feature regularization issues and total loss?

mengmenm / SMIL

Pytorch implementation of SMIL: Multimodal Learning with Severely Missing Modality (AAAI 2021)

84 stars 8 forks source link

Open HYC01 opened 1 month ago

HYC01 commented 1 month ago

why sigma and std is not used in fc1 and fc2?

And why sound_mean_val_mse is not used?