deepseek-ai

/

deepseek-coder-6.7b-instruct

like 309

Does anyone know why this calculator reports such a large VRAM usage for the context size compared to all other models I've checked?

Even deepseek 1.3B or 33B have lower VRAM requirements for the context size. Any explanation?

https://huggingface.co/spaces/NyxKrage/LLM-Model-VRAM-Calculator