Output prediction mask is empty

[junxiant]

hi everyone,

I'm trying to do a 2D segmentation task by using my own pretrained model which i have
trained using pytorch. I have followed the instructions from https://github.com/Project-MONAI/MONAILabel/tree/main/sample-apps/radiology#how-to-add-new-model. My .pt model is under /model/pretrained_{name}.pt .

Everything looks fine and runs without error,
but the output prediction masks are empty.

For example, this is the output using MonaiLabel's:

This is the output when i ran my own test.py using pytorch:

Did i miss out something when trying to implement my own transforms?

[junxiant]

Solved with an issue with Normalization in pre_transforms