{"id":20841757,"url":"https://github.com/jahongir7174/efficientnetv2","last_synced_at":"2025-05-08T22:29:30.210Z","repository":{"id":45449260,"uuid":"357181525","full_name":"jahongir7174/EfficientNetV2","owner":"jahongir7174","description":"EfficientNetV2 implementation using PyTorch","archived":false,"fork":false,"pushed_at":"2022-07-12T05:21:41.000Z","size":78433,"stargazers_count":129,"open_issues_count":1,"forks_count":27,"subscribers_count":2,"default_branch":"master","last_synced_at":"2025-03-31T18:52:02.494Z","etag":null,"topics":["efficientnetv2","imagenet","pytorch","training"],"latest_commit_sha":null,"homepage":"https://arxiv.org/abs/2104.00298","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/jahongir7174.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2021-04-12T12:16:37.000Z","updated_at":"2025-03-10T10:35:53.000Z","dependencies_parsed_at":"2022-07-14T01:10:37.801Z","dependency_job_id":null,"html_url":"https://github.com/jahongir7174/EfficientNetV2","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/jahongir7174%2FEfficientNetV2","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jahongir7174%2FEfficientNetV2/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jahongir7174%2FEfficientNetV2/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jahongir7174%2FEfficientNetV2/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/jahongir7174","download_url":"https://codeload.github.com/jahongir7174/EfficientNetV2/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":253157320,"owners_count":21863094,"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":["efficientnetv2","imagenet","pytorch","training"],"created_at":"2024-11-18T01:21:37.834Z","updated_at":"2025-05-08T22:29:30.186Z","avatar_url":"https://github.com/jahongir7174.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"[EfficientNetV2](https://arxiv.org/abs/2104.00298) implementation using PyTorch\n\n### Steps\n\n* `imagenet` path by changing `data_dir` in `main.py`\n* `bash ./main.sh $ --train` for training model, `$` is number of GPUs\n* `EfficientNet` class in `nets/nn.py` for different versions\n\n### Note\n\n* the default training configuration is for `EfficientNetV2-S`\n\n### Parameters and FLOPS\n\n* `python main.py --benchmark`\n\n```\nNumber of parameters: 21458488\nTime per operator type:\n        1504.95 ms.    80.5982%. Conv\n        225.509 ms.    12.0772%. Sigmoid\n        115.112 ms.     6.1649%. Mul\n        12.7341 ms.   0.681982%. Add\n        7.50523 ms.   0.401946%. AveragePool\n        1.40185 ms.  0.0750768%. FC\n      0.0112697 ms. 0.000603555%. Flatten\n        1867.22 ms in Total\nFLOP per operator type:\n        16.7287 GFLOP.     99.708%. Conv\n      0.0412707 GFLOP.   0.245986%. Mul\n     0.00516096 GFLOP.  0.0307609%. Add\n       0.002561 GFLOP.  0.0152643%. FC\n        16.7777 GFLOP in Total\nFeature Memory Read per operator type:\n        291.409 MB.    51.8224%. Mul\n        224.497 MB.    39.9231%. Conv\n        41.2877 MB.    7.34234%. Add\n        5.12912 MB.   0.912131%. FC\n        562.323 MB in Total\nFeature Memory Written per operator type:\n        165.083 MB.    50.2087%. Mul\n        143.062 MB.    43.5114%. Conv\n        20.6438 MB.    6.27867%. Add\n          0.004 MB. 0.00121657%. FC\n        328.793 MB in Total\nParameter Memory per operator type:\n        79.9537 MB.    93.9773%. Conv\n          5.124 MB.    6.02273%. FC\n              0 MB.          0%. Add\n              0 MB.          0%. Mul\n        85.0777 MB in Total\n```\n\n### Results\n\n* `python main.py --test` for trained model testing\n\n|       name       | resolution | acc@1 | acc@5 | #params |  FLOPS  | resample | training loss |\n|:----------------:|:----------:|:-----:|:-----:|:-------:|:-------:|---------:|--------------:|\n| EfficientNetV2-S |  384x384   | 83.9  | 96.7  |  21.46  | 16.7777 | BILINEAR |  CrossEntropy |\n| EfficientNetV2-S |  384x384   |   -   |   -   |  21.46  | 16.7777 | BILINEAR |      PolyLoss |\n| EfficientNetV2-M |     -      |   -   |   -   |    -    |    -    |        - |             - |\n| EfficientNetV2-L |     -      |   -   |   -   |    -    |    -    |        - |             - |\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjahongir7174%2Fefficientnetv2","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fjahongir7174%2Fefficientnetv2","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjahongir7174%2Fefficientnetv2/lists"}