Dataset
Model Zoo
DeGirum’s fall detection model zoo features models trained with the Ultralytics repository.
Model Architecture |
Input Size |
Precision |
Runtime |
Device Type |
mAP 50-95 |
mAP 50 |
FPS |
yolov8n_relu6_fall |
640x640 |
INT8 |
N2X |
ORCA |
0.4297 |
0.8275 |
108.3 |
yolov8n_relu6_fall |
640x640 |
INT8 |
TFLITE |
EDGETPU |
0.4298 |
0.8279 |
– |
yolov8n_relu6_fall |
640x640 |
INT8 |
OPENVINO |
CPU |
0.4361 |
0.8311 |
– |
yolov8n_relu6_fall |
640x640 |
FP32 |
OPENVINO |
CPU |
0.4333 |
0.8297 |
– |
|
|
|
|
|
|
|
|
yolov8s_relu6_fall |
640x640 |
INT8 |
N2X |
ORCA |
0.4705 |
0.8647 |
39.1 |
yolov8s_relu6_fall |
640x640 |
INT8 |
TFLITE |
EDGETPU |
0.469 |
0.8641 |
– |
yolov8s_relu6_fall |
640x640 |
INT8 |
OPENVINO |
CPU |
0.4651 |
0.8617 |
– |
yolov8s_relu6_fall |
640x640 |
FP32 |
OPENVINO |
CPU |
0.4667 |
0.8629 |
– |