Song

songhan

·

https://songhan.mit.edu

AI & ML interests

efficient AI computing

Recent Activity

authored a paper 6 days ago

Quant VideoGen: Auto-Regressive Long Video Generation via 2-Bit KV-Cache Quantization

upvoted a paper 8 days ago

FreeToken: Efficient Edge-Native MoE Serving with Bandwidth-Adaptive Execution

authored a paper 30 days ago

Sol-Attn: Accelerating Video Generation Inference via On-the-Fly Attention Sparsification

View all activity

Organizations

upvoted a paper 8 days ago

FreeToken: Efficient Edge-Native MoE Serving with Bandwidth-Adaptive Execution

Paper • 2608.16157 • Published 10 days ago • 98

upvoted a collection 10 months ago

SANA-Video

🎬 SANA-Video: Efficient Video Generation with Block Linear Diffusion Transformer • 11 items • Updated 7 days ago • 11

upvoted a paper 10 months ago

OmniVinci: Enhancing Architecture and Data for Omni-Modal Understanding LLM

Paper • 2510.15870 • Published Oct 17, 2025 • 93

upvoted a paper 11 months ago

SANA-Video: Efficient Video Generation with Block Linear Diffusion Transformer

Paper • 2509.24695 • Published Sep 29, 2025 • 54

upvoted a paper about 1 year ago

Scaling RL to Long Videos

Paper • 2507.07966 • Published Jul 10, 2025 • 161

upvoted a paper over 1 year ago

NVILA: Efficient Frontier Visual Language Models

Paper • 2412.04468 • Published Dec 5, 2024 • 62