# AgentFlow: A Modular Toolkit for Scalable RL Research
<!--* B 2021-07-21 internal placeholder *-->
## Overview
`AgentFlow` is a library for composing Reinforcement-Learning agents. The core
features that AgentFlow provides are:
1. tools for slicing, transforming, and composing *specs*
2. tools for encapsulating and composing RL-tasks.
Unlike the standard RL setup, which assumes a single environment and an agent,
`AgentFlow` is designed for the single-embodiment, multiple-task regime. This
was motivated by the robotics use-case, which frequently requires training RL
modules for various skills, and then composing them (possibly with non-learned
controllers too).
Instead of having to implement a separate RL environment for each skill and
combine them ad hoc, with `AgentFlow` you can define one or more `SubTasks`
which *modify* a timestep from a single top-level environment, e.g. adding
observations and defining rewards, or isolating a particular sub-system of the
environment, such as a robot arm.
You then *compose* SubTasks with regular RL-agents to form modules, and use a
set of graph-building operators to define the flow of these modules over time
(hence the name `AgentFlow`).
The graph-building step is entirely optional, and is intended only for use-cases
that require something like a (possibly learnable, possibly stochastic)
state-machine.
<!-- Internal placeholder C -->
### [Components](docs/components.md)
### [Control Flow](docs/control_flow.md)
### [Examples](docs/examples.md)
<!-- Internal placeholder D -->
Raw data
{
"_id": null,
"home_page": "https://github.com/deepmind/dm_robotics/tree/main/py/agentflow",
"name": "dm-robotics-agentflow",
"maintainer": "",
"docs_url": null,
"requires_python": ">=3.7, <3.11",
"maintainer_email": "",
"keywords": "",
"author": "DeepMind",
"author_email": "",
"download_url": "",
"platform": null,
"description": "# AgentFlow: A Modular Toolkit for Scalable RL Research\n\n<!--* B 2021-07-21 internal placeholder *-->\n\n## Overview\n\n`AgentFlow` is a library for composing Reinforcement-Learning agents. The core\nfeatures that AgentFlow provides are:\n\n1. tools for slicing, transforming, and composing *specs*\n2. tools for encapsulating and composing RL-tasks.\n\nUnlike the standard RL setup, which assumes a single environment and an agent,\n`AgentFlow` is designed for the single-embodiment, multiple-task regime. This\nwas motivated by the robotics use-case, which frequently requires training RL\nmodules for various skills, and then composing them (possibly with non-learned\ncontrollers too).\n\nInstead of having to implement a separate RL environment for each skill and\ncombine them ad hoc, with `AgentFlow` you can define one or more `SubTasks`\nwhich *modify* a timestep from a single top-level environment, e.g. adding\nobservations and defining rewards, or isolating a particular sub-system of the\nenvironment, such as a robot arm.\n\nYou then *compose* SubTasks with regular RL-agents to form modules, and use a\nset of graph-building operators to define the flow of these modules over time\n(hence the name `AgentFlow`).\n\nThe graph-building step is entirely optional, and is intended only for use-cases\nthat require something like a (possibly learnable, possibly stochastic)\nstate-machine.\n\n<!-- Internal placeholder C -->\n### [Components](docs/components.md)\n### [Control Flow](docs/control_flow.md)\n### [Examples](docs/examples.md)\n<!-- Internal placeholder D -->\n",
"bugtrack_url": null,
"license": "Apache 2.0",
"summary": "Tools for single-embodiment, multiple-task, Reinforcement Learning",
"version": "0.6.0",
"project_urls": {
"Homepage": "https://github.com/deepmind/dm_robotics/tree/main/py/agentflow"
},
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "e300b29154a63c4fb01aecfe8e34871b0b91c9423dd9465b5d4280c0f85b2c41",
"md5": "203b45b9895fd9e65a92a1149b42b5c6",
"sha256": "bf2ad7188da954edc88e3c8eba32961f3a98410c77dfdd8a3b117c4f1e7b36d1"
},
"downloads": -1,
"filename": "dm_robotics_agentflow-0.6.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "203b45b9895fd9e65a92a1149b42b5c6",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.7, <3.11",
"size": 143830,
"upload_time": "2023-10-31T15:00:12",
"upload_time_iso_8601": "2023-10-31T15:00:12.422436Z",
"url": "https://files.pythonhosted.org/packages/e3/00/b29154a63c4fb01aecfe8e34871b0b91c9423dd9465b5d4280c0f85b2c41/dm_robotics_agentflow-0.6.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-10-31 15:00:12",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "deepmind",
"github_project": "dm_robotics",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "dm-robotics-agentflow"
}
JSON
