mmlab-cv/Agglomerator

Question about training parameters

Opened this issue · 0 comments

Hello,

Could you please tell how you get 32x32 pixels levels images during plot visualization while you have conv_image_size set to 8 in the config file and also tell the same in the paper?