{"id":18527903,"url":"https://github.com/nemonameless/pytorch_summary","last_synced_at":"2025-05-14T21:33:19.141Z","repository":{"id":118413180,"uuid":"179127667","full_name":"nemonameless/pytorch_summary","owner":"nemonameless","description":"Calculate parameters quantity, memory, and flops in simple and complex way.","archived":false,"fork":false,"pushed_at":"2019-03-20T14:04:45.000Z","size":29,"stargazers_count":2,"open_issues_count":0,"forks_count":0,"subscribers_count":2,"default_branch":"master","last_synced_at":"2025-02-17T06:30:26.888Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":false,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/nemonameless.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,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2019-04-02T17:35:20.000Z","updated_at":"2022-12-27T10:18:36.000Z","dependencies_parsed_at":"2023-03-13T12:45:10.062Z","dependency_job_id":null,"html_url":"https://github.com/nemonameless/pytorch_summary","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/nemonameless%2Fpytorch_summary","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/nemonameless%2Fpytorch_summary/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/nemonameless%2Fpytorch_summary/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/nemonameless%2Fpytorch_summary/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/nemonameless","download_url":"https://codeload.github.com/nemonameless/pytorch_summary/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":254231380,"owners_count":22036357,"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":"2024-11-06T17:56:16.393Z","updated_at":"2025-05-14T21:33:19.116Z","avatar_url":"https://github.com/nemonameless.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# PyTorch_summary\nPyTorch summary:\n\nCalculate parameters quantity, memory, run-time, and flops in simple and complex way.\n\n# NEW\n\n- Support run-time test.\n- Support both GPU and CPU for **simple mode** and **complex mode**\n- PyTorch 1.0 and 0.4 supported.\n\n## Useage:\nAdd your model to summary.py \n```\n###### add your model ######\nfrom torchvision import models\nmodel = models.resnet50() \n############################\n```\nRun summary.py\n```\npython3 summary.py --mod simple --gpu -1 --size 3,224,224\n--mod simple|complex|val \n--gpu (gpu id, -1:use cpu)\n--size (input size, split with ,)\n--runtime (if runtime is -1: disable runtime\u003e=1 iters for runtime test.)\n```\n\n## Results:\n\n### simplesum mode:\n\n```\n + Number of params: 25.56M\n + Number of FLOPs: 4.11G\n```\n### run-time test:\n```\nTesting runtime using TITAN X (Pascal)\nRun-Time: 141.8995 ms\n```\n\n### complexsum mode:\n\nfollowing https://github.com/ceykmc/pytorch_model_summary , and add the support to bilinear layer.\n\n```\n module name input shape output shape parameter quantity inference memory(MB) MAdd duration percent\n0 conv1_Conv2d 3 224 224 64 112 112 9408 3.06MB 235,225,088 6.32%\n1 bn1_BatchNorm2d 64 112 112 64 112 112 128 3.06MB 3,211,264 0.91%\n2 relu_ReLU 64 112 112 64 112 112 0 3.06MB 802,816 0.94%\n3 maxpool_MaxPool2d 64 112 112 64 56 56 0 0.77MB 1,605,632 3.58%\n4 layer1.0.conv1_Conv2d 64 56 56 64 56 56 4096 0.77MB 25,489,408 0.45%\n5 layer1.0.bn1_BatchNorm2d 64 56 56 64 56 56 128 0.77MB 802,816 0.21%\n6 layer1.0.conv2_Conv2d 64 56 56 64 56 56 36864 0.77MB 231,010,304 2.75%\n7 layer1.0.bn2_BatchNorm2d 64 56 56 64 56 56 128 0.77MB 802,816 0.22%\n8 layer1.0.conv3_Conv2d 64 56 56 256 56 56 16384 3.06MB 101,957,632 1.38%\n9 layer1.0.bn3_BatchNorm2d 256 56 56 256 56 56 512 3.06MB 