ByteBrief
We're a portrait publication through and through. Turn your phone back and your briefing picks up right where you left it.
(We tried widescreen once. It wasn't us.)
Tencent released AngelSpec, an open-source torch-native training framework for speculative-decoding draft models. It supports autoregressive multi-token prediction and the block-parallel DFlash family. AngelSpec treats workload heterogeneity as a first-class design constraint, specializing structure, training data, and verification depth around it.
Tap to vote and see what everyone thinks.
Summary by ByteBrief