Closed ChristopheKar closed 2 years ago
This allows for models like YOLOv4-P5 with 4 anchors in each YOLO layer to be converted to ONNX and then to TensorRT, by simply parsing the configuration file. Tested for YOLOv4-P5 with dimensions 896 (original config dimensions) and 448.
This allows for models like YOLOv4-P5 with 4 anchors in each YOLO layer to be converted to ONNX and then to TensorRT, by simply parsing the configuration file. Tested for YOLOv4-P5 with dimensions 896 (original config dimensions) and 448.