{"id":23824683,"url":"https://github.com/soran-ghaderi/backpropagation","last_synced_at":"2026-06-29T14:32:16.835Z","repository":{"id":218830315,"uuid":"747455641","full_name":"soran-ghaderi/backpropagation","owner":"soran-ghaderi","description":"Backpropagation and automatic differentiation, and grid search from scratch.","archived":false,"fork":false,"pushed_at":"2024-01-24T01:20:51.000Z","size":13732,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"master","last_synced_at":"2025-11-12T14:34:29.618Z","etag":null,"topics":["automatic-differentiation","backpropagation","backpropagation-neural-network","gridsearch","mlp","neural-network","tuning","tutorial"],"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/soran-ghaderi.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,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2024-01-24T00:43:23.000Z","updated_at":"2024-04-30T15:10:57.000Z","dependencies_parsed_at":null,"dependency_job_id":"497eacf9-0527-45ea-aef1-a42be06fd560","html_url":"https://github.com/soran-ghaderi/backpropagation","commit_stats":null,"previous_names":["soran-ghaderi/backpropagation"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/soran-ghaderi/backpropagation","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/soran-ghaderi%2Fbackpropagation","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/soran-ghaderi%2Fbackpropagation/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/soran-ghaderi%2Fbackpropagation/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/soran-ghaderi%2Fbackpropagation/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/soran-ghaderi","download_url":"https://codeload.github.com/soran-ghaderi/backpropagation/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/soran-ghaderi%2Fbackpropagation/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":34931587,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-06-29T02:00:05.398Z","response_time":58,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"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":["automatic-differentiation","backpropagation","backpropagation-neural-network","gridsearch","mlp","neural-network","tuning","tutorial"],"created_at":"2025-01-02T11:15:25.597Z","updated_at":"2026-06-29T14:32:16.830Z","avatar_url":"https://github.com/soran-ghaderi.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# A from-scratch implementation of the following components with a Keras-like API:\n\n - Automatic differentiation and backpropagation\n - Dense, Sequential, Model layers\n - Adam optimizer, SGD\n - MSE, RMSE, SimpleError\n - Grid Search\n\nThis repository implements a simple multi-regressor MLP for controlling a lunar lander agent. This serves as a tutorial to get started with backpropagation and automatic differentiation. \n\nYou can easily extend this boilerplate to implement Conv1D, Conv2D, etc. layers.\n\nCredits: The automatic differentiation is heavily based on [micrograd](https://github.com/karpathy/micrograd/blob/master/micrograd/engine.py). Additions to Dr. Karpathy's implementation: are as follows:\n - Topological ordering using DP instead of recursion for speed and scalability\n - Support for division (the original implementation does not work)\n - Support for Sigmoid and Softmax activation functions\n - Resolve other minor errors\n\nOther features:\n - Dense layer automatically extracts the input shape and dimension\n - DataProcessor\n - DataLoader\n - Early stopping\n - Save best weights\n\nExmaple:\n\n```python\noriginal_file_path = 'data/ce889_dataCollection.csv'\nnormalized_output_path = 'data/normalized_data.csv'\ndata_processor = DataProcessor(original_file_path)\nnormalized_data = data_processor.normalize(save_path=normalized_output_path)\ndata_wrapper = NormalizedDataWrapper(original_file_path)\ndata_loader = DataLoader(file_path=normalized_output_path, validation_size=0.1)\n# Load data\ndata_loader.load_data(1000)\n# Split data into training, validation, and testing sets\ninputs_train, inputs_val, inputs_test, outputs_train, outputs_val, outputs_test = data_loader.split_data()\n\n\nbeta1 = 0.9\nbeta2 = 0.999\nlr = 0.1\nclip_threshold = 1.0\nactivation_function = 'sigmoid'\n\nmodel = Sequential([\n    Dense(2, activation=activation_function),\n    Dense(32, activation=activation_function),\n    Dense(2) \n])\n\nloss = MeanSquaredError()\noptimizer = Adam(learning_rate=lr, beta1=beta1, beta2=beta2, clip_threshold=clip_threshold)\nmodel.compile(optimizer=optimizer, loss=loss)\n\nmodel.fit(inputs=inputs_train, outputs=outputs_train, val_inputs=inputs_val, val_outputs=outputs_val, batch_size=1, epochs=50, verbose=1,\n          early_stopping={'min_delta': 0.001, 'patience': 5},\n          save_best_model='./weights/best_model_weights.npy')\n\npred = model.predict(inputs_test)\nevaluation_results = model.evaluate(inputs_test, outputs_test)\nprint(\"Evaluation Results after training:\", evaluation_results)\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsoran-ghaderi%2Fbackpropagation","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsoran-ghaderi%2Fbackpropagation","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsoran-ghaderi%2Fbackpropagation/lists"}