{"id":18009141,"url":"https://github.com/kwea123/hindsight_experience_replay","last_synced_at":"2025-03-26T13:31:50.379Z","repository":{"id":109613263,"uuid":"124824418","full_name":"kwea123/hindsight_experience_replay","owner":"kwea123","description":"A tensorflow implementation of hindsight experience replay","archived":false,"fork":false,"pushed_at":"2018-04-19T08:33:31.000Z","size":1746,"stargazers_count":17,"open_issues_count":0,"forks_count":9,"subscribers_count":3,"default_branch":"master","last_synced_at":"2025-03-21T22:22:03.047Z","etag":null,"topics":["deep-q-learning","hindsight-experience-replay","jupyter-notebook","python","reinforcement-learning","tensorflow"],"latest_commit_sha":null,"homepage":"","language":"Jupyter Notebook","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/kwea123.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}},"created_at":"2018-03-12T02:44:42.000Z","updated_at":"2024-10-11T05:24:30.000Z","dependencies_parsed_at":"2023-06-19T20:45:12.644Z","dependency_job_id":null,"html_url":"https://github.com/kwea123/hindsight_experience_replay","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/kwea123%2Fhindsight_experience_replay","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kwea123%2Fhindsight_experience_replay/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kwea123%2Fhindsight_experience_replay/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kwea123%2Fhindsight_experience_replay/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/kwea123","download_url":"https://codeload.github.com/kwea123/hindsight_experience_replay/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":245662930,"owners_count":20652100,"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-q-learning","hindsight-experience-replay","jupyter-notebook","python","reinforcement-learning","tensorflow"],"created_at":"2024-10-30T02:08:49.418Z","updated_at":"2025-03-26T13:31:50.354Z","avatar_url":"https://github.com/kwea123.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Hindsight Experience Replay (HER)\n\nThis repository contains a tensorflow HER implementation and a bit flipping environment as described in [OpenAI's paper](https://arxiv.org/pdf/1707.01495.pdf)\n\nThe implementation includes :\n\n1. In `Hindsight Experience Replay.ipynb` :\n    1.  A DQN and a DDQN agent (which also work on other traditional [gym](https://gym.openai.com/) environments)\n    2.  A bit flipping environment\n    3.  Pre-trained models for 30-bits, 40-bits and 50-bits flipping environments\n2. In `ChaseEnv_DDPG.ipynb` :\n    1. A DDPG agent\n    2. A `ChaseEnv` environment, where a chaser is initialized at a random position in\n       a 2d plane and has to reach a goal in another random position within a certain threshold.\n\n## Benchmarks\n\n*  100% success rate for 30 and 40-bits environments\n*  95% success rate for 50-bits environment (average on 100 tests)\n*  90% success rate for size=5 ChaseEnv (average on 100 tests)\n\n## Customize\nCheck the \"Training\" cell to adjust training parameters and enable/disable HER.\n\n## TODO\n- [x] Optimize the way to concatenate transitions\n- [ ] Parallelize training\n- [x] Train on bit length \u003e 30\n- [x] Implement DDPG\n\n## Extra\n[Here](https://github.com/kwea123/RL/blob/master/ai/unity_test/robot_arm/robot_arm_3d_ddpg_her_sparse.ipynb) is a link to a robot arm reach environment created in Unity, trained with [ML-Agents](https://github.com/Unity-Technologies/ml-agents).\n\nThis environment is trained using DDPG with and without HER, and the comparison is plotted. DDPG+HER performs better.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fkwea123%2Fhindsight_experience_replay","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fkwea123%2Fhindsight_experience_replay","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fkwea123%2Fhindsight_experience_replay/lists"}