Fetching the paper…

Learning to Keep a Promise: Scaling Language Model Decoding Parallelism with Learned Asynchronous Decoding · Around