Hacker News
- Fast Video Generation with Sliding Tile Attention https://hao-ai-lab.github.io/blogs/sta/ 2 comments
- Reasoning Without Hesitating: Efficient Cot Through Certainty Probing https://hao-ai-lab.github.io/blogs/dynasor-cot/ 5 comments
- Consistency LLM: converting LLMs to parallel decoders accelerates inference 3.5x https://hao-ai-lab.github.io/blogs/cllm/ 98 comments