{"id":19459802,"url":"https://github.com/jsflo/convolutionalneuralnetwork-tf","last_synced_at":"2025-10-07T07:10:23.185Z","repository":{"id":187794756,"uuid":"103673944","full_name":"JsFlo/ConvolutionalNeuralNetwork-TF","owner":"JsFlo","description":"Different convolutional nerual network architectures in TensorFlow","archived":false,"fork":false,"pushed_at":"2017-09-16T18:07:51.000Z","size":1160,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":3,"default_branch":"master","last_synced_at":"2025-01-08T02:16:02.850Z","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":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/JsFlo.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}},"created_at":"2017-09-15T15:40:58.000Z","updated_at":"2017-09-15T23:10:11.000Z","dependencies_parsed_at":"2023-08-12T06:52:18.833Z","dependency_job_id":null,"html_url":"https://github.com/JsFlo/ConvolutionalNeuralNetwork-TF","commit_stats":null,"previous_names":["jsflo/convolutionalneuralnetwork-tf"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/JsFlo%2FConvolutionalNeuralNetwork-TF","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/JsFlo%2FConvolutionalNeuralNetwork-TF/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/JsFlo%2FConvolutionalNeuralNetwork-TF/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/JsFlo%2FConvolutionalNeuralNetwork-TF/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/JsFlo","download_url":"https://codeload.github.com/JsFlo/ConvolutionalNeuralNetwork-TF/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":240664480,"owners_count":19837563,"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":"2024-11-10T17:34:09.727Z","updated_at":"2025-10-07T07:10:18.139Z","avatar_url":"https://github.com/JsFlo.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# ConvolutionalNeuralNetwork-TF\n\n## Deep MNIST CNN without Pooling\n### DeepMnistCnnNoPooling.py\nConvolutional Nerual Network without Pooling.\n* 3 Convolutional layers\n* 2 Fully Connected Layers\n\nAdds **non-linearity** after each of the above layers.\n\n[conv -\u003e relu -\u003e conv -\u003e relu -\u003e conv -\u003e relu -\u003e fc -\u003e relu -fc -\u003e relu]\n\nThis model does not use pooling to reduce the size. \nThe **3rd hidden layer** (conv3) uses a **stride** of **2** to reduce the size.\n\nNot using pooling is an idea seen here:[ Striving for Simplicity: The All Convolutional Net](https://arxiv.org/abs/1412.6806)\n\nAccuracy: **99.1**\n\n### Notes\nThese **convolutional** layers do not use **zero-padding** so\nthe sizes of the output volume changes. This means that I had to carefully\ntrack the output volumes.\n\nThe output volumes can be calculated with\n#### ((W - F + 2P)/ (S)) + 1\n* W = Input Volume size\n* F = Filter Size (\"receptive field size\")\n* P = Zero Padding used\n* S = Stride\n\nBecause we use 0 zero-padding it can be simplified to:\n#### ((W - F)/(S)) + 1\n\nAnother **disadvantage** to not using **zero-padding** is that \nthe edges at each convolutional don't get **as many neurons looking** at them\ncompared to a traditional architecture with **zero-padding** at every conv layer \nto keep the volume the same( usually reduced by max-pooling).\n\n## Traditional CNN\nA more **traditional CNN** that relies on **2x2 pooling** to **reduce dimensions**.\nThe **convolutions don't** reshape the **dimensions** because it **zero-pads**.\n\n\u003cimg src=\"conv_example.png\" width=\"600\"\u003e","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjsflo%2Fconvolutionalneuralnetwork-tf","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fjsflo%2Fconvolutionalneuralnetwork-tf","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjsflo%2Fconvolutionalneuralnetwork-tf/lists"}