Fetching the paper…

Accelerating Transformer Inference for Translation via Parallel Decoding · Around