Spaces:
Running
Running
Apply for a GPU community grant: Personal project
#1
by Ujayer - opened
Description
This Space showcases QwenFable (Qwable 27B) — a production-grade 27B language model fully fine-tuned on 105 trillion tokens. The goal is to provide an interactive chat demo where users can try the model's advanced reasoning, software engineering, and coding capabilities in real time.
Justification
This is an open-source research demo for a community-driven model fine-tuned on synthetic reasoning data (Fable 5 + Claude Mythos traces). Making it interactive on ZeroGPU would benefit the open-source AI community and help users evaluate the model without local hardware requirements. The Space is public, MIT-licensed, and intended as an educational/research tool for the broader Hugging Face community.