Rangehow

llada-pretrain-hf. A repository for pretraining a discrete diffusion model (llada), with all components built on the Hugging Face ecosystem.

tofu-search. Python

sft. Python

NFP. Python

light_ifeval. A light repo for supporting eval ifeval for various model (include thinking)

mamba_training. A repo that used to train mamba in huggingface

cardcraft. HTML