Tencent open-sources AngelSpec for speculative decoding
The torch-native release covers both autoregressive multi-token prediction (MTP) and the block-parallel DFlash family, targeting Hy3 models. A companion paper is on arXiv.
2 sources
Daily brief
Get tomorrow's AI brief in your inbox
More stories today
- Hailuo AI journey from 0 to scale highlighted by MiniMax
- Formula 1 adopts agentic AI on AWS to accelerate data operations
- H3 full precision weights showcased on Reddit
- OWASP AI Security Verification Standard offers a framework for secure AI apps
- Intology shows AI models training other models