-
Notifications
You must be signed in to change notification settings - Fork 505
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[hab3_merge] Define different policies for different agents #1530
Conversation
I'm blocked on some of the CI Tests. I don't know how to address this one for instance:
@rpartsey Could you check if I am missing/misunderstanding something in Hydra? |
obs_trans_conf = config.habitat_baselines.rl.policy.obs_transforms | ||
if hasattr(config.habitat_baselines.rl.policy, "obs_transforms"): | ||
|
||
# We assume for now that the observation space is shared among agents |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We don't assume this in general right? Only when using observation transformations?
).articulated_agent.get_joint_transform()[0] | ||
return np.array(joints_pos, dtype=np.float32) | ||
curr_agent = self._sim.get_agent_data(self.agent_id).articulated_agent | ||
if hasattr(curr_agent, "get_joint_transform"): |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Can you check for the type with is instance
instead of hasattr
? This will be more clear what interface you expect here.
…research#1530) * define policies per agent * fix * udpate configs to support multiagent * change configs to support agent name in polic * solving more pytest * correcting test * pytest solve * correct pytets * small comment update
…research#1530) * define policies per agent * fix * udpate configs to support multiagent * change configs to support agent name in polic * solving more pytest * correcting test * pytest solve * correct pytets * small comment update
Motivation and Context
This is needed so that we can define different policy types for each agent. It may be useful to be able to train a multi-agent task where one agent is run by a planner and the other one is run by a neural net, or one has a low level policy and the other has a high level policy.
Cherry-picked from this PR: 26a9696
How Has This Been Tested
Passes Test + End to End Testing.
Types of changes
Checklist