{"id":13784025,"url":"https://github.com/MLI-lab/channel_normalization","last_synced_at":"2025-05-11T19:32:02.907Z","repository":{"id":110075957,"uuid":"207923992","full_name":"MLI-lab/channel_normalization","owner":"MLI-lab","description":null,"archived":false,"fork":false,"pushed_at":"2019-09-11T23:27:34.000Z","size":8868,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":2,"default_branch":"master","last_synced_at":"2024-11-17T20:48:05.074Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Jupyter Notebook","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/MLI-lab.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":"2019-09-11T23:25:34.000Z","updated_at":"2022-10-19T08:50:17.000Z","dependencies_parsed_at":"2023-05-21T02:30:39.159Z","dependency_job_id":null,"html_url":"https://github.com/MLI-lab/channel_normalization","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/MLI-lab%2Fchannel_normalization","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MLI-lab%2Fchannel_normalization/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MLI-lab%2Fchannel_normalization/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MLI-lab%2Fchannel_normalization/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/MLI-lab","download_url":"https://codeload.github.com/MLI-lab/channel_normalization/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":253621157,"owners_count":21937486,"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":"2024-08-03T19:00:34.537Z","updated_at":"2025-05-11T19:32:02.190Z","avatar_url":"https://github.com/MLI-lab.png","language":"Jupyter Notebook","readme":"# Channel normalization in convolutional neural networks\n\nThis folder provides the code for reproducing the results in the paper: \n\n**``Channel Normalization in Convolutional Neural Network avoids Vanishing Gradients''**, by Zhenwei Dai and Reinhard Heckel, ICML workshop 2019.\n\nThe paper is available online [[here]](http://www.reinhardheckel.com/papers/channel_normalization.pdf).\n\n## Installation\n\nThe code is written in python and relies on pytorch. The following libraries are required: \n- python 3\n- pytorch\n- numpy\n- skimage\n- matplotlib\n- scikit-image\n- jupyter\n\n## Citation\n```\n@InProceedings{dai_channel_2019,\n    author    = {Zhenwai Dai and Reinhard Heckel},\n    title     = {Channel Normalization in Convolutional Neural Network avoids Vanishing Gradients},\n    booktitle   = {International Conference on Machine Learning, Deep Phenomena Workshop},\n    year      = {2019}\n}\n```\n\n## Content of the repository\n\n**one_dim_net_convergence_paper.ipynb** includes the code to run gradient descent on deep decoder, multi-channel CNN and linear CNN, and can be used to reproduce Figure 1,2, and 5.\n\n**visualize_loss_function_landscape.ipynb** plots the loss function landscape of multi-channel CNN and linear CNN\n\n**distribution_gradient_linear_network_initialization.ipynb** plots the gradients norm at initialization (with Normal distribution) for a linear CNN, to reproduced Figure 4a and 4b.\n\n**distribution_gradient_CNN_initialization.ipynb** plots the gradients norm at initialization (with Normal distribution) of a multichannel CNN, to reproduced Figure 4c and 4d.\n\n## Licence\n\nAll files are provided under the terms of the Apache License, Version 2.0.\n","funding_links":[],"categories":["2019"],"sub_categories":[],"project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FMLI-lab%2Fchannel_normalization","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FMLI-lab%2Fchannel_normalization","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FMLI-lab%2Fchannel_normalization/lists"}