Skip to content

Latest commit

 

History

History
79 lines (68 loc) · 3.75 KB

CHANGELOG.md

File metadata and controls

79 lines (68 loc) · 3.75 KB
Raw

Changelog

Notable changes to this project will be documented in this file.

The format is based on Keep a Changelog and this project adheres to Semantic Versioning. https://github.com/StanfordVL/GibsonEnv/blob/master/README.md#

0.7.0 - 2021-05-19

  • Now Gibson is published in PYPI
  • Added python utilities to download the core assets data and environments dataset. You can easily download them typing 
      gibson-set-assets-path # This command allows you to set the default Gibson assets folder
  gibson-download-assets-core
  gibson-download-dataset
    on a terminal after Gibson installation. The downloaded data are automatically copied inside the default assets folder (set by the user).
  • Optimized the Github actions for continuous integration.

0.6.0 - 2021-05-11

  • Improved the build procedure and published the Gibson compiled version on pypi (for now test pypi)
  • The build procedure is performed directly from setup.py
  • Added submodule to the repository
  • Added a github action which builds the project (using Python 3.6, 3.7, 3.8, 3.9) and publishes the compiled version on pypi
  • Added Github actions to build Gibson following the manylinux standard. In particular, Gibson is compiled following PEP 571 (manylinux2010) and PEP 599 (manylinux2014).

0.5.0 - 2021-05-03

Added

  • Changed the assets default path: now it must be chosen by the user using gibson-set-assets-path command
  • Deleted the assets data folder: all files are contained in assets-core-v2.tar.gz
  • Created AssetsManager: it stores the assets path and other configuration parameters

Issues

  • The docker image must be fixed.
  • The docker procedure to install Gibson must be updated.

0.4.0 - 2021-04-29

Added

  • Build procedure tested in Ubuntu 20.04 LTS
  • Fixed build issues
  • Fixed dependencies issues
  • Fixed CMake issues
  • Added new environment: TurtlebotNavigateNoPhysicsEnv. In this environment, all physical constraints are deleted. The gravity is set to zero and the collisions between the agent and the environment are ignored. Therefore, the only way to move the agent is to manually set its position at each stage of the simulation.

0.3.1 - 2018-08-11

Added

  • EGL integration, remove X server dependency (solve #16 #24 #25)
  • OpenAI Gym 0.10.5 compatibility
  • Updated rendering filler models, added unfiller models
  • Bug fixes

0.3.0 - 2018-06-18

Added

  • Full dataset
  • ROS integration
  • Misc bug fixes

0.2.1 - 2018-04-18

Bug fixes

Fixed

  • Bug reported by @jackbruce. Random initialization for robot initial position.

0.2.0 - 2018-03-13

MINOR adds and PATCH fixes. Commit

Added

  • Environment offers self.robot API for state, observation, orientation, action, eye, reset, etc.
  • You can define your own environment with customized rewards. See Semantic README for instructions.

Changed

  • env.step(action) now returns as first value obs a dictionary. You can get each component by obs['rgb_filled'], obs['depth'].
  • RL example code modified to reflect the new return value of env.step(action).
  • is_discrete is specified inside configuration file, instead of env class initialization. Keeping it clean.

Fixed

  • Intricacy in environment class inheritance. Now parent class does not assume child class attributes.
  • Issue with rendering rgb_prefilled, rgb_filled.
  • Resolved issue of agent being blocked by a 'transparent wall'.
  • Removed unnecessary logging.

0.1.0 - 2018-02-26

Initial beta release.