mit-han-lab

[ICLR 2024] Efficient Streaming Language Models with Attention Sinks

[MLSys 2024 Best Paper Award] AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration

Efficient vision foundation models for high-resolution generation and perception.

[ICRA'23] BEVFusion: Multi-Task Multi-Sensor Fusion with Unified Bird's-Eye View Representation

[ICCV 2019] TSM: Temporal Shift Module for Efficient Video Understanding

[ICLR 2020] Once for All: Train One Network and Specialize it for Efficient Deployment

[ICML 2023] SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models

A PyTorch-based framework for Quantum Classical Simulation, Quantum Machine Learning, Quantum Neural Networks, Parameterized Quantum Circuits with support for easy deployments on real quantum...

[MICRO'23, MLSys'22] TorchSparse: Efficient Training and Inference Framework for Sparse Convolution on GPUs.

[ICLR 2019] ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware

[NeurIPS 2020] Differentiable Augmentation for Data-Efficient GAN Training

[CVPR 2020] GAN Compression: Efficient Architectures for Interactive Conditional GANs

StreamingVLM: Real-Time Understanding for Infinite Video Streams

TinyChatEngine: On-Device LLM Inference Library

[NeurIPS 2020] MCUNet: Tiny Deep Learning on IoT Devices; [NeurIPS 2021] MCUNetV2: Memory-Efficient Patch-based Inference for Tiny Deep Learning; [NeurIPS 2022] MCUNetV3: On-Device Training Under...

[MLSys'25] QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving; [MLSys'25] LServe: Efficient Long-sequence LLM Serving with Unified Sparse Attention

[CVPR 2021] Anycost GANs for Interactive Image Synthesis and Editing

[CVPR 2024 Highlight] DistriFusion: Distributed Parallel Inference for High-Resolution Diffusion Models

[NeurIPS 2020] MCUNet: Tiny Deep Learning on IoT Devices; [NeurIPS 2021] MCUNetV2: Memory-Efficient Patch-based Inference for Tiny Deep Learning

[IJCV] FastComposer: Tuning-Free Multi-Subject Image Generation with Localized Attention

[NeurIPS 2019, Spotlight] Point-Voxel CNN for Efficient 3D Deep Learning

HART: Efficient Visual Generation with Hybrid Autoregressive Transformer

[ECCV 2020] Searching Efficient 3D Architectures with Sparse Point-Voxel Convolution

[ICLR 2020] Lite Transformer with Long-Short Range Attention

[NeurIPS 2025] Radial Attention: O(nlogn) Sparse Attention with Energy Decay for Long Video Generation

A sparse attention kernel supporting mix sparse patterns

[ICLR 2025] DuoAttention: Efficient Long-Context LLM Inference with Retrieval and Streaming Heads

On-Device Training Under 256KB Memory [NeurIPS'22]

Real-Time VLAs via Future-state-aware Asynchronous Inference.

[NeurIPS 2019] Deep Leakage From Gradients

[ECCV 2018] AMC: AutoML for Model Compression and Acceleration on Mobile Devices

[ICLR 2025] VILA-U: a Unified Foundation Model Integrating Visual Understanding and Generation

[ICML 2024] Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference

[CVPR 2019, Oral] HAQ: Hardware-Aware Automated Quantization with Mixed Precision

Offsite-Tuning: Transfer Learning without Full Model

[ACL'20] HAT: Hardware-Aware Transformers for Efficient Natural Language Processing

[CVPR'22] Lite Pose: Efficient Architecture Design for 2D Human Pose Estimation

[ICML 2025] XAttention: Block Sparse Attention with Antidiagonal Scoring

Code for the papers: “Four Over Six: More Accurate NVFP4 Quantization with Adaptive Block Scaling” and “Adaptive Block-Scaled Data Types”

[MLSys 2021] IOS: Inter-Operator Scheduler for CNN Acceleration

[ASPLOS'26] Taming the Long-Tail: Efficient Reasoning RL Training with Adaptive Drafter

[ECCV 2018] AMC: AutoML for Model Compression and Acceleration on Mobile Devices

[CVPR 2020] APQ: Joint Search for Network Architecture, Pruning and Quantization Policy

[HPCA'21] SpAtten: Efficient Sparse Attention Architecture with Cascade Token and Head Pruning

[CVPR'23] FlatFormer: Flattened Window Attention for Efficient Point Cloud Transformer

[ICLR 2026 Oral] Locality-aware Parallel Decoding for Efficient Autoregressive Image Generation

[CVPR 2026 Highlight] ForeAct: Steering Your VLA with Efficient Visual Foresight Planning

Patch convolution to avoid large GPU memory usage of Conv2D

[CVPR'23] SparseViT: Revisiting Activation Sparsity for Efficient High-Resolution Vision Transformer

Binary Neural Network on IceStick FPGA.