Training mobilenetv3-ssd model with coco datasets, and convert it to ncnn, then i changed the .param file manually. When I use ncnn to inference the image, i got the "detection_out" error, like this
I reference mobilenetv3_ssdlite_voc.param to change the .param, the output are 81 classes.
How to solve this problem, please help me! Thank you very much!
When i test the mobilenetv3_ssdlite_voc.param and mobilenetv3_ssdlite_voc.bin, it also has this problem,
test image is:
Whether the boxes(x=-nan,y=-nan,width=-nan,height=-nan) should be output?
Training mobilenetv3-ssd model with coco datasets, and convert it to ncnn, then i changed the .param file manually. When I use ncnn to inference the image, i got the "detection_out" error, like this
I reference mobilenetv3_ssdlite_voc.param to change the .param, the output are 81 classes.
How to solve this problem, please help me! Thank you very much!