{"id":15402788,"url":"https://github.com/zusez4/rust_rl","last_synced_at":"2025-04-16T03:09:36.392Z","repository":{"id":54811491,"uuid":"268384859","full_name":"ZuseZ4/Rust_RL","owner":"ZuseZ4","description":"A Reinforcement Learning / Neural Network library, written in Rust.","archived":false,"fork":false,"pushed_at":"2021-03-07T19:13:14.000Z","size":11863,"stargazers_count":19,"open_issues_count":2,"forks_count":2,"subscribers_count":4,"default_branch":"master","last_synced_at":"2025-04-16T03:09:03.544Z","etag":null,"topics":["neural-network","rust"],"latest_commit_sha":null,"homepage":"","language":"Rust","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/ZuseZ4.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}},"created_at":"2020-05-31T23:54:44.000Z","updated_at":"2025-03-21T16:09:29.000Z","dependencies_parsed_at":"2022-08-14T03:31:32.394Z","dependency_job_id":null,"html_url":"https://github.com/ZuseZ4/Rust_RL","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/ZuseZ4%2FRust_RL","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZuseZ4%2FRust_RL/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZuseZ4%2FRust_RL/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZuseZ4%2FRust_RL/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/ZuseZ4","download_url":"https://codeload.github.com/ZuseZ4/Rust_RL/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":249188426,"owners_count":21227015,"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":["neural-network","rust"],"created_at":"2024-10-01T16:04:53.476Z","updated_at":"2025-04-16T03:09:36.358Z","avatar_url":"https://github.com/ZuseZ4.png","language":"Rust","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Rust_RL\n\nThis is a test repository to learn about Rust, Neural Networks and Reinforcement Learning.\nThe Neural Network implementations have some real basic optimizations and the forward pass supports parallelization.\nHowever, it comes with some design-flawes and it is significantly limited by not supporting GPUs and not supporting any kind of autodiff. \n\n# FUNCTIONALITY\n1) The following layers have been implemented: Dense, Dropout, Flatten, Reshape  \n2) Convolution_Layer (weight updates work, but don't stack them)  \n3) The following activation functions have been implemented: Softmax, Sigmoid, ReLu, LeakyReLu  \n4) The following loss functions have been implemented: MSE, RMSE, binary_crossentropy, categorical_crossentropy  \n5) The following optimizer have been implemented: SGD (default), Momentum, AdaGrad, RMSProp, Adam  \n6) Networks work for 1d, 2d, or 3d input. Exact input shape has to be given, following layers adjust accordingly.  \n7) Available Datasets: Mnist(-Fashion), Cifar10, Cifar100\n8) Available Agents: Random, Q-Learning, DQN, Double-DQN\n9) Available Environments: TicTacToe, Fortress (https://www.c64-wiki.com/wiki/Fortress_(SSI)\n\n# EXAMPLES\nMNIST (achieving ~98%)\nCIFAR10 (achieving ~49%)\n\nTicTacToe (results in an optimal Agent)\nFortress (... at least DDQN performs significantly better than a random moving bot)\n\n\n# TODO\n1) Add backpropagation of error to conv_layers  \n2) Improve design\n3) Add GPU support for matrix-matrix multiplications\n4) Add some autodiff support.\n\nAt least the last two TODOs will probably stay for some time.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fzusez4%2Frust_rl","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fzusez4%2Frust_rl","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fzusez4%2Frust_rl/lists"}