metadata
license: other
library_name: transformers
base_model: GSAI-ML/LLaDA-8B-Instruct
tags:
- diffusion-language-model
- dllm
- LLaDA
- WINO
- WINO-plus
WINO-plus-LLaDA-8B-Instruct
This repository contains the full merged 8B model weights of WINO+, initialized from LLaDA-8B-Instruct.
WINO+ is a trajectory-injection method for diffusion large language models. It transfers the verified denoising order discovered by WINO into model parameters, enabling faster and higher-quality generation.
The model corresponds to the WINO+ method described in Roll Out and Roll Back: Diffusion LLMs are Their Own Efficiency Teachers.