SDSAT: Accelerating LLM Inference through Speculative Decoding with Semantic Adaptive Tokens

Add code
Apr 01, 2024

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: