Models
Nothing released yet. When we do release something it will be narrow: adapters trained with LoRA or QLoRA on open foundations, for one task each, with the benchmarks and the limitations attached. Until then the work lives in research.
Nothing released yet. When we do release something it will be narrow: adapters trained with LoRA or QLoRA on open foundations, for one task each, with the benchmarks and the limitations attached. Until then the work lives in research.