{"id":13499041,"url":"https://github.com/asmith26/wide_resnets_keras","last_synced_at":"2026-01-08T03:00:48.348Z","repository":{"id":8913837,"uuid":"60194498","full_name":"asmith26/wide_resnets_keras","owner":"asmith26","description":"Keras implementation + pretrained weights for \"Wide Residual Networks\"","archived":false,"fork":false,"pushed_at":"2024-01-18T22:48:51.000Z","size":34567,"stargazers_count":138,"open_issues_count":6,"forks_count":47,"subscribers_count":6,"default_branch":"main","last_synced_at":"2024-10-31T17:39:06.778Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"http://arxiv.org/abs/1605.07146v1","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/asmith26.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":"2016-06-01T16:44:46.000Z","updated_at":"2024-04-01T07:50:11.000Z","dependencies_parsed_at":"2024-10-31T17:32:05.259Z","dependency_job_id":"344620a5-a1ab-4739-a965-c0a7359b65b4","html_url":"https://github.com/asmith26/wide_resnets_keras","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/asmith26%2Fwide_resnets_keras","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/asmith26%2Fwide_resnets_keras/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/asmith26%2Fwide_resnets_keras/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/asmith26%2Fwide_resnets_keras/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/asmith26","download_url":"https://codeload.github.com/asmith26/wide_resnets_keras/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":246135766,"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":"2024-07-31T22:00:27.377Z","updated_at":"2026-01-08T03:00:43.034Z","avatar_url":"https://github.com/asmith26.png","language":"Python","funding_links":[],"categories":["Papers\u0026Codes","Deep Residual Learning","2 目标检测"],"sub_categories":["WRN","Implementations"],"readme":"# Keras implementation of \"Wide Residual Networks\"\nThis repo contains the code to run Wide Residual Networks using Keras.\n- Paper (v1): http://arxiv.org/abs/1605.07146v1 (the authors have since published a v2 of the paper, which introduces slightly different preprocessing and improves the accuracy a little).\n- Original code: https://github.com/szagoruyko/wide-residual-networks\n\n\n## Dependencies:\n- `pip install -r requirements.txt`\n- To plot the architecture of the model used (like the plot of the WRN-16-2 architecture plotted [below](#example-plot)), you need to install `pydot` and `graphviz`. I recommend installing with `conda install -c conda-forge python-graphviz`:\n\n\n## Training Details:\nRun the default configuration (i.e. best configuration for CIFAR10 from original paper/code, WRN-28-10 without dropout) with:\n\n```\n$ python main.py\n```\n\nThere are three configuration sections at the top of `main.py`:\n- [DATA CONFIGURATION](https://github.com/asmith26/wide_resnets_keras/blob/master/main.py#L34-48): Containing data details.\n- [NETWORK/TRAINING CONFIGURATION](https://github.com/asmith26/wide_resnets_keras/blob/master/main.py#L50-87): Includes the main parameters the authors experimented with.\n- [OUTPUT CONFIGURATION](https://github.com/asmith26/wide_resnets_keras/blob/master/main.py#L89-97): Defines paths regarding where to save model/checkpoint weights and plots.\n\n\n## Results and Trained models:\n- ***WRN-28-10 no dropout***:\n  - Using these values in **main.py**, I obtained a **test loss = 0.31** and **test accuracy = 0.93**. This test error (i.e. 1 - 0.93 = **7%**) is a little higher than the reported result (Table 4 states the same model obtains a test error of *4.97%*); see the note below for a likely explanation.\n  - You can find the trained weights for this model at **models/WRN-28-10.h5**, whilst **[models/test.py](https://github.com/asmith26/wide_resnets_keras/blob/master/models/test.py)** provides an example of running these weights against the test set.\n\n**Note:** I have not followed the exact same preprocessing and data augmentation steps used in the paper, in particular:\n\n- \"global *contrast* normalization\", and\n- \"random crops from image padded by 4 pixels on each side, filling missing pixels with reflections of original image\", which appears to be implemented in [this file](https://github.com/szagoruyko/wide-residual-networks/blob/8b166cc15fa8a598490ce0ae66365bf165dffb75/augmentation.lua).\n\nIdeally, we will add such methods directly to the [Keras image preprocessing script](https://github.com/fchollet/keras/blob/master/keras/preprocessing/image.py).\n\n\n## \u003ca name=\"example-plot\"\u003eWRN-16-2 Architecture\u003c/a\u003e\n![WRN-16-2 Architecture](models/WRN-16-2.png?raw=true \"WRN-16-2 Architecture\")\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fasmith26%2Fwide_resnets_keras","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fasmith26%2Fwide_resnets_keras","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fasmith26%2Fwide_resnets_keras/lists"}