Synth¶


Action Space	`Discrete(7)`
Observation Space	`Dict('direction': Discrete(4), 'image': Box(0, 255, (7, 7, 3), uint8), 'mission': MissionSpace(<function BabyAIMissionSpace._gen_mission at 0x7fb62a848820>, None))`
Reward Range	`(0, 1)`
Creation	`gymnasium.make("BabyAI-SynthS5R2-v0")`

Description¶

Union of all instructions from PutNext, Open, Goto and PickUp. The agent may need to move objects around. The agent may have to unlock the door, but only if it is explicitly referred by the instruction.

Competencies: Maze, Unblock, Unlock, GoTo, PickUp, PutNext, Open

Mission Space¶

“go to the {color} {type}”

“pick up a/the {color} {type}”

“open the {color} door”

“put the {color} {type} next to the {color} {type}”

{color} is the color of the box. Can be “red”, “green”, “blue”, “purple”, “yellow” or “grey”.

{type} is the type of the object. Can be “ball”, “box” or “key”.

Action Space¶

Num	Name	Action
0	left	Turn left
1	right	Turn right
2	forward	Move forward
3	pickup	Pick up an object
4	drop	Unused
5	toggle	Unused
6	done	Unused

Observation Encoding¶

Each tile is encoded as a 3 dimensional tuple: (OBJECT_IDX, COLOR_IDX, STATE)
OBJECT_TO_IDX and COLOR_TO_IDX mapping can be found in minigrid/core/constants.py
STATE refers to the door state with 0=open, 1=closed and 2=locked

Rewards¶

A reward of ‘1 - 0.9 * (step_count / max_steps)’ is given for success, and ‘0’ for failure.

Termination¶

The episode ends if any one of the following conditions is met:

The agent achieves the task.
Timeout (see max_steps).

Registered Configurations¶

BabyAI-Synth-v0
BabyAI-SynthS5R2-v0