# z-lab/dflash

DFlash: Block Diffusion for Flash Speculative Decoding

Repository: https://github.com/z-lab/dflash
Canonical: https://ross.abutalabs.com/products/dflash
Homepage: https://dflash.z-lab.ai
Language: Python
License: MIT
License Family: permissive
Last push: 2026-08-18T21:14:10+00:00

## Health v2 (maintenance only)
Score: 71/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 66, longevity 17
- inputs: {"age_days": 241, "days_push": 15, "days_rel": 15, "gap_med": null, "n_releases_24m": 1}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5967, forks 414 (observed 2026-08-28T04:09:33.779300+00:00)

## What it is
DFlash is a lightweight block diffusion model used as a draft model for speculative decoding of large language models, drafting entire token blocks in a single parallel forward pass. It ships as a pip-installable Python package with checkpoints for many model families and integrates with Transformers, MLX, SGLang, vLLM, oMLX, and llama.cpp serving backends.

## Use cases
- speed up LLM inference with speculative decoding
- draft multiple tokens in parallel for a target LLM
- serve Qwen or Llama models with faster generation in SGLang
- benchmark speculative decoding speedups against EAGLE-3
- run lossless LLM acceleration on Apple Silicon with MLX
- use a diffusion-based draft model with vLLM or llama.cpp

## When to choose
- you want lossless LLM inference speedups beyond autoregressive drafters like EAGLE-3
- your target model has an official DFlash checkpoint (Qwen, Gemma, Llama, Kimi, MiniMax, GLM, GPT-OSS)
- you serve models via SGLang, vLLM, oMLX, or llama.cpp OpenAI-compatible servers

## When to avoid
- your target model family has no available DFlash checkpoint
- you need training or fine-tuning of the draft model rather than inference
- you cannot run a GPU or Apple Silicon backend

## Facets
- artifact type: library
- maturity: active
- function: llm-inference, machine-learning, deep-learning
- domain: large-language-models, machine-learning, deep-learning, performance
- platform: python
- tags: speculative-decoding, block-diffusion, llm-acceleration, draft-model, sglang, vllm, mlx, transformers, inference-optimization, linux, macos, gpu, docker

## Member repositories
- z-lab/dflash (main) score 71

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:33.779300+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:50:09.809566+00:00, confidence not recorded.
  - readme: https://github.com/z-lab/dflash (fetched 2026-08-28T04:09:33.779300+00:00, sha 546a0c4c4b2d)
  - homepage: https://dflash.z-lab.ai (fetched 2026-08-29T08:46:05.717230+00:00, sha 546f39c97c3c)
  - registry_pypi: https://pypi.org/pypi/dflash/json (fetched 2026-08-29T08:46:05.719993+00:00, sha fa904852f2c7)
- Data as of 2026-08-30T08:39:29.467469+00:00.