{"id":30197587,"url":"https://github.com/mplantady/twinstickrl","last_synced_at":"2026-06-14T21:32:34.903Z","repository":{"id":307922620,"uuid":"146819617","full_name":"mplantady/TwinStickRL","owner":"mplantady","description":"Twin-stick shooter simulation using reinforcement learning in Unity","archived":false,"fork":false,"pushed_at":"2018-09-27T22:21:18.000Z","size":54500,"stargazers_count":2,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"master","last_synced_at":"2025-08-13T06:37:32.254Z","etag":null,"topics":["reinforcement-learning","unityml"],"latest_commit_sha":null,"homepage":"","language":"C#","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/mplantady.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,"zenodo":null}},"created_at":"2018-08-31T00:05:52.000Z","updated_at":"2019-04-24T12:25:16.000Z","dependencies_parsed_at":"2025-08-03T03:47:32.739Z","dependency_job_id":null,"html_url":"https://github.com/mplantady/TwinStickRL","commit_stats":null,"previous_names":["mplantady/twinstickrl"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/mplantady/TwinStickRL","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mplantady%2FTwinStickRL","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mplantady%2FTwinStickRL/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mplantady%2FTwinStickRL/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mplantady%2FTwinStickRL/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/mplantady","download_url":"https://codeload.github.com/mplantady/TwinStickRL/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mplantady%2FTwinStickRL/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":34339194,"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":"2026-06-14T02:00:07.365Z","response_time":62,"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":["reinforcement-learning","unityml"],"created_at":"2025-08-13T06:24:39.352Z","updated_at":"2026-06-14T21:32:34.898Z","avatar_url":"https://github.com/mplantady.png","language":"C#","funding_links":[],"categories":[],"sub_categories":[],"readme":"# TwinStickRL\n\n![TwinStick](https://github.com/mplantady/TwinStickRL/blob/master/Preview.gif)\n\n## Installation\nUnityML is already included in the source, but you need to follow this to install the plugin TensorflowSharp:\nhttps://github.com/llSourcell/Unity_ML_Agents/blob/master/docs/Using-TensorFlow-Sharp-in-Unity-(Experimental).md\n\n## Environment\nInspired by the game \"Geometry Wars\" a ship need to survive in a small arena.\nThe ship can move and shoot simultaneously in any direction.\n\n## AI Agent\nThe agent has 2 brains, one that control the movement and the other aim and shoot.\nThey both get a small reward at each step (0.1) and a negative score when the ship is killed.\n\n## Scenes\n- \"Main\" is a single arena to test an agent in fullscreen\n- \"Train\" is a gym with multiple simultaneous arenas to speed-up training\n\n## Training\nYou must build the game with one brain in external and the other in internal mode. (TwinMove or TwinAim)\n\nThe 2 brains are trained separately using this command:\n\n### Move Brain\npython learn.py ./build/TwinMove.exe --curriculum=./curricula/Twin_move.json --run-id=Twin_move --train\n\n### Aim Brain\npython learn.py ./build/TwinAim.exe --curriculum=./curricula/Twin_aim.json --run-id=Twin_aim --train\n\n### Curriculum variables\nTo help the training some variables change over time depending on reward thresholds\n- spawn_rate: number of step between 2 enemy spawns\n- min_enemy_distance: minimum distance allowed to spawn\n- max_enemy_count: maximum of simultaneous enemy\n- reset_center_prob: probability that the agent spawn at the center of the arena\n- skip_internal_brain: disable the other brain during early stage of the training\n\n## Informations\nThis repository contains UnityML 0.4 (https://github.com/Unity-Technologies/ml-agents)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmplantady%2Ftwinstickrl","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fmplantady%2Ftwinstickrl","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmplantady%2Ftwinstickrl/lists"}