I am doing image segmentation in 2D using nnUNet. I have 3D tif files, images and masks.
in ordrer to train the model and make the data fit to the pipeline of nnUNet I am slicing the tif files so each frame will be a separate tif file. so if an image and its matching mask are 150 at the first dimension,then after slicing I will get 150 frames (each with the same 2D shape).

Now I am preparing the data so it fits the pipline with name formatting etc.
Then, I run training on a dataset, let’s sat dataset1.

Now, if I want to make prediction on dataset2, by using the model that was trained on dataset1, how do I do that?
I am using the following command for prediction:
nnUNetv2_predict -d Dataset1 -i /path/to/test_set -o /path/to/predictions -f 0 1 2 3 4 -tr nnUNetTrainer -c 2d -p nnUNetPlans
I enter a path to the test set but how does the model know which training set was used for training? is it using the last one I have traind or how can I spicify which one I want to use?

thank you in advance

I tried to predict a test set by using the model that was trained on another train set from another image. And I am not sure if the results I get are correct since I do not know which trained model the prediction is using.

I also tried to find an answer for it here: nnUNet

but unfortentaly I have not find one.