SpecLA: Efficient Speculative Decoding for Linear-Attention Models

Add code
Jul 18, 2026

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: