{"id":25518290,"url":"https://github.com/sidray-infinity/dumbtf","last_synced_at":"2025-07-07T18:33:25.059Z","repository":{"id":54830111,"uuid":"258236128","full_name":"Sidray-Infinity/DumbTF","owner":"Sidray-Infinity","description":"A deep learning library for dumb people. ","archived":false,"fork":false,"pushed_at":"2024-05-03T20:54:44.000Z","size":669,"stargazers_count":2,"open_issues_count":3,"forks_count":1,"subscribers_count":1,"default_branch":"master","last_synced_at":"2025-04-10T23:51:44.568Z","etag":null,"topics":["deep-learning","flask","machine-learning","material-ui","reactjs"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"gpl-3.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/Sidray-Infinity.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,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2020-04-23T14:43:09.000Z","updated_at":"2024-04-30T04:16:04.000Z","dependencies_parsed_at":"2024-05-03T22:03:33.955Z","dependency_job_id":null,"html_url":"https://github.com/Sidray-Infinity/DumbTF","commit_stats":{"total_commits":61,"total_committers":5,"mean_commits":12.2,"dds":0.5737704918032787,"last_synced_commit":"4736e6010cbc8c7c8cacd5b877183a9e1368f917"},"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Sidray-Infinity%2FDumbTF","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Sidray-Infinity%2FDumbTF/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Sidray-Infinity%2FDumbTF/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Sidray-Infinity%2FDumbTF/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Sidray-Infinity","download_url":"https://codeload.github.com/Sidray-Infinity/DumbTF/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248317706,"owners_count":21083528,"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":["deep-learning","flask","machine-learning","material-ui","reactjs"],"created_at":"2025-02-19T16:25:03.525Z","updated_at":"2025-04-10T23:51:46.877Z","avatar_url":"https://github.com/Sidray-Infinity.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cp align=center\u003e\u003cimg src=\"images/logo.png\"\n     alt=\"logo\"\n     style=\"height: 180px; width: 200px;\" /\u003e\u003c/p\u003e\n\nDumbTF\n---\nA deep learning library for dumb people.\n\nFocus\n---\n\n- Implement additional layers, to be used along with convolutions\n- Explore LAMB : https://towardsdatascience.com/an-intuitive-understanding-of-the-lamb-optimizer-46f8c0ae4866\n- Convolutions are still slow, despite using the numpy stride trick and einsum.\n\n## Installation\n\n`pip install Dumb-TF`\n\n## Status\n---\n\n- ### Regression (Boston Housing Dataset)\n  Network Architecture\n    ```python \n    model = Model()\n    model.add(Dense(128, input_shape=13, activation='relu'))\n    model.add(Dense(64, input_shape=128, activation='relu'))\n    model.add(Dense(1, input_shape=64, activation='linear'))\n\n    model.compile(loss=\"mse\", optimizer=\"mini_batch_gd\", lr=0.01)\n\n    loss, loss_pe = model.fit(x_train, y_train, epochs=20, batch_size=64)\n    ```\n  \u003cimg src=\"images/regression.png\"\n  alt=\"Regression Graph\"\n  style=\"margin-right: 10px; height:400px; width: 800px;\" /\u003e\n\n- ### Binary Classification (Breast Cancer Dataset)\n  Network Architecture\n    ```python\n    model = Model()\n    model.add(Dense(64, input_shape=30, activation='relu'))\n    model.add(Dense(32, input_shape=64, activation='relu'))\n    model.add(Dense(1, input_shape=32, activation='sigmoid'))\n\n    model.compile(loss=\"bce\", optimizer=\"mini_batch_gd\", lr=0.01)\n\n    loss, loss_pe = model.fit(x, y, epochs=150, batch_size=64)\n    ```\n  \u003cimg src=\"images/bce.png\"\n  alt=\"BCE Graph\"\n  style=\"margin-right: 10px; height:400px; width: 800px;\" /\u003e\n\n- ### Multiclass Classification (MNIST)\n  Network Architecture\n    ```python\n    model = Model()\n    model.add(Dense(32, input_shape=784, activation='relu'))\n    model.add(Dense(10, input_shape=32, activation='softmax'))\n\n    model.compile(loss=\"cce\", optimizer=\"mini_batch_gd\", lr=0.01)\n\n    loss, loss_pe = model.fit(x_train, y_train, epochs=10, batch_size=64)\n    ```\n  \u003cimg src=\"images/cce.png\"\n  alt=\"BCE Graph\"\n  style=\"margin-right: 10px; height:400px; width: 800px;\" /\u003e\n\n- ### Convolution Network (MNIST)\n  Network Architecture\n    ```python\n    model = Model()\n    model.add(Conv2D(32, 3, 'relu', (28, 28, 1)))\n    model.add(Conv2D(32, 3, 'relu', (26, 26, 32)))\n    model.add(Flatten())\n    model.add(Dense(10, input_shape=26*26*32, activation='softmax'))\n\n    model.compile(loss='cce', optimizer='mini_batch_gd')\n\n    loss, loss_pe = model.fit(x_train, y_train, epochs=10, batch_size=64)\n    ```\n  \u003cimg src=\"images/conv.png\"\n  alt=\"BCE Graph\"\n  style=\"margin-right: 10px; height:400px; width: 800px;\" /\u003e\n\n## References\n\n- http://neuralnetworksanddeeplearning.com/chap2.html\n- https://brilliant.org/wiki/backpropagation/#$\n- https://towardsdatascience.com/batch-mini-batch-stochastic-gradient-descent-\n- https://www.youtube.com/watch?v=i94OvYb6noo\n- https://aimatters.wordpress.com/2020/06/14/derivative-of-softmax-layer/\n- https://towardsdatascience.com/deriving-the-backpropagation-equations-from-scratch-part-1-343b300c585a\n\n### For Convolutions\n\n* https://www.sicara.ai/blog/2019-10-31-convolutional-layer-convolution-kernel\n* https://towardsdatascience.com/beginners-guide-to-understanding-convolutional-neural-networks-ae9ed58bb17d\n* https://www.youtube.com/watch?v=8rrHTtUzyZA\n\n(FAST CONVOLUTIONS)\n\n* https://laurentperrinet.github.io/sciblog/posts/2017-09-20-the-fastest-2d-convolution-in-the-world.html\n* https://towardsdatascience.com/how-are-convolutions-actually-performed-under-the-hood-226523ce7fbf\n* https://jessicastringham.net/2017/12/31/stride-tricks/\n\n(Backpropagation in convolutions)\n\n* https://medium.com/@pavisj/convolutions-and-backpropagations-46026a8f5d2c\n* https://www.jefkine.com/general/2016/09/05/backpropagation-in-convolutional-neural-networks/\n\n(MaxPool2D)\n\n* https://wiseodd.github.io/techblog/2016/07/18/convnet-maxpool-layer/#:~:text=Maxpool%20backward\u0026text=We%20let%20the%20gradient%20pass,max%20operation%20do%20in%20backpropagation\n  .\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsidray-infinity%2Fdumbtf","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsidray-infinity%2Fdumbtf","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsidray-infinity%2Fdumbtf/lists"}