also, care to drop the link to the model? how big is it? i can run about 14b models on my GPU
I ran it on duck.ai
https://huggingface.co/meta-llama/Llama-4-Scout-17B-16E-Instruct
Please Login to reply.
goodness, that's a big model, 44gb?
oh, that was the "old" version, the new version is 39. i'm skeptical that this runs fast on my 16gb GPU tho
