Spaces:

---

Duplicated from  UNESCO/nllb

kevinpro

/

R-PRM-Demo

like 3

Running on Zero

I am requesting a community grant to host a public, interactive demo for my 7B Process Reward Model (R-PRM), kevinpro/R-PRM-7B-DPO.

This model is designed to evaluate the quality of a language model's reasoning steps, not just the final answer. The demo will serve as a valuable tool for the open-source community, allowing researchers and developers to better understand and build upon process-level feedback mechanisms.

A GPU is needed to run the 7B model with responsive inference for a public-facing demo. Thank you for your consideration.

https://huggingface.co/kevinpro/R-PRM-7B-DPO