Urgent.News

What's breaking now, across hundreds of outlets.

AI

D-REAL: Acceptance-Length Optimization with Unbiased Gradients for Speculative Decoding

Speculative decoding efficiency depends on the number of consecutively accepted draft tokens, yet token-level training objectives do not directly optimize this quantity. We introduce D-REAL (Drafter Reinforcement of Expected Acceptance Length), a post-training method that directly optimizes expected acceptance length for a fixed decoding context. D-REAL weights position-wise draft-target…

We haven't written up this one. arXiv cs.AI has the full story — the link below goes straight to it.

Source: arxiv.org

Read the original at arxiv.org →

Report a problem with this story

More in AI

More from Saturday 7 November →