microsoft

/

Magma-8B

like 350

Follow

Microsoft 10.9k

Hey, do you have any suggestions on how to deploy this model?
What are good settings to get a nice efficiency/performance tradeoff?
What framework would you suggest?

I saw this request on vllm: https://github.com/vllm-project/vllm/issues/13870

Thanks so much, can't wait to try this out!