Pre-optimized and pre-compield models in Hugging Face Hub

Furiosa LLM
1

Hi folks,

Since the 2025.2 release, we’ve uploaded the pre-optimized and pre-compield models in Hugging Face Hub.

For example, without an artifact compilation step, you can quickly run as follows:

furiosa-llm serve furiosa-ai/Llama-3.1-8B-Instruct-FP8

You can find more available pre-compiled models at furiosa-ai (FuriosaAI).