Can I run Phi-3-mini-4k-instruct on a NVIDIA H200?
Yes.Runs
You can run this comfortably. On a NVIDIA H200 (141 GB VRAM), Phi-3-mini-4k-instruct needs about 10 GB (FP16) and should generate at roughly ~230 tok/s. Recommended: Run on your GPU with FP16.
microsoft/Phi-3-mini-4k-instructWhy
- At full precision this model needs ~10 GB, which fits your GPU VRAM.
How to run Phi-3-mini-4k-instruct on a H200
Serve with vLLM (OpenAI-compatible API)
# Fast production serving on http://localhost:8000/v1
vllm serve microsoft/Phi-3-mini-4k-instructPython (Transformers)
from transformers import pipeline
pipe = pipeline("text-generation", model="microsoft/Phi-3-mini-4k-instruct", device_map="auto")
print(pipe("Hello", max_new_tokens=50))