how to get the class label from output data

[pango99]

Hi,in postProcess() function,you extract the score & rect infor from output data, but how to get the class label info if my model has multi class?

[XuHao9166]

Hi,in postProcess() function,you extract the score & rect infor from output data, but how to get the class label info if my model has multi class?

do you finish this problem?how to get the class label?

[pango99]

Hi,in postProcess() function,you extract the score & rect infor from output data, but how to get the class label info if my model has multi class?

do you finish this problem?how to get the class label?

No,I still dont know how to get,now I'm using caffe version yolo

[XuHao9166]

Hi,in postProcess() function,you extract the score & rect infor from output data, but how to get the class label info if my model has multi class?

do you finish this problem?how to get the class label?

No,I still dont know how to get,now I'm using caffe version yolo caffe version yolo is good choose,but it hard to train model. i only use pytorch or darknet . how you train caffe model by yourself? i means train and change to onnx.

[pango99]

Hi,in postProcess() function,you extract the score & rect infor from output data, but how to get the class label info if my model has multi class?

do you finish this problem?how to get the class label?

No,I still dont know how to get,now I'm using caffe version yolo caffe version yolo is good choose,but it hard to train model. i only use pytorch or darknet . how you train caffe model by yourself? i means train and change to onnx.

I train model using darknet,and then convert yolo to caffe,so the model can be accelerated by tensorrt;the convert script can be found on github

[ghost]

I solve it with

int max_index = -1; float max_value = FLT_MIN; for (int k = 5; k < CATEGORY + 5; k++) { if (ptr[k] > max_value) { max_value = ptr[k]; max_index = k - 5; } }