Closed dreamedrainbow closed 4 years ago
Hello, I got this issue as well when I use my own .pth file (pretrained vgg16 model on VOC datasets). And I changed some layers of the original vgg16 model structure. Have you solved this problem already? I would so appreciate it if you can share your experience with me.
Hi,there! When running test_net.py and demo.py on coco dataset with faster_rcnn_1_10_9771.pth(the pretrained resnet101 model on coco dataset provided by jwyang), I encounter the same error below :
I tried the most useful way described here: to __C.ANCHOR_SCALES = [4,8,16,32] also tried modifying the classes of pascal voc dataset to the classes of coco datasets
but the error still here. Anyone help?