ModelHarmonizing Accuracy and Efficiency for Speculative Decoding
HASS
A speculative decoding method designed to balance accuracy and efficiency, used as a comparison baseline for RCSD.
Timeline (last 90 days)
Oct 22
Compared with the HASS method, RCSD improves the average acceptance length by 35%.
Unverified50%
Oct 22
HASS (Harmonizing Accuracy and Efficiency for Speculative Decoding) is a representative baseline method for training draft models in speculative decoding, mainly based on distribution-matching training objectives.
Unverified50%
All Facts (2)
Unverified
Compared with the HASS method, RCSD improves the average acceptance length by 35%.
50%UnverifiedHASS (Harmonizing Accuracy and Efficiency for Speculative Decoding) is a representative baseline method for training draft models in speculative decoding, mainly based on distribution-matching training objectives.
50%