Song Han — all works

34 works · ← Song Han

SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size ↗arXiv (Cornell University) · no summary yetForrest Iandola et al.2016 · 6,402 citations
TSM: Temporal Shift Module for Efficient Video Understanding ↗IEEE/CVF International Conference on Computer Vision (ICCV) · no summary yetJi Lin, Chuang Gan, Song Han2019 · 2,109 citations
AMC: AutoML for Model Compression and Acceleration on Mobile Devices ↗arXiv:1802.03494 · no summary yetYihui He; Ji Lin; Zhijian Liu; Hanrui Wang; Li-Jia Li;, Song Han2018 · 1,444 citations
EIE: Efficient Inference Engine on Compressed Deep Neural Network ↗arXiv:1602.01528 · no summary yetSong Han et al.2016 · 1,193 citations
HAQ: Hardware-Aware Automated Quantization With Mixed Precision ↗IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · no summary yetKuan Wang et al.2019 · 1,024 citations
SpAtten: Efficient Sparse Attention Architecture with Cascade Token and Head Pruning ↗arXiv:2012.09852 · no summary yetHanrui Wang, Zhekai Zhang, Song Han2020 · 534 citations
ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware ↗arXiv:1812.00332 · no summary yetHan Cai, Ligeng Zhu, Song Han2018 · 479 citations
VILA: On Pre-training for Visual Language ModelsIEEE Computer Society Conference on Computer Vision and Pattern RecognitionJi Lin, Hongxu Yin, Wei Ping, et al.2023 · 381 citations
EfficientViT: Lightweight Multi-Scale Attention for High-Resolution Dense Prediction ↗IEEE/CVF International Conference on Computer Vision (ICCV) · no summary yetHan Cai et al.2023 · 363 citations
HAT: Hardware-Aware Transformers for Efficient Natural Language ProcessingAnnual Meeting of the Association for Computational Linguistics (ACL)Hanrui Wang, Zhanghao Wu, Zhijian Liu, et al.2020 · 294 citations
LongLoRA: Efficient Fine-tuning of Long-Context Large Language ModelsarXiv:2309.12307Yukang Chen, Shengju Qian, Haotian Tang, et al.2023 · 268 citations
Once-for-All: Train One Network and Specialize it for Efficient Deployment ↗arXiv:1908.09791 · no summary yetHan Cai et al.2019 · 252 citations
Quest: Query-Aware Sparsity for Efficient Long-Context LLM InferencearXiv:2406.10774Jiaming Tang, Yilong Zhao, Kan Zhu, et al.2024 · 203 citations
QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving ↗MLSys 2025 · no summary yetYujun Lin et al.2024 · 120 citations
DuoAttention: Efficient Long-Context LLM Inference with Retrieval and Streaming HeadsarXiv:2410.10819Guangxuan Xiao, Jiaming Tang, Jingwei Zuo, et al.2024 · 114 citations
NVILA: Efficient Frontier Visual Language Models ↗IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · no summary yetZhijian Liu et al.2025 · 73 citations
XAttention: Block Sparse Attention with Antidiagonal ScoringarXiv:2503.16428Ruyi Xu, Guangxuan Xiao, Haofeng Huang, et al.2025 · 59 citations
Sparse VideoGen: Accelerating Video Diffusion Transformers with Spatial-Temporal Sparsity ↗arXiv:2502.01776 · no summary yetHaocheng Xi et al.2025 · 56 citations
LongVILA: Scaling Long-Context Visual Language Models for Long Videos ↗arXiv:2408.10188 · no summary yetYukang Chen et al.2024 · 49 citations
Sparse VideoGen2: Accelerate Video Generation with Sparse Attention via Semantic-Aware Permutation ↗arXiv:2505.18875 · no summary yetShuo Yang et al.2025 · 44 citations
Radial Attention: O(n n) Sparse Attention with Energy Decay for Long Video Generation ↗arXiv (Cornell University) · no summary yetLi, Xingyang et al.2025 · 23 citations
Efficient Spatially Sparse Inference for Conditional GANs and Diffusion Models ↗arXiv:2211.02048 · no summary yetMuyang Li; Ji Lin; Chenlin Meng; Stefano Ermon; Song Han;, Jun-Yan Zhu2022 · 19 citations
MLSys: The New Frontier of Machine Learning Systems ↗arXiv (Cornell University) · no summary yetAlexander Ratner et al.2019 · 19 citations
Twilight: Adaptive Attention Sparsity with Hierarchical Top-p Pruning ↗neural information processing systems · no summary yetChaofan Lin et al.2025 · 17 citations
WorldModelBench: Judging Video Generation Models As World Models ↗neural information processing systems · no summary yetDacheng Li et al.2025 · 11 citations
FlatFormer: Flattened Window Attention for Efficient Point Cloud Transformer ↗arXiv:2301.08739 · no summary yetZhijian Liu et al.2023 · 11 citations
Lightning OPD: Efficient Post-Training for Large Reasoning Models with Offline On-Policy Distillation ↗arXiv:2604.13010 · no summary yetYecheng Wu, Song Han, Han Cai2026 · 10 citations
Scaling Vision Pre-Training to 4K Resolution ↗IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · no summary yetBaifeng Shi et al.2025 · 7 citations
Wolf: Dense Video Captioning with a World Summarization Framework ↗arXiv (Cornell University) · no summary yetBoyi Li et al.2024 · 6 citations
Radial Attention: O(n n) Sparse Attention with Energy Decay for Long Video Generation ↗neural information processing systems · no summary yetXingyang Li et al.2025 · 3 citations
Accelerating Large-Scale Reasoning Model Inference with Sparse Self-Speculative Decoding ↗arXiv:2512.01278 · no summary yetYilong Zhao et al.2025 · 2 citations
Attend Before Attention: Efficient and Scalable Video Understanding via Autoregressive Gazing ↗arXiv (Cornell University) · no summary yetBaifeng Shi et al.2026 · 1 citation
FreeToken: Efficient Edge-Native MoE Serving with Bandwidth-Adaptive Execution ↗arXiv:2608.16157 · no summary yetShuo Yang et al.2026 · 1 citation
FoundationMotion: Auto-Labeling and Reasoning about Spatial Movement in Videos ↗arXiv (Cornell University) · no summary yetGan, Yulu et al.2025 · 1 citation

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/authors/song-han/works/

🧑‍🔬 Song Han — all works

Song Han — all works