128 citations · 487 across the 23 of their papers we have counts for
23 papers
SysNoise: Exploring and Benchmarking Training-Deployment System Inconsistency
Yan Wang, Yuhang Li, Ruihao Gong +8
Extensive studies have shown that deep learning models are vulnerable to adversarial and natural noises, yet little is known about model robustness on noises caused by different sy…
UniHCP: A Unified Model for Human-Centric Perceptions
Yuanzheng Ci, Yizhou Wang, Meilin Chen +7
Human-centric perceptions (e.g., pose estimation, human parsing, pedestrian detection, person re-identification, etc.) play a key role in industrial applications of visual models.…
Improving Long-tailed Object Detection with Image-Level Supervision by Multi-Task Collaborative Learning
Bo Li, Yongqiang Yao, Jingru Tan +4
Data in real-world object detection often exhibits the long-tailed distribution. Existing solutions tackle this problem by mitigating the competition between the head and tail cate…
The Equalization Losses: Gradient-Driven Training for Long-tailed Object Recognition
Jingru Tan, Bo Li, Xin Lu +4
Long-tail distribution is widely spread in real-world applications. Due to the extremely small ratio of instances, tail categories often show inferior accuracy. In this paper, we f…
Mimose: An Input-Aware Checkpointing Planner for Efficient Training on GPU
Jianjin Liao, Mingzhen Li, Qingxiao Sun +8
Larger deep learning models usually lead to higher model quality with an ever-increasing GPU memory footprint. Although tensor checkpointing techniques have been proposed to enable…
Outlier Suppression: Pushing the Limit of Low-bit Transformer Language Models
Xiuying Wei, Yunchen Zhang, Xiangguo Zhang +5
Transformer architecture has become the fundamental element of the widespread natural language processing~(NLP) models. With the trends of large NLP models, the increasing memory a…
Towards Frame Rate Agnostic Multi-Object Tracking
Weitao Feng, Lei Bai, Yongqiang Yao +2
Multi-Object Tracking (MOT) is one of the most fundamental computer vision tasks that contributes to various video analysis applications. Despite the recent promising progress, cur…
QDrop: Randomly Dropping Quantization for Extremely Low-bit Post-Training Quantization
Xiuying Wei, Ruihao Gong, Yuhang Li +2
Recently, post-training quantization (PTQ) has driven much attention to produce efficient neural networks without long-time retraining. Despite its low cost, current PTQ works tend…
Equalized Focal Loss for Dense Long-Tailed Object Detection
Bo Li, Yongqiang Yao, Jingru Tan +4
Despite the recent success of long-tailed object detection, almost all long-tailed object detectors are developed based on the two-stage paradigm. In practice, one-stage detectors…
INTERN: A New Learning Paradigm Towards General Vision
Jing Shao, Siyu Chen, Yangguang Li +24
Enormous waves of technological innovations over the past several years, marked by the advances in AI technologies, are profoundly reshaping the industry and the society. However,…
MQBench: Towards Reproducible and Deployable Model Quantization Benchmark
Yuhang Li, Mingzhu Shen, Jian Ma +6
Model quantization has emerged as an indispensable technique to accelerate deep learning inference. While researchers continue to push the frontier of quantization algorithms, exis…
Supervision Exists Everywhere: A Data Efficient Contrastive Language-Image Pre-training Paradigm
Yangguang Li, Feng Liang, Lichen Zhao +5
Recently, large-scale Contrastive Language-Image Pre-training (CLIP) has attracted unprecedented attention for its impressive zero-shot recognition ability and excellent transferab…
RobustART: Benchmarking Robustness on Architecture Design and Training Techniques
Shiyu Tang, Ruihao Gong, Yan Wang +9
Deep neural networks (DNNs) are vulnerable to adversarial noises, which motivates the benchmark of model robustness. Existing benchmarks mainly focus on evaluating defenses, but th…
Incorporating Convolution Designs into Visual Transformers
Kun Yuan, Shaopeng Guo, Ziwei Liu +3
Motivated by the success of Transformers in natural language processing (NLP) tasks, there emerge some attempts (e.g., ViT and DeiT) to apply Transformers to the vision domain. How…
Diversifying Sample Generation for Accurate Data-Free Quantization
Xiangguo Zhang, Haotong Qin, Yifu Ding +6
Quantization has emerged as one of the most prevalent approaches to compress and accelerate neural networks. Recently, data-free quantization has been widely studied as a practical…
BRECQ: Pushing the Limit of Post-Training Quantization by Block Reconstruction
Yuhang Li, Ruihao Gong, Xu Tan +6
We study the challenging task of neural network quantization without end-to-end retraining, called Post-training Quantization (PTQ). PTQ usually requires a small subset of training…
MixMix: All You Need for Data-Free Compression Are Feature and Data Mixing
Yuhang Li, Feng Zhu, Ruihao Gong +5
User data confidentiality protection is becoming a rising challenge in the present deep learning research. Without access to data, conventional data-driven model compression faces…
Once Quantization-Aware Training: High Performance Extremely Low-bit Architecture Search
Mingzhu Shen, Feng Liang, Ruihao Gong +6
Quantization Neural Networks (QNN) have attracted a lot of attention due to their high efficiency. To enhance the quantization accuracy, prior works mainly focus on designing advan…
Efficient Bitwidth Search for Practical Mixed Precision Neural Network
Yuhang Li, Wei Wang, Haoli Bai +3
Network quantization has rapidly become one of the most widely used methods to compress and accelerate deep neural networks. Recent efforts propose to quantize weights and activati…
Towards Unified INT8 Training for Convolutional Neural Network
Feng Zhu, Ruihao Gong, Fengwei Yu +5
Recently low-bit (e.g., 8-bit) network quantization has been extensively studied to accelerate the inference. Besides inference, low-bit training with quantized gradients can furth…
Forward and Backward Information Retention for Accurate Binary Neural Networks
Haotong Qin, Ruihao Gong, Xianglong Liu +4
Weight and activation binarization is an effective approach to deep neural network compression and can accelerate the inference by leveraging bitwise operations. Although many bina…
Differentiable Soft Quantization: Bridging Full-Precision and Low-Bit Neural Networks
Ruihao Gong, Xianglong Liu, Shenghu Jiang +5
Hardware-friendly network quantization (e.g., binary/uniform quantization) can efficiently accelerate the inference and meanwhile reduce memory consumption of the deep neural netwo…
POI: Multiple Object Tracking with High Performance Detection and Appearance Feature
Fengwei Yu, Wenbo Li, Quanquan Li +3
Detection and learning based appearance feature play the central role in data association based multiple object tracking (MOT), but most recent MOT works usually ignore them and on…