search

JohnleeHIT / Brats2019

[BrainLes2019] Multi-step cascaded network for brain tumor segmentations (tensorflow)
https://arxiv.org/abs/1908.05887
MIT License
95 stars 24 forks source link

You load pre-weights only for the first mode named unet,did you? #5

Open zwlshine opened 4 years ago

zwlshine commented 4 years ago

In your code file "oprations.py", you extract some layers for fine-tuning, and I found these layers only from the first mode unet. Does it means just do fine-tuning for unet mode , but not do for unet_resnet mode?

New learner of tensorflow, I don't really understand fine-tuning in tensorflow, thank you very much!

JohnleeHIT commented 4 years ago

Yes, you are right, I only fine tune some of the layers of the mode unet,

zwlshine commented 4 years ago

Yes, you are right, I only fine tune some of the layers of the mode unet,

Why not use the pre-training weights for all modes, maybe can get better dice result.