6.5k citations · 18.4k across the 52 of their papers we have counts for
25 papers · 1 filter
PVNAS: 3D Neural Architecture Search with Point-Voxel Convolution
Zhijian Liu, Haotian Tang, Shengyu Zhao +2
3D neural networks are widely used in real-world applications (e.g., AR/VR headsets, self-driving cars). They are required to be fast and accurate; however, limited hardware resour…
MCUNetV2: Memory-Efficient Patch-based Inference for Tiny Deep Learning
Ji Lin, Wei-Ming Chen, Han Cai +2
Tiny deep learning on microcontroller units (MCUs) is challenging due to the limited memory size. We find that the memory bottleneck is due to the imbalanced memory distribution in…
Network Augmentation for Tiny Deep Learning
Han Cai, Chuang Gan, Ji Lin +1
We introduce Network Augmentation (NetAug), a new training method for improving the performance of tiny neural networks. Existing regularization techniques (e.g., data augmentation…
TSM: Temporal Shift Module for Efficient and Scalable Video Understanding on Edge Device
Ji Lin, Chuang Gan, Kuan Wang +1
The explosive growth in video streaming requires video understanding at high accuracy and low computation cost. Conventional 2D CNNs are computationally cheap but cannot capture te…
LocTex: Learning Data-Efficient Visual Representations from Localized Textual Supervision
Zhijian Liu, Simon Stent, Jie Li +2
Computer vision tasks such as object detection and semantic/instance segmentation rely on the painstaking annotation of large training datasets. In this paper, we propose LocTex th…
PatchNet -- Short-range Template Matching for Efficient Video Processing
Huizi Mao, Sibo Zhu, Song Han +1
Object recognition is a fundamental problem in many video processing tasks, accurately locating seen objects at low computation cost paves the way for on-device video recognition.…
Anycost GANs for Interactive Image Synthesis and Editing
Ji Lin, Richard Zhang, Frieder Ganz +2
Generative adversarial networks (GANs) have enabled photorealistic image synthesis and editing. However, due to the high computational cost of large-scale generators (e.g., StyleGA…
Hardware-Centric AutoML for Mixed-Precision Quantization
Kuan Wang, Zhijian Liu, Yujun Lin +2
Model quantization is a widely used technique to compress and accelerate deep neural network (DNN) inference. Emergent DNN hardware accelerators begin to support mixed precision (1…
TinyTL: Reduce Activations, Not Trainable Parameters for Efficient On-Device Learning
Han Cai, Chuang Gan, Ligeng Zhu +1
On-device learning enables edge devices to continually adapt the AI models to new data, which requires a small memory footprint to fit the tight memory constraint of edge devices.…
MCUNet: Tiny Deep Learning on IoT Devices
Ji Lin, Wei-Ming Chen, Yujun Lin +3
Machine learning on tiny IoT devices based on microcontroller units (MCU) is appealing but challenging: the memory of microcontrollers is 2-3 orders of magnitude smaller even than…
Differentiable Augmentation for Data-Efficient GAN Training
Shengyu Zhao, Zhijian Liu, Ji Lin +2
The performance of generative adversarial networks (GANs) heavily deteriorates given a limited amount of training data. This is mainly because the discriminator is memorizing the e…
GAN Compression: Efficient Architectures for Interactive Conditional GANs
Muyang Li, Ji Lin, Yaoyao Ding +3
Conditional Generative Adversarial Networks (cGANs) have enabled controllable image synthesis for many vision and graphics applications. However, recent cGANs are 1-2 orders of mag…
Training Kinetics in 15 Minutes: Large-scale Distributed Training on Videos
Ji Lin, Chuang Gan, Song Han
Deep video recognition is more computationally expensive than image recognition, especially on large-scale datasets like Kinetics [1]. Therefore, training scalability is essential…
Point-Voxel CNN for Efficient 3D Deep Learning
Zhijian Liu, Haotian Tang, Yujun Lin +1
We present Point-Voxel CNN (PVCNN) for efficient, fast 3D deep learning. Previous work processes 3D data using either voxel-based or point-based NN models. However, both approaches…
HAQ: Hardware-Aware Automated Quantization with Mixed Precision
Kuan Wang, Zhijian Liu, Yujun Lin +2
Model quantization is a widely used technique to compress and accelerate deep neural network (DNN) inference. Emergent DNN hardware accelerators begin to support mixed precision (1…
TSM: Temporal Shift Module for Efficient Video Understanding
Ji Lin, Chuang Gan, Song Han
The explosive growth in video streaming gives rise to challenges on performing video understanding at high accuracy and low computation cost. Conventional 2D CNNs are computational…
Efficient Sparse-Winograd Convolutional Neural Networks
Xingyu Liu, Jeff Pool, Song Han +1
Convolutional Neural Networks (CNNs) are computationally intensive, which limits their application on mobile devices. Their energy is dominated by the number of multiplies needed t…
AMC: AutoML for Model Compression and Acceleration on Mobile Devices
Yihui He, Ji Lin, Zhijian Liu +3
Model compression is a critical technique to efficiently deploy neural network models on mobile devices which have limited computation resources and tight power budgets. Convention…
Deep Gradient Compression: Reducing the Communication Bandwidth for Distributed Training
Yujun Lin, Song Han, Huizi Mao +2
Large-scale distributed training requires significant communication bandwidth for gradient exchange that limits the scalability of multi-node training, and requires expensive high-…
Deep Generative Adversarial Networks for Compressed Sensing Automates MRI
Morteza Mardani, Enhao Gong, Joseph Y. Cheng +8
Magnetic resonance image (MRI) reconstruction is a severely ill-posed linear inverse task demanding time and resource intensive computations that can substantially trade off {\it a…
DSD: Dense-Sparse-Dense Training for Deep Neural Networks
Song Han, Jeff Pool, Sharan Narang +9
Modern deep neural networks have a large number of parameters, making them very hard to train. We propose DSD, a dense-sparse-dense training flow, for regularizing deep neural netw…
Generate Image Descriptions based on Deep RNN and Memory Cells for Images Features
Shijian Tang, Song Han
Generating natural language descriptions for images is a challenging task. The traditional way is to use the convolutional neural network (CNN) to extract image features, followed…
SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
Forrest N. Iandola, Song Han, Matthew W. Moskewicz +3
Recent research on deep neural networks has focused primarily on improving accuracy. For a given accuracy level, it is typically possible to identify multiple DNN architectures tha…
EIE: Efficient Inference Engine on Compressed Deep Neural Network
Song Han, Xingyu Liu, Huizi Mao +4
State-of-the-art deep neural networks (DNNs) have hundreds of millions of connections and are both computationally and memory intensive, making them difficult to deploy on embedded…