Verified record · from the answer index

speculative-decoding

Accelerate LLM inference using speculative decoding, Medusa multiple heads, and lookahead decoding techniques. Use when optimizing inference speed (1.5-3.6× speedup), reducing latency for real-time applications, or deploying models with lim

skill:orchestra-research/ai-research-skills/speculative-decoding · published by orchestra-research · source repository