search

clovaai / voxceleb_trainer

In defence of metric learning for speaker recognition
MIT License
1.03k stars 272 forks source link

Fast ResNet-34? #51

Closed XinhaoMei closed 4 years ago

XinhaoMei commented 4 years ago

Hi, I didn't find Fast ResNet-34 in the paper. After comparing it with the Thin ResNet-34 in the code, are the differences between them only the input features(Spectrogram for Thin ResNet-34, MelSpectrogram for Fast ResNet-34) and the lack of maxpool layer in Fast ResNet-34?

joonson commented 4 years ago

Yes, these are the only differences.

gancx commented 1 year ago

Hi, I didn't find Fast ResNet-34 in the paper. After comparing it with the Thin ResNet-34 in the code, are the differences between them only the input features(Spectrogram for Thin ResNet-34, MelSpectrogram for Fast ResNet-34) and the lack of maxpool layer in Fast ResNet-34?

I wonder where you can the code of Thin ResNet34. i found the implementation is different from that in the original paper.