Spaces:
Sleeping
Sleeping
Commit ·
2ced530
1
Parent(s): 49913c2
fix: reduce step penalty from -1 to -0.1
Browse filesStep penalty of -1 was punishing the model for surviving longer,
making death the optimal strategy. Reduced to -0.1.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
src/tetris_env/server/game_engine.py
CHANGED
|
@@ -37,7 +37,7 @@ LINE_REWARDS = {
|
|
| 37 |
4: 1500, # "Tetris!" — the dream
|
| 38 |
}
|
| 39 |
|
| 40 |
-
STEP_PENALTY = -1
|
| 41 |
HOLE_PENALTY_MULT = -5
|
| 42 |
GAME_OVER_PENALTY = -500
|
| 43 |
|
|
|
|
| 37 |
4: 1500, # "Tetris!" — the dream
|
| 38 |
}
|
| 39 |
|
| 40 |
+
STEP_PENALTY = -0.1
|
| 41 |
HOLE_PENALTY_MULT = -5
|
| 42 |
GAME_OVER_PENALTY = -500
|
| 43 |
|