Jonas Jankaitis

Results 2 comments of Jonas Jankaitis

Yolov9-s is slower than Yolov9-c. Why is this?

I ran some yolov7 vs yolov9 tests. Inference times on RTX 3050: yolov7.pt - 19 ms; yolov7-tiny.pt - 9 ms; yolov9-c-converted.pt - 19 ms; yolov9-m-converted.pt - 17 ms; yolov9-s-converted.pt -...

Significant Precision Loss After Converting to FP16 in TensorRT Inference

in export_onnx.py, changing opset_version=16 to opset_version=17 fixed it for me