Unranked—
Known for
Fine-tuning an open model without writing a training loop per architecture.
One interface for fine-tuning more than a hundred language and vision models, covering the tuning methods most teams reach for.
52.0index score · Open Source
Stars72.7 ×0.25
Forks78.8 ×0.33
Activity18.2 ×0.42
Each factor shows where this entry sits among the 33 ranked on this board. Weights are shared out across the factors we have data for.
A training framework rather than something you run to serve or use a model. It sits beside what this board measures — the tools people deploy and talk to — so it is listed rather than ordered against them.