speculative-decoding

Speculative-decoding Articles

2 articles in this category

DFlash on RTX 3090: 207 tok/s Qwen3.5-27B with Speculative Decoding

May 6, 2026 · 7 min read

Run Qwen3.5-27B at 3.43x autoregressive speed on a single RTX 3090. Lucebox's DFlash port brings block-diffusion speculative decoding to GGUF — build, download weights, and start generating in under 20 minutes.

local-ai speculative-decoding

Gemma 4 MTP Drafters: Speculative Decoding for Local LLMs

May 6, 2026 · 4 min read

Google released official MTP drafter models for Gemma 4. A small companion model guesses tokens ahead, the big model verifies — same quality, nearly 3x speed on the same hardware.

local-ai youtube