weight file appears to be corrupted?

emma-sjwang / BEAL

code for paper Boundary and Entropy-driven Adversarial Learning for Fundus Image Segmentation

MIT License

121 stars 25 forks source link

weight file appears to be corrupted? #6

Closed JACKYLUO1991 closed 4 years ago

JACKYLUO1991 commented 4 years ago

Thanks for your work. There was a problem with downloading the weights. Is it now possible to download the weights? Or is it my network transmission? I look forward to hearing from you...

JACKYLUO1991 commented 4 years ago

And why +1 in utils/Utils.py in line 110?

emma-sjwang commented 4 years ago

After label, the background is set as 0. The regiosn in the foreground will be set as 1,2,3... But the area id is from 0. The background is already 0. We only need to set other non-largest areas to be 0.

JACKYLUO1991 commented 4 years ago

thanks

emma-sjwang commented 4 years ago

Have you downloaded the weight files sucessful? The weights are stored in Google drive. If you are in mainland, you may need VPN to access it.

JACKYLUO1991 commented 4 years ago

@EmmaW8 I can download, but the weight file is damaged, it doesn't matter, I have retrained a version, thank you for your guidance.