Open siretru opened 2 weeks ago
Can I use the NvDsInferYolo parse function in Deepstream for a YoloV8 model built with yoloNMS?
Yes
The yoloNMS plugin has been renamed to EfficientNMSX to facilitate integration with TensorRT. For detection models, I use EfficientNMS, and for segmentation models, I use EfficientNMSX (formerly yoloNMS) and the roiAlign plugin.
Both plugins are supported by NvDsInferYolo.
The export of the model in End2End format is implemented in this repository: https://github.com/levipereira/ultralytics
EfficientNMSX (yoloNMS) returns detection indices necessary for segmentation models, while EfficientNMS does not return detection indices and is generally used only in detection models.
https://github.com/levipereira/TensorRT/tree/release/10.0/plugin/efficientNMSPlugin
Hello,
First of all, thank you for all these repositories. The work accomplished is just great.
I would like to ask you 1 question:
-Can I use the NvDsInferYolo parse function in Deepstream for a YoloV8 model built with yoloNMS?
Thank you !