How can I get NMS_THRESH and CONF_THRESH inferenced from config files

[theiot]

@marcoslucianops

Hi, I want to make nvdsparsebbox_Yolo more flexible by remove：

static const int NUM_CLASSES_YOLO = 80;
#define NMS_THRESH 0.45
#define CONF_THRESH 0.25

Right now I just cannot get NMS_THRESH from config file "config_infer_primary_yoloV5s.txt:
[class-attrs-all]
nms-iou-threshold=0.45
pre-cluster-threshold=0.25

Do you have any suggestion?

Thanks,

[marcoslucianops]

Hi, i'm working on it now (parse some fuctions based on cfg and config.txt file) and I will update the repo soon.

I tested the mAP and FPS in some YOLO models, and I recommend you see my readme.md. YOLOv4 performs better when compared to YOLOv5 on TensorRT/DeepStream.

With my files, you can directly convert YOLOv4 to DeepStream TensorRT model.

[theiot]

Hi, i'm working on it now (parse some fuctions based on cfg and config.txt file) and I will update the repo soon.

I tested the mAP and FPS in some YOLO models, and I recommend you see my readme.md. YOLOv4 performs better when compared to YOLOv5 on TensorRT/DeepStream.

With my files, you can directly convert YOLOv4 to DeepStream TensorRT model.

Glad to hear and looking forward to it.

[marcoslucianops]

CONF_THRESH (pre-cluster-threshold)

const float kCONF_THRESH = detectionParams.perClassThreshold[0];

NUM_CLASSES_YOLO (num-detected-classes) -> Not needed in YOLOv5 code

const uint kNUM_CLASSES = detectionParams.numClassesConfigured;

NMS_THRESH cannot be obtained from config.txt file in current DeepStream code.