PCA-Net
PCA-Net copied to clipboard
why take first 32 channels as attention map in model.py at line 112
could you tell me why just take 32 channels feature map as attention map for fusioning?
@TIan1874