkpoint to Hugging Face format

5

I have looked around for a script that could convert MaxText Gemma and Gemma 2 checkpoints to Hugging Face format but i have not find anything related. This may related...

Exporting Keras Llama Checkpoint to HF

7

Hi, As the Llama3 is popular model, it would be great if we can have a script that export Llama Keras checkpoint to HF. The code is a already exist...