eagle-hq/EAGLE-4

Official Implementation of EAGLE-1 (ICML'24), EAGLE-2 (EMNLP'24), and EAGLE-3 (NeurIPS'25).

Official Implementation of DART (DART: Low-Latency Parallel Drafting with Continuity-Aware Tree Pruning for Speculative Decoding, EMNLP26).

JetSpec: Breaking the Scaling Ceiling of Speculative Decoding with Causal Parallel Tree Drafting

A lightweight inference engine supporting speculative speculative decoding (SSD).

DFlash: Block Diffusion for Flash Speculative Decoding

RTP-LLM: Alibaba's high-performance LLM inference engine for diverse applications.

[ICML 2024] Break the Sequential Dependency of LLM Inference Using Lookahead Decoding

TokenSpeed is a speed-of-light LLM inference engine.

Inferflow is an efficient and highly configurable inference engine for large language models (LLMs).