{"id":13499010,"url":"https://github.com/nutszebra/prelu_net","last_synced_at":"2025-03-29T03:32:13.784Z","repository":{"id":201706505,"uuid":"76662139","full_name":"nutszebra/prelu_net","owner":"nutszebra","description":"Implementation of PReLUNet by chainer (Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification: https://arxiv.org/abs/1502.01852)","archived":false,"fork":false,"pushed_at":"2017-02-02T06:20:29.000Z","size":56,"stargazers_count":12,"open_issues_count":1,"forks_count":2,"subscribers_count":2,"default_branch":"master","last_synced_at":"2024-10-31T17:38:56.259Z","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":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/nutszebra.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":"2016-12-16T15:00:59.000Z","updated_at":"2023-07-24T12:59:35.000Z","dependencies_parsed_at":"2024-03-31T20:15:20.768Z","dependency_job_id":null,"html_url":"https://github.com/nutszebra/prelu_net","commit_stats":null,"previous_names":["nutszebra/prelu_net"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/nutszebra%2Fprelu_net","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/nutszebra%2Fprelu_net/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/nutszebra%2Fprelu_net/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/nutszebra%2Fprelu_net/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/nutszebra","download_url":"https://codeload.github.com/nutszebra/prelu_net/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":246135717,"owners_count":20729056,"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":"2024-07-31T22:00:25.939Z","updated_at":"2025-03-29T03:32:13.476Z","avatar_url":"https://github.com/nutszebra.png","language":"Python","funding_links":[],"categories":["Papers\u0026Codes"],"sub_categories":["PReLU-nets"],"readme":"# What's this\nImplementation of PReLUNet by chainer  \n\n# Dependencies\n\n    git clone https://github.com/nutszebra/prelu_net.git\n    cd prelu_net\n    git submodule init\n    git submodule update\n\n# How to run\n    python main.py -g 0\n\n# Details about my implementation\nAll hyperparameters and network architecture are the same as in [[1]][Paper] except for some parts.\n\n* Data augmentation  \nTrain: Pictures are randomly resized in the range of [256, 512], then 224x224 patches are extracted randomly and are normalized locally. Horizontal flipping is applied with 0.5 probability.  \nTest: Pictures are resized to 384x384, then they are normalized locally. Single image test is used to calculate total accuracy.  \n\n* SPP net\nInstead of spp, I use global average pooling.\n\n* Learning rate schedule\nLearning rate is divided by 10 at [150, 225] epoch. The total number of epochs is 300.\n\n# Cifar10 result\n| network                                                   | total accuracy (%) |\n|:----------------------------------------------------------|-------------------:|\n| my implementation(model A)                                | 94.98              |\n\n\u003cimg src=\"https://github.com/nutszebra/prelu_net/blob/master/loss.jpg\" alt=\"loss\" title=\"loss\"\u003e\n\u003cimg src=\"https://github.com/nutszebra/prelu_net/blob/master/accuracy.jpg\" alt=\"total accuracy\" title=\"total accuracy\"\u003e\n\n# References\nDelving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification [[1]][Paper]  \n\n[paper]: https://arxiv.org/abs/1502.01852 \"Paper\"\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fnutszebra%2Fprelu_net","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fnutszebra%2Fprelu_net","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fnutszebra%2Fprelu_net/lists"}