[rllib] [rfc] add contrib module and guideline for merging #3565
Conversation
ericl
changed the title
|
Test FAILed. |
|
Test FAILed. |
|
Test FAILed. |
|
Test FAILed. |
| - should include tuned hyperparameter examples and documentation | ||
| - should offer functionality not present in existing algorithms | ||
|
|
||
| * Fully integrated algorithms (`rllib/agents <https://github.com/ray-project/ray/tree/master/python/ray/rllib/agents>`__) have the following additional requirements: |
There was a problem hiding this comment.
Do we want contributions of non-fully integrated algorithms?
There was a problem hiding this comment.
I see it as strictly positive given the guidelines here. A lot of the value from having a suite of integrated algorithms comes just from having a uniform API to run them. Leveraging the other features of RLlib is a second order benefit that comes with fuller integration.
|
Test FAILed. |
|
Test FAILed. |
|
Test FAILed. |
|
Test FAILed. |
| @@ -20,6 +20,7 @@ APEX-DQN **Yes** `+parametric`_ No **Yes** No | |||
| APEX-DDPG No **Yes** **Yes** No | |||
| ES **Yes** **Yes** No No | |||
| ARS **Yes** **Yes** No No | |||
| QMIX **Yes** No **Yes** **Yes** | |||
| return RandomAgent | ||
|
|
||
| CONTRIBUTED_ALGORITHMS = { | ||
| "contrib/RandomAgent": _import_random_agent, |
There was a problem hiding this comment.
on the docs, show two agents (is currently confusing as is)
| .. code-block:: bash | ||
|
|
||
| rllib train --run=contrib/RandomAgent --env=CartPole-v0 | ||
| tensorboard --logdir=~/ray_results |
There was a problem hiding this comment.
should add a link for questions
|
(will merge after lint) |
|
Test FAILed. |
|
Test FAILed. |
|
Test FAILed. |
|
jenkins, retest this please |
|
Test FAILed. |
What do these changes do?
This adds guidelines for merging code into
rllib/contribvsrllib/agents. Also, clean up the agent import code to make registration easier.