HappyColor / SpeechFormer

Official implement of SpeechFormer written in Python (PyTorch).
75 stars 7 forks source link

Hi, I'm having an overfit problem #7

Open xuzhuochang opened 10 months ago

xuzhuochang commented 10 months ago

I ran your code on the DAIC data set, the model is SpeechFormer-S, the audio features are logmel, and there is overfitting. Specifically, the loss in the training phase can be reduced, but the loss in the testing phase is getting higher and higher. Could you please give me some suggestions?