ML-GSAI/Diffusion-LLM-Papers
FreeCurated papers on diffusion language models — LLaDA, Dream, MMaDA, consistency sampling, fast inference; 169 stars, actively maintained (2026) 
About ML-GSAI/Diffusion-LLM-Papers
A curated collection of papers on diffusion language models, maintained by the ML-GSAI community. The repository categorizes research into theoretical basis, foundation models (e.g., LLaDA, Dream, Seed Diffusion), multimodal models (including understanding, unified, and speech/ASR), fast sampling techniques with KV caching and advanced sampling methods, and reinforcement learning for diffusion LLMs. It aims to provide a structured reference for researchers and practitioners, with contributions accepted via pull requests.
Key Features
Pros & Cons
- Comprehensive collection of diffusion language model papers in one place
- Well-organized categories for easy navigation
- Actively maintained with recent papers (up to August 2025)
- Community-driven, allowing contributions from the research community
- May not cover all diffusion LLM papers due to time constraints (acknowledged by maintainers)
- Only includes papers up to the last commit date; requires manual curation
- Requires GitHub familiarity to browse and contribute