SegDiff

SegDiff copied to clipboard

lg_loss_scale continue declining in the training step on the MODU Dataset

1

Thanks for your good work! When I was training the MODU dataset, I encountered a situation where lg_loss_scale kept declining and the gradient kept being unstable. Please tell me how...

CV-poo

Dear authors: When i try to run the vaih.py with image_size = 512 with changing the script_util.py and adding the if image_size = 512: if deeper_net: channel_mult = (1,1,1,2,2,4,4,8) else:...

egcin

i don't know why this NaN appears during training. Is it caused by the dataset? | grad_norm | 2.34e+05 | | lg_loss_scale | -0.821 | | loss | 5.99e+04 |...

1245179435