-
Notifications
You must be signed in to change notification settings - Fork 93
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
1 changed file
with
51 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,51 @@ | ||
# OvercookedV2 | ||
|
||
OvercookedV2 is an extended version of the original Overcooked environment. It introduces meaningful partial observability and increased stochasticity to enable more complex coordination challenges. | ||
|
||
## Key Features | ||
|
||
- Configurable agent view radius | ||
- Multiple ingredients and recipes | ||
- Asymmetric information through recipe indicators | ||
- Randomized starting positions and directions | ||
- Grounded communication via button recipe indicators | ||
- Flexible layout creation | ||
|
||
## Custom Layouts | ||
|
||
Create custom layouts using ASCII strings: | ||
|
||
```python | ||
layout = """ | ||
WWPWW | ||
0A A1 | ||
L R | ||
WBWXW | ||
""" | ||
recipes = [[0,0,1], [0,1,1]] | ||
custom_layout = Layout.from_string(layout, possible_recipes=recipes) | ||
``` | ||
|
||
## Pre-configured Layouts | ||
|
||
See `layouts.py` for a variety of pre-configured layouts: | ||
|
||
- Adaptations of original Overcooked layouts | ||
- Extended Cat-Dog problem layouts | ||
- Test-time coordination challenge layouts | ||
|
||
## Observations | ||
|
||
Observations are structured as a width x height x num_channels tensor, with partial observability based on the configured view radius. | ||
|
||
## Rewards | ||
|
||
- 20 points for correct deliveries | ||
- Optional -20 points for incorrect deliveries | ||
- Shaped rewards for actions aligned with the current recipe | ||
|
||
## Visualization and Interactive Play | ||
|
||
- JIT-compiled rendering pipeline for efficient episode visualization | ||
- Interactive mode for playing alongside trained policies | ||
|