SJTU-DENG-Lab/Discrete-Diffusion-Forcing
Discrete Diffusion Forcing (D2F): dLLMs Can Do Faster-Than-AR Inference
This project helps AI developers and researchers significantly speed up large language model (LLM) inference. It takes an existing diffusion-based LLM and provides a method to generate text, code, or other sequences much faster. The result is quicker responses from LLMs, making them more practical for real-time applications, while maintaining high generation quality.
244 stars.
Use this if you are building or deploying large language models and need to accelerate their text generation speed, especially for high-throughput applications, without sacrificing output quality.
Not ideal if you are working with traditional autoregressive LLMs and are not familiar with diffusion models or advanced model training/inference techniques.
Stars
244
Forks
17
Language
Python
License
MIT
Category
Last pushed
Feb 03, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/SJTU-DENG-Lab/Discrete-Diffusion-Forcing"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
ljleb/sd-mecha
Executable State Dict Recipes
declare-lab/tango
A family of diffusion models for text-to-audio generation.
Li-Jinsong/DAEDAL
[ICLR 2026] Official repository of "Beyond Fixed: Training-Free Variable-Length Denoising for...
SalesforceAIResearch/CoDA
Salesforce AI Research's open diffusion language model
ZhanqiuHu/flash-dlm-experimental
Implementation of Flash-DLM (paper: FlashDLM: Accelerating Diffusion Language Models via...