{"id":13499080,"url":"https://github.com/wenxinxu/ResNeXt-in-tensorflow","last_synced_at":"2025-03-29T04:30:30.990Z","repository":{"id":97989826,"uuid":"82236363","full_name":"wenxinxu/ResNeXt-in-tensorflow","owner":"wenxinxu","description":"This is an implementation of ResNeXt (by Xie et al.) in tensorflow","archived":false,"fork":false,"pushed_at":"2018-02-18T23:27:13.000Z","size":1402,"stargazers_count":78,"open_issues_count":9,"forks_count":24,"subscribers_count":8,"default_branch":"master","last_synced_at":"2024-08-01T22:50:06.830Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"","language":"Python","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/wenxinxu.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,"governance":null,"roadmap":null,"authors":null}},"created_at":"2017-02-16T23:31:04.000Z","updated_at":"2024-01-04T16:11:32.000Z","dependencies_parsed_at":"2023-03-27T11:01:21.842Z","dependency_job_id":null,"html_url":"https://github.com/wenxinxu/ResNeXt-in-tensorflow","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/wenxinxu%2FResNeXt-in-tensorflow","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/wenxinxu%2FResNeXt-in-tensorflow/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/wenxinxu%2FResNeXt-in-tensorflow/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/wenxinxu%2FResNeXt-in-tensorflow/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/wenxinxu","download_url":"https://codeload.github.com/wenxinxu/ResNeXt-in-tensorflow/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":222455936,"owners_count":16987574,"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":"2024-07-31T22:00:28.154Z","updated_at":"2024-10-31T17:31:31.051Z","avatar_url":"https://github.com/wenxinxu.png","language":"Python","funding_links":[],"categories":["Papers\u0026Codes"],"sub_categories":["ResNext"],"readme":"# ResNeXt in Tensorflow\n\nThis is an implementation of [ResNeXt](https://arxiv.org/abs/1611.05431) in tensorflow. The tensorboard visualization of the detailed model structure (ResNeXt-29, 4x64d as example) is shown in [graph](https://github.com/wenxinxu/ResNeXt-in-tensorflow/blob/master/figure/graph.png), [block](https://github.com/wenxinxu/ResNeXt-in-tensorflow/blob/master/figure/graph_block.png), and [split](https://github.com/wenxinxu/ResNeXt-in-tensorflow/blob/master/figure/graph_block_split.png). \n\nI implemented the blocks with both methods in Figure 3b (split-\u003econcat). \n\n**Figure 3c (grouped convolutions) is not available yet. ** \n\ntensorflow 1.0.0 is needed here. They re-named lots of functions, so the previous versions won't work...\n\n## Quick guide\n\nYou can run cifar10_train.py to train a ResNeXt on cifar10 and see how it works from the screen output (the code will download the data for you if you don't have them yet). It’s better to speicify a version identifier before running, since the training logs, checkpoints, and error.csv file will be saved in the folder with name logs_$version. \n\n`python cifar10_train.py --version='test'`\n\nYou may alter the shape of the model via the hyper-parameters. Typically a ResNeXt is represented as 'ResNeXt-a, b*c'. \n\na is the total layer, which is defined by 9 * FLAGS.num_resnext_blocks + 2.\n\nb is the cardinality, which is defined by FLAGS.cardinality.\n\nc is the number of channels in each split, which is defined by FLAGS.block_unit_depth\n\nTo play with these hyper-parameters, you could either change inside the hyper-parameters.py or use the following commands:\n```\npython cifar10_train.py --version='test',num_resnext_blocks=3,cardinality=4\n```\nMore explanations about the hyper-parameters can be found [here](https://github.com/wenxinxu/resnet-in-tensorflow#hyper-parameters)\n\n### Files included\n1. hyper-parameters.py defines the hyper-parameters related to train, ResNeXt structure, data augmentation, etc.\n\n2. cifar10_input.py includes the data I/O, pre-processing of images and data augmentation\n\n3. resNeXt.py is the main body of ResNeXt network\n\n4. cifar10_train.py is responsible for the training and validation\n\n## TODO:\n1. Train the model and generate the learning curve\n2. Data augmentation: cv2 is not compatible w/ tensorflow 1.0.0, so I need to:\n\n    a. Wait for a new version of opencv\n\n    or\n\n    b. Use the queueRunner in tensorflow as data I/O and implement data augmentaion\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fwenxinxu%2FResNeXt-in-tensorflow","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fwenxinxu%2FResNeXt-in-tensorflow","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fwenxinxu%2FResNeXt-in-tensorflow/lists"}