Closed leeqiaogithub closed 2 years ago
There is a slight problem in file resnet.py. It seems the parameter "pretrained" is no use. I wonder if it was a mistake or deliberate.If it's on purpose,Why don't we need to load pre-training weights?
The pre-trained loading in resnet.py is useless since the loading of pretrained DeepLabV3+ includes the parameters of ResNet.
I see,thanks for your reply
There is a slight problem in file resnet.py. It seems the parameter "pretrained" is no use. I wonder if it was a mistake or deliberate.If it's on purpose,Why don't we need to load pre-training weights?