Skip to content
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

First Pass of MAIRL #8

Closed
wants to merge 43 commits into from
Closed

First Pass of MAIRL #8

wants to merge 43 commits into from

Conversation

lalitlal
Copy link

Major In progress

  1. Create Reward Fn and Value Fn in Discriminator [DONE]
  • Inside init() or forward() -> added inside forward()
  1. [discriminator_irl.py] - modified discriminator.py [DONE]
  • Calculate log_p_tau, log_q_tau, log_pq for external losses
  1. [mgail.py] - modify how our discriminator is changed [DONE]
  1. [mgail.py] - modify al_loss to take into account new [DONE] discriminator output

  2. Modify expert data to include expert action probs [TODO]

  • Need this for lprobs
  1. Modify ER to account for new field 'action_probs' [TODO]

TO DO

  • Testing that we didn't break existing code
  • Testing the MAIRL actually compiles
  • Testing that MAIRL actually works as expected

Yip Sang Leung and others added 30 commits February 26, 2021 17:02
Garage scripts to train and run expert policy
first pass of using generated data
- Introduce WithActionObs and OnlyPartialObjAndColor wrappers
- Change max_kl_step to 0.001
- Change the hidden sizes of the policy
@lalitlal lalitlal closed this Mar 28, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants