New
yinsongxu/llm2jev v0.2.0 — adds image Box judgment
yinsongxu/llm2jev is a tool that judges using the yes/no logits of the prefill, and v0.2.0 adds image bounding box judgment.
Projects
yinsongxu/llm2jev
Turn local language models into Jev-style structured decision models. Get results from text and images with prefill alone—no token-by-token decoding required.
View on GitHub ↗- stars
- ★ 449
- Forks
- 50
- Language
- Python
- License
- Apache-2.0
- Latest release
- v0.2.0
Posts
1No matching stories.