Open gztleo opened 6 years ago
Dear, I think your architecture of u-net is not correct for the beginning layers. I use it to train the same dataset, but the loss doesn't converge.Could you upload your trained model?Thank you!
for keras has updated to 2.1.5 now
Dear, I think your architecture of u-net is not correct for the beginning layers. I use it to train the same dataset, but the loss doesn't converge.Could you upload your trained model?Thank you!