`evorl.envs.jaxmarl`¶

Module Contents¶

Classes¶

`JaxMARLAdapter`
`MABraxAdapter`

Functions¶

`create_mabrax_env`
`create_wrapped_mabrax_env`
`get_random_actions`
`jaxmarl_space_to_evorl_space`

Data¶

`ma_brax_env_list`
`supported_jaxmarl_env_list`

API¶

class evorl.envs.jaxmarl.JaxMARLAdapter(env: jaxmarl.environments.MultiAgentEnv)¶

Bases: evorl.envs.multi_agent_env.MultiAgentEnvAdapter

property action_space: collections.abc.Mapping[evorl.types.AgentID, evorl.envs.space.Space]¶

property agents: list[evorl.types.AgentID]¶

property obs_space: collections.abc.Mapping[evorl.types.AgentID, evorl.envs.space.Space]¶

reset(key: chex.PRNGKey) → evorl.envs.env.EnvState¶

step(state: evorl.envs.env.EnvState, action: evorl.types.Action) → evorl.envs.env.EnvState¶

class evorl.envs.jaxmarl.MABraxAdapter(env: jaxmarl.environments.MultiAgentEnv)¶

Bases: evorl.envs.jaxmarl.JaxMARLAdapter

reset(key: chex.PRNGKey) → evorl.envs.env.EnvState¶

step(state: evorl.envs.env.EnvState, action: evorl.types.Action) → evorl.envs.env.EnvState¶

evorl.envs.jaxmarl.create_mabrax_env(env_name: str, **kwargs) → evorl.envs.jaxmarl.JaxMARLAdapter¶

evorl.envs.jaxmarl.create_wrapped_mabrax_env(env_name: str, episode_length: int = 1000, parallel: int = 1, autoreset_mode: evorl.envs.wrappers.training_wrapper.AutoresetMode = AutoresetMode.NORMAL, **kwargs) → evorl.envs.jaxmarl.JaxMARLAdapter¶

evorl.envs.jaxmarl.get_random_actions(env: jaxmarl.environments.MultiAgentEnv)¶

evorl.envs.jaxmarl.jaxmarl_space_to_evorl_space(space)¶

evorl.envs.jaxmarl.ma_brax_env_list¶: (‘ant_4x2’, ‘halfcheetah_6x1’, ‘hopper_3x1’, ‘humanoid_9|8’, ‘walker2d_2x3’)

evorl.envs.jaxmarl.supported_jaxmarl_env_list¶: (‘MPE_simple_v3’, ‘MPE_simple_tag_v3’, ‘MPE_simple_world_comm_v3’, ‘MPE_simple_spread_v3’, ‘MPE_simp…