{"id":15600405,"url":"https://github.com/avik-pal/deeplearningbenchmarks","last_synced_at":"2026-03-17T09:33:32.019Z","repository":{"id":102318132,"uuid":"140796028","full_name":"avik-pal/DeepLearningBenchmarks","owner":"avik-pal","description":"Benchmarks across Deep Learning Frameworks in Julia and Python","archived":false,"fork":false,"pushed_at":"2020-02-26T05:55:52.000Z","size":737,"stargazers_count":25,"open_issues_count":2,"forks_count":0,"subscribers_count":6,"default_branch":"master","last_synced_at":"2025-09-09T13:06:11.866Z","etag":null,"topics":["benchmark","computer-vision","conv2d","flux","gpu","julia","machine-learning","pytorch"],"latest_commit_sha":null,"homepage":null,"language":"Julia","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/avik-pal.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,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2018-07-13T04:15:43.000Z","updated_at":"2025-01-11T01:31:50.000Z","dependencies_parsed_at":null,"dependency_job_id":"243b8c75-c520-46ba-be37-30e4d19b50c0","html_url":"https://github.com/avik-pal/DeepLearningBenchmarks","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/avik-pal/DeepLearningBenchmarks","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/avik-pal%2FDeepLearningBenchmarks","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/avik-pal%2FDeepLearningBenchmarks/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/avik-pal%2FDeepLearningBenchmarks/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/avik-pal%2FDeepLearningBenchmarks/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/avik-pal","download_url":"https://codeload.github.com/avik-pal/DeepLearningBenchmarks/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/avik-pal%2FDeepLearningBenchmarks/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":30620740,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-03-17T08:10:05.930Z","status":"ssl_error","status_checked_at":"2026-03-17T08:10:04.972Z","response_time":56,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"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":["benchmark","computer-vision","conv2d","flux","gpu","julia","machine-learning","pytorch"],"created_at":"2024-10-03T02:02:45.416Z","updated_at":"2026-03-17T09:33:32.003Z","avatar_url":"https://github.com/avik-pal.png","language":"Julia","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Popular Computer Vision Model Benchmarks\n\n## Input Dimensions\n1. Batch Size = 8, Image = 3 x 224 x 224 (IF NOTHING SPECIFIED / CPU USED)\n2. Batch Size = 4, Image = 3 x 224 x 224\n    * Resnet 101\n    * Resnet 152\n\n## GPU USED --- Titan 1080Ti 12 GB\n|Model|Framework|Forward Pass|Backward Pass|Total Time|Inference|\n|:---:|:---:|:---:|:---:|:---:|:---:|\n|VGG16|Pytorch 0.4.1|__0.0245 s__|__0.0606 s__|__0.0852 s__|__0.0234 s__|\n||Flux 0.6.8+|0.0287 s|0.0760 s|0.1047 s|0.0288 s|\n|VGG16 BN|Pytorch 0.4.1|__0.0271 s__|__0.0672 s__|__0.0943 s__|__0.0273 s__|\n||Flux 0.6.8+|0.0333 s|0.0818 s|0.1151 s|0.0327 s|\n|VGG19|Pytorch 0.4.1|__0.0281 s__|__0.0741 s__|__0.1021 s__|__0.0280 s__|\n||Flux 0.6.8+|0.0355 s|0.0923 s|0.1278 s|0.0356 s|\n|VGG19 BN|Pytorch 0.4.1|__0.0321 s__|__0.0812 s__|__0.1134 s__|__0.0325 s__|\n||Flux 0.6.8+|0.0377 s|0.0965 s|0.1342 s|0.0371 s|\n|Resnet18|Pytorch 0.4.1|__0.0064 s__|__0.0125 s__|__0.0190 s__|__0.0050 s__|\n||Flux 0.6.8+|0.0079 s|0.0218 s|0.0297 s|0.0079 s|\n|Resnet34|Pytorch 0.4.1|__0.0092 s__|__0.0216 s__|__0.0307 s__|__0.0092 s__|\n||Flux 0.6.8+|0.0137 