3,211,264 0.78%\n10 layer1.0.relu_ReLU 256 56 56 256 56 56 0 3.06MB 802,816 0.53%\n11 layer1.0.downsample 64 56 56 256 56 56 16896 6.12MB 105,168,896 2.13%\n12 layer1.1 256 56 56 256 56 56 70400 12.25MB 441,147,392 6.28%\n13 layer1.2 256 56 56 256 56 56 70400 12.25MB 441,147,392 6.24%\n14 layer2.0.conv1_Conv2d 256 56 56 128 56 56 32768 1.53MB 205,119,488 1.94%\n15 layer2.0.bn1_BatchNorm2d 128 56 56 128 56 56 256 1.53MB 1,605,632 0.20%\n16 layer2.0.conv2_Conv2d 128 56 56 128 28 28 147456 0.38MB 231,110,656 2.65%\n17 layer2.0.bn2_BatchNorm2d 128 28 28 128 28 28 256 0.38MB 401,408 0.08%\n18 layer2.0.conv3_Conv2d 128 28 28 512 28 28 65536 1.53MB 102,359,040 0.96%\n19 layer2.0.bn3_BatchNorm2d 512 28 28 512 28 28 1024 1.53MB 1,605,632 0.25%\n20 layer2.0.relu_ReLU 512 28 28 512 28 28 0 1.53MB 401,408 0.26%\n21 layer2.0.downsample 256 56 56 512 28 28 132096 3.06MB 206,725,120 1.93%\n22 layer2.1 512 28 28 512 28 28 280064 6.12MB 438,939,648 4.36%\n23 layer2.2 512 28 28 512 28 28 280064 6.12MB 438,939,648 4.47%\n24 layer2.3 512 28 28 512 28 28 280064 6.12MB 438,939,648 4.44%\n25 layer3.0.conv1_Conv2d 512 28 28 256 28 28 131072 0.77MB 205,320,192 1.48%\n26 layer3.0.bn1_BatchNorm2d 256 28 28 256 28 28 512 0.77MB 802,816 0.14%\n27 layer3.0.conv2_Conv2d 256 28 28 256 14 14 589824 0.19MB 231,160,832 2.02%\n28 layer3.0.bn2_BatchNorm2d 256 14 14 256 14 14 512 0.19MB 200,704 0.08%\n29 layer3.0.conv3_Conv2d 256 14 14 1024 14 14 262144 0.77MB 102,559,744 0.77%\n30 layer3.0.bn3_BatchNorm2d 1024 14 14 1024 14 14 2048 0.77MB 802,816 0.28%\n31 layer3.0.relu_ReLU 1024 14 14 1024 14 14 0 0.77MB 200,704 0.15%\n32 layer3.0.downsample 512 28 28 1024 14 14 526336 1.53MB 206,123,008 1.76%\n33 layer3.1 1024 14 14 1024 14 14 1117184 3.06MB 437,835,776 3.97%\n34 layer3.2 1024 14 14 1024 14 14 1117184 3.06MB 437,835,776 4.06%\n35 layer3.3 1024 14 14 1024 14 14 1117184 3.06MB 437,835,776 4.05%\n36 layer3.4 1024 14 14 1024 14 14 1117184 3.06MB 437,835,776 4.03%\n37 layer3.5 1024 14 14 1024 14 14 1117184 3.06MB 437,835,776 4.02%\n38 layer4.0.conv1_Conv2d 1024 14 14 512 14 14 524288 0.38MB 205,420,544 1.42%\n39 layer4.0.bn1_BatchNorm2d 512 14 14 512 14 14 1024 0.38MB 401,408 0.14%\n40 layer4.0.conv2_Conv2d 512 14 14 512 7 7 2359296 0.10MB 231,185,920 2.41%\n41 layer4.0.bn2_BatchNorm2d 512 7 7 512 7 7 1024 0.10MB 100,352 0.13%\n42 layer4.0.conv3_Conv2d 512 7 7 2048 7 7 1048576 0.38MB 102,660,096 1.03%\n43 layer4.0.bn3_BatchNorm2d 2048 7 7 2048 7 7 4096 0.38MB 401,408 0.44%\n44 layer4.0.relu_ReLU 2048 7 7 2048 7 7 0 0.38MB 100,352 0.08%\n45 layer4.0.downsample 1024 14 14 2048 7 7 2101248 0.77MB 205,821,952 2.48%\n46 layer4.1 2048 7 7 2048 7 7 4462592 1.53MB 437,283,840 5.26%\n47 layer4.2 2048 7 7 2048 7 7 4462592 1.53MB 437,283,840 5.27%\n48 avgpool_AvgPool2d 2048 7 7 2048 1 1 0 0.01MB 100,352 0.05%\n49 fc_Linear 2048 1000 2049000 0.00MB 4,095,000 0.24%\n===============================================================================================================================\ntotal parameters quantity: 25,557,032\ntotal memory: 109.69MB\ntotal MAdd: 8,219,737,624\n\n```\n### validation mode\n To make sure the result is correct, a validation mothod from https://github.com/sksq96/pytorch-summary is added.\n To use this validation method, please install *torchsummary*\n `pip3 install torchsummary`","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fnemonameless%2Fpytorch_summary","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fnemonameless%2Fpytorch_summary","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fnemonameless%2Fpytorch_summary/lists"}