P-EAGLE: Parallel-Drafting EAGLE with Scalable Training
Paper • 2602.01469 • Published • 3
SpecDec
-
-
Direct Alignment of Draft Model for Speculative Decoding with Chat-Fine-Tuned LLMs
Paper • 2403.00858 • Published • 1 -
Gumiho: A Hybrid Architecture to Prioritize Early Tokens in Speculative Decoding
Paper • 2503.10135 • Published -
LK Losses: Direct Acceptance Rate Optimization for Speculative Decoding
Paper • 2602.23881 • Published • 18