repoGitHubTrust 82 · PrimaryPublished 20h agoLive · 19h ago
sgl-project/SpecForge
Train speculative decoding models effortlessly and port them smoothly to SGLang serving.
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Implements
Covers
Related across the graph
paperWhen Is a Draft Accepted? A Theory of Acceptance in Speculative DecodingpaperBlockPilot: Instance-Adaptive Policy Learning for Diffusion-based Speculative Decodingnews[Research] JetSpec: Speculative Decoding with Parallel Tree Drafting Enables up to 9.64x Lossless LLM Inference Speedup with more than 1000TPSnewsDSpark: Speculative decoding accelerates LLM inference [pdf]paperSpeculative decoding with draft models
