Hydra: Sequentially-Dependent Draft Heads for Medusa Decoding
Paper
•
2402.05109
•
Published
•
1
MTP
-
-
Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads
Paper • 2401.10774 • Published • 57 -
Better & Faster Large Language Models via Multi-token Prediction
Paper • 2404.19737 • Published • 78 -
Exploring the Latent Capacity of LLMs for One-Step Text Generation
Paper • 2505.21189 • Published • 60