Qwen3 8B DFlash b16 Paper GitHub Blog DFlash is a novel speculative decoding method that utilizes a lightweight block diffusion model for drafting. It enables efficient, high quality parallel drafting that pushes the limits of inference speed. This model is the drafter component. It must be used in conjunction with the target model Qwen/Qwen3 8B . 🚀 Quick Start SGLang Installation Launch Server Usage vLLM Installation Launch Server Usage Transformers This model requires trust remote code=True to load the custom architecture for block diffusion generation. Installation Ensure you have transformers and torch installed. Our evaluation is conducted with torch==2.9.0 and transformers=4.57.3. Inference The following example demonstrates how to load the DFlash drafter and the Qwen3 8B target model to perform speculative decoding. Evaluation DFlash achieves up to 6.17x lossless acceleration for Qwen3 8B , making it nearly 2.5x faster than the state of the art speculative decoding method EAGLE 3. Check out our GitHub repository to see how to reproduce the results. Citation If you find DFlash useful for your research or applications, please cite our project.
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy