This repository contains the $R^2$-dLLM-Dream model, an accelerated Diffusion Large Language Model (dLLM) introduced in the paper
$R^2$-dLLM: Accelerating Diffusion Large Language Models via Spatio-Temporal Redundancy Reduction.
$R^2$-dLLM is a unified framework for reducing decoding redundancy in dLLMs from both inference and training perspectives. This specific checkpoint is a redundancy-aware supervised fine-tuned (SFT) version based on
Dream-v0-Instruct-7B.
Diffusion Large Language Models (dLLMs) enable parallel token prediction but often suffer from high inference latency due to recurring spatial and temporal redundancy during decoding. $R^2$-dLLM addresses this through:
Experiments demonstrate that $R^2$-dLLM can reduce the number of decoding steps by up to 88% compared to existing strategies while maintaining competitive generation quality.
1@article{du2026r,
2 title={$R^{2}$-dLLM: Accelerating Diffusion Large Language Models via Spatio-Temporal Redundancy Reduction},
3 author={Du, Zhenbang and Xia, Kejing and Zhong, Xinrui and Fu, Yonggan and Oswald, Nicolai and Ji, Binfei and Khailany, Brucek and Molchanov, Pavlo and Lin, Yingyan},
4 journal={arXiv preprint arXiv:2604.18995},
5 year={2026}
6}
