{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/adabatch-adaptive-batch-sizes-for-training","title":"AdaBatch: Adaptive Batch Sizes for Training Deep Neural Networks","arxiv_id":"1712.02029","date":"2017-12-06","proceeding":null,"authors":["Aditya Devarakonda","Maxim Naumov","Michael Garland"],"abstract":"Training deep neural networks with Stochastic Gradient Descent, or its\nvariants, requires careful choice of both learning rate and batch size. While\nsmaller batch sizes generally converge in fewer training epochs, larger batch\nsizes offer more parallelism and hence better computational efficiency. We have\ndeveloped a new training approach that, rather than statically choosing a\nsingle batch size for all epochs, adaptively increases the batch size during\nthe training process. Our method delivers the convergence rate of small batch\nsizes while achieving performance similar to large batch sizes. We analyse our\napproach using the standard AlexNet, ResNet, and VGG networks operating on the\npopular CIFAR-10, CIFAR-100, and ImageNet datasets. Our results demonstrate\nthat learning with adaptive batch sizes can improve performance by factors of\nup to 6.25 on 4 NVIDIA Tesla P100 GPUs while changing accuracy by less than 1%\nrelative to training with fixed batch sizes.","url_abs":"http://arxiv.org/abs/1712.02029v2","url_pdf":"http://arxiv.org/pdf/1712.02029v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"adabatch-adaptive-batch-sizes-for-training","repo_url":"https://github.com/GXU-GMU-MICCAI/AdaBatch-numerical-experiments","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"computational-efficiency","task_name":"Computational Efficiency"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"bottleneck-residual-block","method_name":"Bottleneck Residual Block"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"},{"method_slug":"grouped-convolution","method_name":"Grouped Convolution"},{"method_slug":"kaiming-initialization","method_name":"Kaiming Initialization"},{"method_slug":"local-response-normalization","method_name":"Local Response Normalization"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"residual-block","method_name":"Residual Block"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1712.02029","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}