search

uci-cbcl / NoduleNet

[MICCAI' 19] NoduleNet: Decoupled False Positive Reduction for Pulmonary Nodule Detection and Segmentation
Other
187 stars 58 forks source link

How to use the pretrained model? #42

Open yjsb opened 2 years ago

yjsb commented 2 years ago

I downloaded the file '200.ckpt' on Google Drive, and I changed its name to '001.ckpt' in my project. I trained the model for 50 epoches on fold 3, but I get a poor mean dice(exclude fn); 0.74, which in the paper is 0.83. So it maybe something wrong in my training procedure, can u help me, pls?

yjsb commented 2 years ago

The learning rate I used in my training procedure was 0.0001, I think it is small enough.

yjsb commented 2 years ago

I also tested it on 0_val.csv, and got the mean dice(exclude fn): 0.742. The model is download file '200.ckpt'.