An Adaptive and Momental Bound Method for Stochastic Learning

Ding, Jianbang; Ren, Xuancheng; Luo, Ruixuan; Sun, Xingming

doi:10.48550/arxiv.1910.12249

Cited by 9 publications

(13 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We implemented UCondDGCN on the PyTorch platform [33] and conducted experiments on a single NVIDIA TITAN V GPU. We optimized the model by the AdaMod optimizer [10] for 110 epochs with a batch size of 256, in which the learning rate was initially Table 1: Quantitative comparisons with state-of-the-art methods on Human3.6M under protocol #1 and protocol #2, where methods marked with † are video-based; T denotes the number of input frames; and CPN and HR-Net denote the input 2D poses are estimated by [5] and [41], respectively. The best and second-best results are marked in bold and underlined, respectively.…”

Section: Methodsmentioning

confidence: 99%

Conditional Directed Graph Convolution for 3D Human Pose Estimation

Zhang

Zhan

et al. 2021

Proceedings of the 29th ACM International Conference on Multimedia

View full text Add to dashboard Cite

Input 2D poseEstimated 3D pose by UGCN Estimated 3D pose by our method Figure 1: Given a sequence of 2D human poses estimated by an off-the-shelf 2D pose estimator, e.g., HR-Net [41], our method can produce more precise 3D poses compared with state-of-the-art approach, UGCN [49]. We show the results under three different viewpoints as indicated by the 3D orientation markers. And ground-truth 3D poses are shown in gray as a reference.

show abstract

Section: Methodsmentioning

confidence: 99%

Conditional Directed Graph Convolution for 3D Human Pose Estimation

Zhang

Zhan

et al. 2021

Proceedings of the 29th ACM International Conference on Multimedia

View full text Add to dashboard Cite

show abstract

“…Some research shows that the learning rate is one of the most important hyperparameters for deep learning because it directly affects the gradient descents and controls the speed of network convergence to the point of global minima by navigation through a non-convex loss surface [6]. How find an optimal learning rate becomes a huge challenge for our experiments.…”

Section: Learning Rate Schedulermentioning

confidence: 99%

CV 3315 Is All You Need : Semantic Segmentation Competition

Akide¹,

Wang²

2022

Preprint

View full text Add to dashboard Cite

This competition 1 focus on Urban-Sense Segmentation based on the vehicle camera view. Class highly unbalanced Urban-Sense images dataset challenge the existing solutions and further studies. Deep Conventional neural network-based semantic segmentation methods such as encoder-decoder architecture and multi-scale and pyramidbased approaches become flexible solutions applicable to real-world applications. In this competition, we mainly review the literature and conduct experiments on transformerdriven methods especially SegFormer [26], to achieve an optimal trade-off between performance and efficiency. For example, SegFormer-B0 achieved 74.6% mIoU with the smallest FLOPS, 15.6G, and the largest model, SegFormer-B5 archived 80.2% mIoU. According to multiple factors, including individual case failure analysis, individual class performance, training pressure and efficiency estimation, the final candidate model for the competition is SegFormer-B2 with 50.6 GFLOPS and 78.5% mIoU evaluated on the testing set 2 .

show abstract

“…As for optimizers, we used very popular Adam optimizer. We also tried recently introduced diffGrad [16] and Adamod [17] but they didn't provide us with any improvements (we didn't perform a hyper-parameter tuning). Comparison of optimizers depicted on Fig.…”

Section: Model Optimizationmentioning

confidence: 99%

Deep Residual 3D U-Net for Joint Segmentation and Texture Classification of Nodules in Lung

Rassadin¹

2020

Lecture Notes in Computer Science

View full text Add to dashboard Cite

In this work we present a method for lung nodules segmentation, their texture classification and subsequent follow-up recommendation from the CT image of lung. Our method consists of neural network model based on popular U-Net architecture family but modified for the joint nodule segmentation and its texture classification tasks and an ensemble-based model for the followup recommendation. This solution was evaluated within the LNDb medical imaging challenge and produced the best nodule segmentation result on the final leaderboard.

show abstract

An Adaptive and Momental Bound Method for Stochastic Learning

Cited by 9 publications

References 0 publications

Conditional Directed Graph Convolution for 3D Human Pose Estimation

Conditional Directed Graph Convolution for 3D Human Pose Estimation

CV 3315 Is All You Need : Semantic Segmentation Competition

Deep Residual 3D U-Net for Joint Segmentation and Texture Classification of Nodules in Lung

Contact Info

Product

Resources

About