Finetuning Hardware Recomendations
greyweb opened this issue · comments
Hi,
I am trying to finetune GPT-J 6B from HF converted weights. It would be great to know some recommendations on the finetuning compute widely used/ suggested for GPT-J 6B.
Model parallel transformers in JAX and Haiku
greyweb opened this issue · comments
Hi,
I am trying to finetune GPT-J 6B from HF converted weights. It would be great to know some recommendations on the finetuning compute widely used/ suggested for GPT-J 6B.