{"id":16889415,"url":"https://github.com/lyken17/sparsenet","last_synced_at":"2025-03-17T06:31:42.171Z","repository":{"id":66151039,"uuid":"107478240","full_name":"Lyken17/SparseNet","owner":"Lyken17","description":"[ECCV 2018] Sparsely Aggreagated Convolutional Networks https://arxiv.org/abs/1801.05895","archived":false,"fork":false,"pushed_at":"2018-10-10T21:25:29.000Z","size":1138,"stargazers_count":124,"open_issues_count":4,"forks_count":25,"subscribers_count":7,"default_branch":"master","last_synced_at":"2025-02-27T19:04:34.309Z","etag":null,"topics":["computer-vision","convolutional-neural-networks","deep-learning"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"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/Lyken17.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}},"created_at":"2017-10-19T00:29:09.000Z","updated_at":"2025-01-21T03:57:54.000Z","dependencies_parsed_at":"2023-05-06T03:17:00.241Z","dependency_job_id":null,"html_url":"https://github.com/Lyken17/SparseNet","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/Lyken17%2FSparseNet","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Lyken17%2FSparseNet/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Lyken17%2FSparseNet/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Lyken17%2FSparseNet/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Lyken17","download_url":"https://codeload.github.com/Lyken17/SparseNet/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":243847062,"owners_count":20357317,"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":["computer-vision","convolutional-neural-networks","deep-learning"],"created_at":"2024-10-13T16:57:12.288Z","updated_at":"2025-03-17T06:31:41.486Z","avatar_url":"https://github.com/Lyken17.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# SparseNet\nSparsely Aggregated Convolutional Networks [[PDF](https://arxiv.org/abs/1801.05895)]\n\n[Ligeng Zhu](https://lzhu.me), [Ruizhi Deng](http://www.sfu.ca/~ruizhid/), [Michael Maire](http://ttic.uchicago.edu/~mmaire/), [Zhiwei Deng](http://www.sfu.ca/~zhiweid/), [Greg Mori](http://www.cs.sfu.ca/~mori/), [Ping Tan](https://www.cs.sfu.ca/~pingtan/)\n\n# What is SparseNet?\nSparseNet is a  network architecture that only aggregates previous layers with exponential offset, for example, i - 1, i - 2, i - 4, i - 8, i - 16 ... \n\n![](images/dense_and_sparse.png)\n\n# Why use SparseNet?\nThe connectivity pattern yields state-of-the-art arruacies on small dataset CIFAR/10/100. On large scale ILSVRC 2012 (ImageNet) dataset, SparseNet achieves similar accuracy as ResNet and DenseNet, while only using much less parameters. \n\n# Better Performance\n\n\u003ctable\u003e\n\u003ctr\u003e\u003cth\u003e Without BC \u003c/th\u003e\u003cth\u003e With BC \u003c/th\u003e\u003c/tr\u003e\n\u003ctr\u003e\u003ctd\u003e\n\nArchitecture | Params | CIFAR 100\n--- | --- | ---\nDenseNet-40-12  | 1.1M | 24.79\nDenseNet-100-12 | 7.2M | 20.97\nDenseNet-100-24 | 28.28M | 19.61\n--- | --- | ---\nSparseNet-40-24  | 0.76M | 24.65\nSparseNet-100-36 | 5.65M | 20.50\nSparseNet-100-{16,32,64} | 7.22M | 19.49\n\n\n\u003c/td\u003e\u003ctd\u003e\n\nArchitecture | Params | CIFAR 100\n--- | --- | ---\nDenseNet-100-12 | 0.8M | 22.62\nDenseNet-250-24 | 15.3M | 17,6\nDenseNet-190-40 | 25.6M | 17.53\n--- | --- | ---\nSparseNet-100-24  | 1.46M | 22.12\nSparseNet-100-{16,32,64} | 4.38M | 19.71\nSparseNet-100-{32,64,128} | 16.72M | 17.71\n\n\n\u003c/td\u003e\u003c/tr\u003e \u003c/table\u003e\n\n\n## Efficient Parameter Utilization\n* Parameter efficiency on CIFAR\n  ![](images/cropped_two-weights-int.jpg)\n\n* Paramter efficiency on ImageNet\n\n  We notice sparsenet shows comparable efficiency even compared with pruned models.\n  ![](images/imagenet_efficiency.png)\n  \n  \n# Pretrained model\nRefer for [source folder](src/).\n\n# Cite\nIf SparseNet helps your research, please cite our work :) \n\n```\n@article{DBLP:journals/corr/abs-1801-05895,\n  author    = {Ligeng Zhu and\n               Ruizhi Deng and\n               Michael Maire and\n               Zhiwei Deng and\n               Greg Mori and\n               Ping Tan},\n  title     = {Sparsely Aggregated Convolutional Networks},\n  journal   = {CoRR},\n  volume    = {abs/1801.05895},\n  year      = {2018},\n  url       = {http://arxiv.org/abs/1801.05895},\n  archivePrefix = {arXiv},\n  eprint    = {1801.05895},\n  biburl    = {https://dblp.org/rec/bib/journals/corr/abs-1801-05895},\n  bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Flyken17%2Fsparsenet","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Flyken17%2Fsparsenet","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Flyken17%2Fsparsenet/lists"}