dtype: float32 in base model vs. dtype: bfloat16 in the instruction fine-tuned model
#32
by
tanliboy
- opened
In this base model, the dtype is float32; however, in the instruction fine-tuned model, the dtype is (https://huggingface.co/google/gemma-2-9b-it/blob/main/config.json#L29).
Is this inconsistency intentional or a bug?