{"id":13499019,"url":"https://github.com/KaimingHe/resnet-1k-layers","last_synced_at":"2025-03-29T03:32:13.943Z","repository":{"id":38360933,"uuid":"55462616","full_name":"KaimingHe/resnet-1k-layers","owner":"KaimingHe","description":"Deep Residual Networks with 1K Layers","archived":false,"fork":false,"pushed_at":"2017-05-24T17:35:48.000Z","size":4,"stargazers_count":915,"open_issues_count":1,"forks_count":251,"subscribers_count":66,"default_branch":"master","last_synced_at":"2025-03-28T08:09:39.340Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"","language":"Lua","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/KaimingHe.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2016-04-05T02:50:30.000Z","updated_at":"2025-03-26T06:48:27.000Z","dependencies_parsed_at":"2022-08-25T04:50:30.868Z","dependency_job_id":null,"html_url":"https://github.com/KaimingHe/resnet-1k-layers","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/KaimingHe%2Fresnet-1k-layers","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/KaimingHe%2Fresnet-1k-layers/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/KaimingHe%2Fresnet-1k-layers/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/KaimingHe%2Fresnet-1k-layers/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/KaimingHe","download_url":"https://codeload.github.com/KaimingHe/resnet-1k-layers/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":246135734,"owners_count":20729056,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2024-07-31T22:00:26.174Z","updated_at":"2025-03-29T03:32:13.911Z","avatar_url":"https://github.com/KaimingHe.png","language":"Lua","funding_links":[],"categories":["Papers\u0026Codes","Model Zoo"],"sub_categories":["PreActResNet","Convolutional Networks"],"readme":"# Deep Residual Networks with 1K Layers\n\nBy [Kaiming He](http://kaiminghe.com), [Xiangyu Zhang](https://scholar.google.com/citations?user=yuB-cfoAAAAJ\u0026hl=en), [Shaoqing Ren](http://home.ustc.edu.cn/~sqren/), [Jian Sun](http://research.microsoft.com/en-us/people/jiansun/).\n\nMicrosoft Research Asia (MSRA).\n\n## Table of Contents\n0. [Introduction](#introduction)\n0. [Notes](#notes)\n0. [Usage](#usage)\n\n\n\n## Introduction\n\nThis repository contains re-implemented code for the paper \"Identity Mappings in Deep Residual Networks\" (http://arxiv.org/abs/1603.05027). This work enables training quality **1k-layer** neural networks in a super simple way.\n\n*Acknowledgement*: This code is re-implemented by Xiang Ming from Xi'an Jiaotong Univeristy for the ease of release.\n\n**Seel Also:** Re-implementations of **ResNet-200 [a] on ImageNet** from Facebook AI Research (FAIR): https://github.com/facebook/fb.resnet.torch/tree/master/pretrained\n\nRelated papers:\n\n\t[a]\t@article{He2016,\n\t\t\tauthor = {Kaiming He and Xiangyu Zhang and Shaoqing Ren and Jian Sun},\n\t\t\ttitle = {Identity Mappings in Deep Residual Networks},\n\t\t\tjournal = {arXiv preprint arXiv:1603.05027},\n\t\t\tyear = {2016}\n\t\t}\n\t\n\t[b] @article{He2015,\n\t\t\tauthor = {Kaiming He and Xiangyu Zhang and Shaoqing Ren and Jian Sun},\n\t\t\ttitle = {Deep Residual Learning for Image Recognition},\n\t\t\tjournal = {arXiv preprint arXiv:1512.03385},\n\t\t\tyear = {2015}\n\t\t}\n\t\n\n\t\n## Notes\n\n0. This code is based on the implementation of Torch ResNets (https://github.com/facebook/fb.resnet.torch).\n\n0. The experiments in the paper were conducted in Caffe, whereas this code is re-implemented in Torch. We observed similar results within reasonable statistical variations.\n\n0. To fit the 1k-layer models into memory without modifying much code, we simply reduced the mini-batch size to 64, noting that results in the paper were obtained with a mini-batch size of 128. Less expectedly, the results with the mini-batch size of 64 are slightly better:\n\n\tmini-batch |CIFAR-10 test error (%): (median (mean+/-std))\n\t:---------:|:------------------:\n\t128 (as in [a]) | 4.92 (4.89+/-0.14)\n\t64 (as in this code)| **4.62** (4.69+/-0.20)\n\n0. Curves obtained by running this code with a mini-batch size of 64 (training loss: y-axis on the left; test error: y-axis on the right):\t\n![resnet1k](https://cloud.githubusercontent.com/assets/11435359/14414142/68714c82-ffc0-11e5-8b1b-657fdb3d96a6.png)\n\t\n## Usage\n\n0. Install Torch ResNets (https://github.com/facebook/fb.resnet.torch) following instructions therein.\n0. Add the file resnet-pre-act.lua from this repository to ./models.\n0. To train ResNet-1001 as of the form in [a]:\n```\nth main.lua -netType resnet-pre-act -depth 1001 -batchSize 64 -nGPU 2 -nThreads 4 -dataset cifar10 -nEpochs 200 -shareGradInput false\n```\n**Note**: ``shareGradInput=true'' is not valid for this model yet.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FKaimingHe%2Fresnet-1k-layers","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FKaimingHe%2Fresnet-1k-layers","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FKaimingHe%2Fresnet-1k-layers/lists"}