s|0.0313 s|0.0450 s|0.0151 s|\n|Resnet50|Pytorch 0.4.1|__0.0155 s__|__0.0351 s__|__0.0506 s__|0.0152 s|\n||Flux 0.6.8+|0.0205 s|0.1795 s|0.2000 s|-|\n|Resnet101|Pytorch 0.4.1|0.0297 s|__0.0379 s__|__0.0676 s__|0.0298 s|\n||Flux 0.6.8+|__0.0215 s__|0.0616 s|0.0831 s|__0.0208 s__|\n|Resnet152|Pytorch 0.4.1|0.0431 s|__0.05337 s__|__0.0965 s__|0.0429 s|\n||Flux 0.6.8+|__0.0308 s__|0.0807 s|0.1115 s|__0.0298 s__|\n\n## CPU USED --- Intel(R) Xeon(R) Silver 4114 CPU @ 2.20GHz\n|Model|Framework|Forward Pass|Backward Pass|Total Time|Inference|\n|:---:|:---:|:---:|:---:|:---:|:---:|\n|VGG16|Pytorch 0.4.1|__6.6024 s__|__9.4336 s__|__16.036 s__|__6.4216 s__|\n||Flux 0.6.8+|10.458 s|10.245 s|20.703 s|10.111 s|\n|VGG16 BN|Pytorch 0.4.1|__7.0793 s__|__9.0536 s__|__16.132 s__|__6.7909 s__|\n||Flux 0.6.8+|29.633 s|18.649 s|49.282 s|24.047 s|\n|VGG19|Pytorch 0.4.1|__8.3075 s__|__10.899 s__|__19.207 s__|__8.0593 s__|\n||Flux 0.6.8+|12.226 s|12.457 s|24.683 s|12.029 s|\n|VGG19 BN|Pytorch 0.4.1|__8.7794 s__|__12.739 s__|__21.519 s__|__8.4044 s__|\n||Flux 0.6.8+|28.518 s|21.464 s|49.982 s|22.649 s|\n\u003c!-- |Resnet18|Pytorch 0.4.1|||||\n||Flux 0.6.8+|||||\n|Resnet34|Pytorch 0.4.1|||||\n||Flux 0.6.8+|||||\n|Resnet50|Pytorch 0.4.1|||||\n||Flux 0.6.8+|||||\n|Resnet101|Pytorch 0.4.1|||||\n||Flux 0.6.8+|||||\n|Resnet152|Pytorch 0.4.1|||||\n||Flux 0.6.8+||||| --\u003e\n\n# Individual Layer Benchmarks\n\n## Layer Descriptions\n1. Conv3x3/1 = Conv2d, 3x3 Kernel, 1x1 Padding, 1x1 Stride\n2. Conv5x5/1 = Conv2d, 5x5 Kernel, 2x2 Padding, 1x1 Stride\n3. Conv3x3/2 = Conv2d, 3x3 Kernel, 1x1 Padding, 2x2 Stride\n4. Conv5x5/2 = Conv2d, 5x5 Kernel, 2x2 Padding, 2x2 Stride\n5. Dense = 1024 =\u003e 512\n6. BatchNorm = BatchNorm2d\n\n## GPU USED --- Titan 1080Ti 12 GB\n|Layer|Framework|Forward Pass|Backward Pass|Total Time|\n|:---:|:---:|:---:|:---:|:---:|\n|Conv3x3/1|Pytorch 0.4.1|0.2312 ms|__0.5359 ms__|__0.7736 ms__|\n||Flux 0.6.8+|__0.1984 ms__|0.7640 ms|0.9624 ms|\n|Conv5x5/1|Pytorch 0.4.1|0.2667 ms|__0.5345 ms__|__0.8299 ms__|\n||Flux 0.6.8+|__0.2065 ms__|0.8075 ms|1.014 ms|\n|Conv3x3/2|Pytorch 0.4.1|0.1170 ms|__0.2203 ms__|__0.3376 ms__|\n||Flux 0.6.8+|__0.0927 ms__|0.5988 ms|0.6915 ms|\n|Conv5x5/2|Pytorch 0.4.1|0.1233 ms|__0.2162 ms__|__0.3407 ms__|\n||Flux 0.6.8+|__0.0941 ms__|0.6515 ms|0.7456 ms|\n|Dense|Pytorch 0.4.1|0.0887 ms|__0.1523 ms__|__0.2411 ms__|\n||Flux 0.6.8+|__0.0432 ms__|0.2044 ms|0.2476 ms|\n|BatchNorm|Pytorch 0.4.1|__0.1096 ms__|__0.1999 ms__|__0.3095 ms__|\n||Flux 0.6.8+|0.2211 ms|0.2849 ms|0.5060 ms|\n\n\u003c!-- ## CPU USED --- Intel(R) Xeon(R) Silver 4114 CPU @ 2.20GHz\n|Layer|Framework|Forward Pass|Backward Pass|Total Time|\n|:---:|:---:|:---:|:---:|:---:|\n|Conv3x3/1|Pytorch 0.4.1||||\n||Flux 0.6.8+||||\n|Conv5x5/1|Pytorch 0.4.1||||\n||Flux 0.6.8+||||\n|Conv3x3/2|Pytorch 0.4.1||||\n||Flux 0.6.8+||||\n|Conv5x5/2|Pytorch 0.4.1||||\n||Flux 0.6.8+||||\n|Dense|Pytorch 0.4.1||||\n||Flux 0.6.8+||||\n|BatchNorm|Pytorch 0.4.1||||\n||Flux 0.6.8+|||| --\u003e\n\n# NOTE\n\nTo reproduce the benchmarks checkout `Flux 0.6.8+` __avik-pal/cudnn_batchnorm__ and `CuArrays` __master__.\nSince the Batchnorm GPU is broken for Flux 0.6.8+ master so we cannot perform the benchmarks using that.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Favik-pal%2Fdeeplearningbenchmarks","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Favik-pal%2Fdeeplearningbenchmarks","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Favik-pal%2Fdeeplearningbenchmarks/lists"}