dice on training set 0.85 and val set is 0.007?

MrBlankness / LightM-UNet

Pytorch implementation of "LightM-UNet: Mamba Assists in Lightweight UNet for Medical Image Segmentation"

https://arxiv.org/abs/2403.05246

Apache License 2.0

279 stars 25 forks source link

dice on training set 0.85 and val set is 0.007? #25

Open Caipengzhou opened 5 months ago

Caipengzhou commented 5 months ago

Hello, I follow the steps of readme to train and reason, but the results on the training set (dice:0.85) and the verification set (dice:0.007) are very different. Note: The training set and the val set are fixed, there is no random division, what is the reason?

shayan1999 commented 2 months ago

Maybe your model is became overfit or your data is imbalance(maybe data augmentation can help you here), etc

p.n I considered that you want to use this model for another dataset