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.

Sign up for GitHub

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

Jump to bottom

improve gym env / tune hps #71

Merged
merged 33 commits into from
Feb 1, 2024
Merged

improve gym env / tune hps #71

merged 33 commits into from
Feb 1, 2024

Commits on Jan 23, 2024

  1. use pid to speed up real env reset

    @Armandpl
    Armandpl committed Jan 23, 2024
    Configuration menu
    Copy the full SHA
    9086aed View commit details
    Browse the repository at this point in the history

  2. setup deadzone wrapper

    @Armandpl
    Armandpl committed Jan 23, 2024
    Configuration menu
    Copy the full SHA
    01451f4 View commit details
    Browse the repository at this point in the history

  3. setup tqdc + deadzone sim exp

    @Armandpl
    Armandpl committed Jan 23, 2024
    Configuration menu
    Copy the full SHA
    9e7d5fa View commit details
    Browse the repository at this point in the history

  4. add angles to the obs when limiting them; add time feat wrapper

    @Armandpl
    Armandpl committed Jan 23, 2024
    Configuration menu
    Copy the full SHA
    347befd View commit details
    Browse the repository at this point in the history

  5. check for nan instead of none bc none gets converted to nan by np.arr… 

    …ay w/ dtype float32
    @Armandpl
    Armandpl committed Jan 23, 2024
    Configuration menu
    Copy the full SHA
    cd9cf1c View commit details
    Browse the repository at this point in the history

Commits on Jan 24, 2024

  1. setup real exp

    @Armandpl
    Armandpl committed Jan 24, 2024
    Configuration menu
    Copy the full SHA
    5c4b6c4 View commit details
    Browse the repository at this point in the history

  2. add deadzone wrapper

    @Armandpl
    Armandpl committed Jan 24, 2024
    Configuration menu
    Copy the full SHA
    e2d6360 View commit details
    Browse the repository at this point in the history

Commits on Jan 25, 2024

  1. use cos for env reset bc pendulum can do multiple turns

    @Armandpl
    Armandpl committed Jan 25, 2024
    Configuration menu
    Copy the full SHA
    99a80dd View commit details
    Browse the repository at this point in the history

  2. add a center to the deadzone wrapper to allow for a zero action

    @Armandpl
    Armandpl committed Jan 25, 2024
    Configuration menu
    Copy the full SHA
    df6cf31 View commit details
    Browse the repository at this point in the history

  3. add exp reward

    @Armandpl
    Armandpl committed Jan 25, 2024
    Configuration menu
    Copy the full SHA
    b5bdca6 View commit details
    Browse the repository at this point in the history

  4. setup tqc + exp reward sim experiment

    @Armandpl
    Armandpl committed Jan 25, 2024
    Configuration menu
    Copy the full SHA
    3e00fae View commit details
    Browse the repository at this point in the history

  5. add sbx and jax-metal

    @Armandpl
    Armandpl committed Jan 25, 2024
    Configuration menu
    Copy the full SHA
    06116f7 View commit details
    Browse the repository at this point in the history

  6. fix base wrapper; make sim exp closer to real w/ n_envs=1

    @Armandpl
    Armandpl committed Jan 25, 2024
    Configuration menu
    Copy the full SHA
    7c9f844 View commit details
    Browse the repository at this point in the history

  7. add timeout to motor stopping, hack to avoid pid oscillations instead… 

    … of tuning it
    @Armandpl
    Armandpl committed Jan 25, 2024
    Configuration menu
    Copy the full SHA
    22c5a43 View commit details
    Browse the repository at this point in the history

  8. setup rew sweep

    @Armandpl
    Armandpl committed Jan 25, 2024
    Configuration menu
    Copy the full SHA
    8a62cdc View commit details
    Browse the repository at this point in the history

Commits on Jan 29, 2024

  1. investigate tqc action looking like gaussian noise

    @Armandpl
    Armandpl committed Jan 29, 2024
    Configuration menu
    Copy the full SHA
    1b925e6 View commit details
    Browse the repository at this point in the history

  2. fix default args

    @Armandpl
    Armandpl committed Jan 29, 2024
    Configuration menu
    Copy the full SHA
    d650c03 View commit details
    Browse the repository at this point in the history

  3. add max act to deadzone wrapper

    @Armandpl
    Armandpl committed Jan 29, 2024
    Configuration menu
    Copy the full SHA
    0f6c7ad View commit details
    Browse the repository at this point in the history

  4. setup hp tune sweep for tqc

    @Armandpl
    Armandpl committed Jan 29, 2024
    Configuration menu
    Copy the full SHA
    5503027 View commit details
    Browse the repository at this point in the history

  5. actually use early stopping

    @Armandpl
    Armandpl committed Jan 29, 2024
    Configuration menu
    Copy the full SHA
    5b7e11d View commit details
    Browse the repository at this point in the history

Commits on Jan 30, 2024

  1. use sb3_contrib master to train

    @Armandpl
    Armandpl committed Jan 30, 2024
    Configuration menu
    Copy the full SHA
    a46febd View commit details
    Browse the repository at this point in the history

  2. use progressbar

    @Armandpl
    Armandpl committed Jan 30, 2024
    Configuration menu
    Copy the full SHA
    e3d3761 View commit details
    Browse the repository at this point in the history

Commits on Jan 31, 2024

  1. add option to use either sbx or sbx tqc

    @Armandpl
    Armandpl committed Jan 31, 2024
    Configuration menu
    Copy the full SHA
    2186f79 View commit details
    Browse the repository at this point in the history

  2. add new reward taking theta into account

    @Armandpl
    Armandpl committed Jan 31, 2024
    Configuration menu
    Copy the full SHA
    df26804 View commit details
    Browse the repository at this point in the history

  3. remove unused code

    @Armandpl
    Armandpl committed Jan 31, 2024
    Configuration menu
    Copy the full SHA
    f679592 View commit details
    Browse the repository at this point in the history

  4. make progressbar a param

    @Armandpl
    Armandpl committed Jan 31, 2024
    Configuration menu
    Copy the full SHA
    722456e View commit details
    Browse the repository at this point in the history

  5. setup new rew sweep

    @Armandpl
    Armandpl committed Jan 31, 2024
    Configuration menu
    Copy the full SHA
    e751650 View commit details
    Browse the repository at this point in the history

  6. fix error in reward

    @Armandpl
    Armandpl committed Jan 31, 2024
    Configuration menu
    Copy the full SHA
    2d76294 View commit details
    Browse the repository at this point in the history

  7. setup real training

    @Armandpl
    Armandpl committed Jan 31, 2024
    Configuration menu
    Copy the full SHA
    5c933d0 View commit details
    Browse the repository at this point in the history

  8. fix sweep setup

    @Armandpl
    Armandpl committed Jan 31, 2024
    Configuration menu
    Copy the full SHA
    9759dc8 View commit details
    Browse the repository at this point in the history

Commits on Feb 1, 2024

  1. add theta reward

    @Armandpl
    Armandpl committed Feb 1, 2024
    Configuration menu
    Copy the full SHA
    5986850 View commit details
    Browse the repository at this point in the history

  2. clear gsde notebook output

    @Armandpl
    Armandpl committed Feb 1, 2024
    Configuration menu
    Copy the full SHA
    9070298 View commit details
    Browse the repository at this point in the history

  3. fix default reward key

    @Armandpl
    Armandpl committed Feb 1, 2024
    Configuration menu
    Copy the full SHA
    bcda11d View commit details
    Browse the repository at this point in the history