SPIRe: Boosting LLM Inference Throughput with Speculative Decoding

Add code
Apr 08, 2025

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: