Filters








403 Hits in 4.5 sec

A Multigrid Method for Efficiently Training Video Models [article]

Chao-Yuan Wu, Ross Girshick, Kaiming He, Christoph Feichtenhofer, Philipp Krähenbühl
<span title="2020-06-10">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Following standard practice for training image models, video model training assumes a fixed mini-batch shape: a specific number of clips, frames, and spatial size. However, what is the optimal shape?  ...  Training competitive deep video models is an order of magnitude slower than training their counterpart image models.  ...  We would like to thank Haoqi Fan for helping with the code release and Ishan Nigam, Santhosh K. Ramakrishnan, and Xingyi Zhou for helpful comments on an earlier draft.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1912.00998v2">arXiv:1912.00998v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/ik6pfij57rco3logehwuiy7ioy">fatcat:ik6pfij57rco3logehwuiy7ioy</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200612002946/https://arxiv.org/pdf/1912.00998v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1912.00998v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Accelerating the Training of Video Super-Resolution Models [article]

Lijian Lin, Xintao Wang, Zhongang Qi, Ying Shan
<span title="2022-05-17">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Despite that convolution neural networks (CNN) have recently demonstrated high-quality reconstruction for video super-resolution (VSR), efficiently training competitive VSR models remains a challenging  ...  Training is accelerated by such a multigrid training strategy, as most of computation is performed on smaller spatial and shorter temporal shapes.  ...  In summary, we make the following contributions: -We propose a multigrid training strategy for efficient VSR training.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2205.05069v2">arXiv:2205.05069v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/gqnj2mfjdrdapbcqjwtylvvpd4">fatcat:gqnj2mfjdrdapbcqjwtylvvpd4</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220525014316/https://arxiv.org/pdf/2205.05069v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/06/0b/060bac73d0c6cab65b91f77364645a0afe142d38.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2205.05069v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Multigrid Predictive Filter Flow for Unsupervised Learning on Videos [article]

Shu Kong, Charless Fowlkes
<span title="2019-04-02">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We introduce multigrid Predictive Filter Flow (mgPFF), a framework for unsupervised learning on videos.  ...  We develop a multigrid coarse-to-fine modeling strategy that avoids the requirement of learning large filters to capture large displacement.  ...  Shu Kong personally thanks Teng Liu and Etthew Kong who initiated this research, and the academic uncle Alexei A. Efros for the encouragement and discussion.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1904.01693v1">arXiv:1904.01693v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/ssyaqyvl5bg7dgh4yml7zlziqu">fatcat:ssyaqyvl5bg7dgh4yml7zlziqu</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200828073030/https://arxiv.org/pdf/1904.01693v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/22/5d/225dc18db507147de068710839941900996a7329.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1904.01693v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Automatic Image Tagging Model Based on Multigrid Image Segmentation and Object Recognition

Woogyoung Jun, Yillbyung Lee, Byoung-Min Jun
<span title="">2014</span> <i title="Hindawi Limited"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/4xiibn6qsbh3pj5zdvqplxb37m" style="color: black;">Advances in Multimedia</a> </i> &nbsp;
Since there are still lots of limitations in automatic image tagging models, we propose efficient automatic image tagging model using multigrid based image segmentation and feature extraction method.  ...  Our method is tested with Corel dataset and the result showed that our model performance is efficient and effective compared to other models.  ...  We find out image segmentation technique and propose a multigrid based image segmentation method.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1155/2014/857682">doi:10.1155/2014/857682</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/ozpxmjtiabgwvbrev6t5vmgguy">fatcat:ozpxmjtiabgwvbrev6t5vmgguy</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200215072820/http://downloads.hindawi.com/journals/am/2014/857682.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/15/fd/15fde490e655ab86fda428ee0dee5de25b87bd9e.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1155/2014/857682"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="unlock alternate icon" style="background-color: #fb971f;"></i> hindawi.com </button> </a>

Learning across scales - A multiscale method for Convolution Neural Networks [article]

Eldad Haber, Lars Ruthotto, Elliot Holtham, Seong-Hwan Jun
<span title="2017-06-22">2017</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
network approximates the data-label relation for given training data.  ...  The second class of multiscale methods connects shallow and deep networks and leads to new training strategies that gradually increase the depths of the CNN while re-using parameters for initializations  ...  Such a process can be particularly efficient when considering the classification of videos on mobile devices where expanding the video to high resolution can be computationally prohibitive.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1703.02009v2">arXiv:1703.02009v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/f2phovve4bcrbhdxpznofaatja">fatcat:f2phovve4bcrbhdxpznofaatja</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200917232809/https://arxiv.org/pdf/1703.02009v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/d6/59/d659ef861254a1465c1c52df10b95c4c7aa501a2.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1703.02009v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Distributed Multigrid Neural Solvers on Megavoxel Domains [article]

Aditya Balu, Sergio Botelho, Biswajit Khara, Vinay Rao, Chinmay Hegde, Soumik Sarkar, Santi Adavani, Adarsh Krishnamurthy, Baskar Ganapathysubramanian
<span title="2021-04-29">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
First, we accelerate training a large model via a method analogous to the multigrid technique used in numerical linear algebra.  ...  This approach is deployed to train a generalized 3D Poisson solver that scales well to predict output full-field solutions up to the resolution of 512x512x512 for a high dimensional family of inputs.  ...  A multigrid-inspired training scheme for training the networks at higher resolutions.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2104.14538v1">arXiv:2104.14538v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/ctqujbs7sraidktqejuh33qonq">fatcat:ctqujbs7sraidktqejuh33qonq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210502084252/https://arxiv.org/pdf/2104.14538v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/7c/5e/7c5e6b769549a83a40d4b7174c2e86d1f3835da8.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2104.14538v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Learning to Optimize Multigrid PDE Solvers [article]

Daniel Greenfeld, Meirav Galun, Ron Kimmel, Irad Yavneh, Ronen Basri
<span title="2019-08-05">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We train a neural network once for the entire class of PDEs, using an efficient and unsupervised loss function.  ...  Constructing fast numerical solvers for partial differential equations (PDEs) is crucial for many scientific disciplines. A leading technique for solving large-scale PDEs is using multigrid methods.  ...  These methods require separate training for each new equation.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1902.10248v3">arXiv:1902.10248v3</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/dd4njs76wnhw7auvcw6xc6izlq">fatcat:dd4njs76wnhw7auvcw6xc6izlq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20191013064752/https://arxiv.org/pdf/1902.10248v1.pdf" title="fulltext PDF download [not primary version]" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <span style="color: #f43e3e;">&#10033;</span> <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/bd/ea/bdea7d67840f6b8a71ba47b00e6423212a1e3654.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1902.10248v3" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Layer-Parallel Training with GPU Concurrency of Deep Residual Neural Networks via Nonlinear Multigrid [article]

Andrew C. Kirby, Siddharth Samsi, Michael Jones, Albert Reuther, Jeremy Kepner, Vijay Gadepally
<span title="2020-08-30">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
A Multigrid Full Approximation Storage algorithm for solving Deep Residual Networks is developed to enable neural network parallelized layer-wise training and concurrent computational kernel execution  ...  This work demonstrates a 10.2x speedup over traditional layer-wise model parallelism techniques using the same number of compute units.  ...  MULTIGRID METHODS Multigrid (MG) methods are iterative algorithms that accelerate the solution convergence of a system of equations.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2007.07336v2">arXiv:2007.07336v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/oababftxvbbcjhprxv3ujmceye">fatcat:oababftxvbbcjhprxv3ujmceye</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200909072735/https://arxiv.org/pdf/2007.07336v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2007.07336v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Optimization-Based Algebraic Multigrid Coarsening Using Reinforcement Learning [article]

Ali Taghibakhshi, Scott MacLachlan, Luke Olson, Matthew West
<span title="2022-01-04">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In this paper, we consider the complementary problem of learning to coarsen graphs for a multigrid solver, a necessary step in developing fully learnable AMG methods.  ...  The efficiency of the multigrid solver depends critically on this selection and many selection methods have been developed over the years.  ...  While dueling architectures have benefits for training, they are also valuable for graph problems where the Q and V values are dependent on the size of the graph but A is not, as is the case for multigrid  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2106.01854v3">arXiv:2106.01854v3</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/2wz7eozrlfglfpi777zzzwyjpi">fatcat:2wz7eozrlfglfpi777zzzwyjpi</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210615050209/https://arxiv.org/pdf/2106.01854v1.pdf" title="fulltext PDF download [not primary version]" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <span style="color: #f43e3e;">&#10033;</span> <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/a7/1c/a71ca120b8666199172cb7627bc16b28bb03e6b4.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2106.01854v3" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Multigrid Neural Memory [article]

Tri Huynh, Michael Maire, Matthew R. Walter
<span title="2020-08-15">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Such networks have an implicit capacity for internal attention; augmented with memory, they learn to read and write specific memory locations in a dynamic data-dependent manner.  ...  Our hierarchical spatial organization, parameterized convolutionally, permits efficient instantiation of large-capacity memories, while multigrid topology provides short internal routing pathways, allowing  ...  We thank Gordon Kindlmann for his support in pursuing this project, Chau Huynh for her help with the code, and Pedro Savarese and Hai Nguyen for fruitful discussions.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1906.05948v4">arXiv:1906.05948v4</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/hshirzwqa5evjn5ti3o7m3pyke">fatcat:hshirzwqa5evjn5ti3o7m3pyke</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200819004142/https://arxiv.org/pdf/1906.05948v4.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1906.05948v4" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Parallel Training of GRU Networks with a Multi-Grid Solver for Long Sequences [article]

Gordon Euhyun Moon, Eric C. Cyr
<span title="2022-03-07">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In this paper, we present a novel parallel training scheme (called parallel-in-time) for GRU based on a multigrid reduction in time (MGRIT) solver.  ...  Experimental results on the HMDB51 dataset, where each video is an image sequence, demonstrate that the new parallel training scheme achieves up to 6.5× speedup over a serial approach.  ...  As the size of each image in the video is too large to directly use it as an input for a RNN-based model, we use a pre-trained CNN model to generate low-dimensional input features for the GRU network.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2203.04738v1">arXiv:2203.04738v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/cd7lddptwvclphfo3qy7aggn5y">fatcat:cd7lddptwvclphfo3qy7aggn5y</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220313051945/https://arxiv.org/pdf/2203.04738v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/0c/7f/0c7fa4d8935a05afadebe0a2daf21ef1d91daf49.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2203.04738v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Connections between Numerical Algorithms for PDEs and Neural Networks [article]

Tobias Alt, Karl Schrader, Matthias Augustin, Pascal Peter, Joachim Weickert
<span title="2022-03-21">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Moreover, we present U-net architectures that implement multigrid techniques for learning efficient solutions of partial differential equation models, and motivate uncommon design choices such as trainable  ...  Our considerations serve as a basis for explaining the success of popular neural architectures and provide a blueprint for developing new mathematically well-founded neural building blocks.  ...  They belong to the most efficient numerical methods for PDE-related problems and have been successfully applied to various tasks such as image denoising [13] , inpainting [70] , video compression [62  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2107.14742v2">arXiv:2107.14742v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/6nhb6b6ryjcpfizgi2bywdcxdq">fatcat:6nhb6b6ryjcpfizgi2bywdcxdq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220327165704/https://arxiv.org/pdf/2107.14742v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/11/28/1128b76da2bdc5e0275361b5bc3d4d8c38fe9038.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2107.14742v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Stochastic Backpropagation: A Memory Efficient Strategy for Training Video Models [article]

Feng Cheng, Mingze Xu, Yuanjun Xiong, Hao Chen, Xinyu Li, Wei Li, Wei Xia
<span title="2022-03-31">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We propose a memory efficient method, named Stochastic Backpropagation (SBP), for training deep neural networks on videos.  ...  Experiments show that SBP can be applied to a wide range of models for video tasks, leading to up to 80.0% GPU memory saving and 10% training speedup with less than 1% accuracy drop on action recognition  ...  However, how to integrate it with other efficient model training strategies, such as Multigrid [51] , to speed up the training is still not well explored. Potential Negative Impact.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2203.16755v1">arXiv:2203.16755v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/uq3zd76ijzar5etuigndch7scu">fatcat:uq3zd76ijzar5etuigndch7scu</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220521101104/https://arxiv.org/pdf/2203.16755v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/83/2a/832aafe1d5b7d0476dfafdb47f39af5c84d8c782.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2203.16755v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

In Defense of Image Pre-Training for Spatiotemporal Recognition [article]

Xianhang Li, Huiyu Wang, Chen Wei, Jieru Mei, Alan Yuille, Yuyin Zhou, Cihang Xie
<span title="2022-05-03">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Image pre-training, the current de-facto paradigm for a wide range of visual tasks, is generally less favored in the field of video recognition.  ...  The code and models are available at https://github.com/UCSC-VLAA/Image-Pretraining-for-Video.  ...  Given the appearance prior, our proposed method aims to perform efficient pre-training and achieve an acceptable cost balance compared with the training from scratch. Efficient 3D Video Recognition.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2205.01721v1">arXiv:2205.01721v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/p7q45kbdlzgvdj5hgzdkbnuiyy">fatcat:p7q45kbdlzgvdj5hgzdkbnuiyy</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220506220240/https://arxiv.org/pdf/2205.01721v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/16/83/1683b02fc3cf4227ad9398dd581c0fd2e10a9493.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2205.01721v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Feature Space Optimization for Semantic Video Segmentation

Abhijit Kundu, Vibhav Vineet, Vladlen Koltun
<span title="">2016</span> <i title="IEEE"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/ilwxppn4d5hizekyd3ndvy2mii" style="color: black;">2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)</a> </i> &nbsp;
Shen, A. van dan Hengel, and I. Reid. Effi- video segmentation with exemplar-based object rea- cient piecewise training of deep structured models for soning.  ...  Semantic video segmentation: Exploring inference ef- [22] J. W. Ruge and K. Stüben. Algebraic multigrid. In ficiency. In ISOCC, 2015. 5, 6, 7 Multigrid Methods.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/cvpr.2016.345">doi:10.1109/cvpr.2016.345</a> <a target="_blank" rel="external noopener" href="https://dblp.org/rec/conf/cvpr/KunduVK16.html">dblp:conf/cvpr/KunduVK16</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/vup36eavdrfgroapaw3r5hscru">fatcat:vup36eavdrfgroapaw3r5hscru</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20161213092043/http://www.cv-foundation.org:80/openaccess/content_cvpr_2016/papers/Kundu_Feature_Space_Optimization_CVPR_2016_paper.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/98/d3/98d3b78336c3ab4196d23d01b3e07cd86a44d091.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/cvpr.2016.345"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> ieee.com </button> </a>
&laquo; Previous Showing results 1 &mdash; 15 out of 403 results