Koulanurag Ma Gym Versions Save

A collection of multi agent environments based on OpenAI gym.

v0.0.14

11 months ago
  • checks/support for 3.10/3.11

v0.0.13

1 year ago
  • minor build changes

v0.0.11

1 year ago
  • dependency versions upgraded
  • non-critical initialization bug resolved in switch corridor

v0.0.10

1 year ago

minor version update to pyglet

v0.0.9

1 year ago
  • added checks for validating episode initialization and overflow.
  • Existing pre-trained agents calling reset() after all(dones) wouldn't be affected.
  • added reference to papers citing ma-gym

v0.0.8

2 years ago
  • Fixed multiple issues with Combat ( Thanks to @tjuHaoXiaotian )

v0.0.7

2 years ago
  • added TrafficJunction env (Thanks to @rafaelmp2 )
  • added clock version of switch and checkers.
  • reward eased for switch env. by having no step cost once the agent reaches the cell.
  • added on-hot representation for checkers observation.

v0.0.6

2 years ago
  • code refactor
  • small fix in switch environment( disabled agents are removed)
  • no change in performance of agents

v0.0.5

3 years ago
  • corrected packaging of sub-packages for pip install

v0.0.4

3 years ago