Пожалуйста, обратите внимание, что пользователь заблокирован
https://artificialanalysis.ai/modelsIs there a site/repo that maintains a mapping of models and their minimal compute hardware requirement to run at a sane token/s rate?
https://huggingface.co/spaces/ArtificialAnalysis/LLM-Performance-Leaderboard
https://llm.extractum.io/list/
I can't give you any specifics here. However, it will require a lot of resources. Probably four or more than eight NVIDIA A100What cluster of GPUs would I need to get to run the mega model?