AI GRAND PRIX
RACE
GATE
—
LAP
1/2
0:00
SPEED
0
mph
LAP HISTORY
TRAINING DATA
RL POLICY MODEL
best_model (recommended)
final_model (2M steps)
checkpoint 1.7M steps
checkpoint 1.5M steps
checkpoint 1.0M steps
checkpoint 500K steps
VISION MODEL
YOLO + U-Net (full pipeline)
YOLO only (detector)
U-Net only (segmentation)
MODEL INFO
Policy
best_model
Steps
2.0M
Reward
+1,120
Vision
YOLO+U-Net
YOLO mAP
99.5%
ONNX
3/3 ready
0 mph
RACE COMPLETE
—
RUN AGAIN
NEW RACE
PX4 AUTOPILOT v1.14.3 — DEBUG