meta-llama

/

Llama-3.2-11B-Vision-Instruct

like 300

Hi experts,

I see HF version of checkpoint and config of llama 3.2 11b vision uses image size 560. But in Meta's repo it claims 448 for 11B and 560 for 90B. I wonder if it's a mistake in HF config?
https://github.com/meta-llama/llama-models/blob/main/models/sku_list.py#L662