trikwi/chsa-p14-dpo-adapter
Adaptateur DPO — projet CHSA P14. Depot prive.
| modele de base | Qwen/Qwen3-1.7B-Base @ ea980cb0a6c2ae4b936e82123acc929f1cec04c1 |
| parent | trikwi/chsa-p14-sft-v2-adapter @ 5d23e67e4604924ed1457728945f8b0fde6175a9 |
| donnees | trikwi/chsa-p14-dpo-data @ 190b6eaeaad786b3fc947a5de0a8bd07eef0fde3 |
| train / validation | 4000 / 500 paires |
| beta | 0.1 · graine 14 · lot effectif 16 |
| duree | 149 min |
| eval_loss | 0.5341736674308777 |
| W&B | https://wandb.ai/richardalexandrehugou-opencassrooms/chsa-p14/runs/npdnhkjz |
Preferences issues majoritairement d'un juge automatise. Aucune validation clinique. Ne pas utiliser en situation de soin.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for trikwi/chsa-p14-dpo-adapter
Base model
Qwen/Qwen3-1.7B-Base