Model: promotion/ronpo-qwen3-8b-fair-inpo-avg-s42 Source: Original Platform
base_model, library_name, tags
| base_model | library_name | tags | |||
|---|---|---|---|---|---|
| Qwen/Qwen3-8B | transformers |
|
Qwen3-8B fair-demo checkpoint: inpo_avg
Validation-selected candidate: inpo_b. Exact evaluator, selection, sweep, and training metadata are stored under experiment/.
Description
Languages
Jinja
100%