Several trained models to compare the differences between each method. Each model has a complete description of hyperparams with wandb reports.
G
G-reen
AI & ML interests
SFT, DPO, ORPO, LLMs, text-generation
Recent Activity
updated
a model
11 days ago
G-reen/SmolLM3-3B-SFT
published
a model
11 days ago
G-reen/SmolLM3-3B-SFT
updated
a model
about 1 month ago
G-reen/Qwen2.5-3B-W8A8FP
Organizations
None yet