UC San Diego's DFlash replaces autoregressive drafting with a lightweight block diffusion model for speculative decoding. It drafts whole token blocks in a single forward pass and conditions on target hidden features through KV injection. The paper reports up to 6.08x lossless speedup on Qwen3-8B, while NVIDIA reports up to 15x throughput on Blackwell at fixed interactivity. DFlash ships 20 checkpoints and supports SGLang, vLLM, and TensorRT-LLM.
The post DFlash Speculative Decoding Drafts Whole Token Blocks in Parallel for Up to 15x Higher Throughput on NVIDIA Blackwell appeared first on MarkTechPost.
source & further reading
marktechpost.com — original article
Mistral OCR 4 Brings Citation-Ready Structured Output to RAG, Agentic, and Enterprise Search Pipelines
How to Use NVIDIA Canary-1B-v2 for ASR, Translation, and Automatic SRT Subtitle Export in Python
Prime Intellect Releases prime-rl 0.6.0 to Train Trillion-Parameter MoE Models on Agentic RL Workloads