mobilenetv2 backbone is wrong

Xiaoccer / MobileFaceNet_Pytorch

MobileFaceNets: Efficient CNNs for Accurate Real-Time Face Verification on Mobile Devices

355 stars 89 forks source link

Open anguoyang opened 4 years ago

anguoyang commented 4 years ago

Given groups=1, weight of size 512 128 1 1, expected input[2, 320, 4, 4] to have 128 channels, but got 320 channels instead

anguoyang commented 4 years ago

Mobilenetv1 is correct, but V2 is wrong

Xiaoccer commented 4 years ago

Sorry, I didn't understand your question too much, can you describe it in detail？