caffe2pytorch-tsn
caffe2pytorch-tsn copied to clipboard
size mismatch for new_fc.weight
When i used the pretrain model of ucf101_rgb.pth, there is an error "size mismatch for new_fc.weight: copying a param with shape torch.Size([101, 1024]) from checkpoint, the shape in current model is torch.Size([101, 2048])". How can i solve it ?
Ok, i solve the problem. The reason is that basemodel i used resnet101, but the pretrained model is used BNInception.