How to correctly run Synapse dataset with Mamba-UNet?

[ngodongnguyen]

Hello,
First of all, thank you very much for sharing this great work.
I’m trying to run Mamba-UNet on the Synapse dataset, but I’m not sure if my current training & testing steps are correct. I would like to ask for clarification from the authors and the community.

Here is what I did:

Step 1: Environment setup
I installed all the dependencies as described in the repository.
Step 2: Training
python train_fully_supervised_2D_VIM.py
--root_path ../data/Synapse
--exp Synapse/VIM
--model mambaunet
--max_iterations 10000
--batch_size 28
--num_classes 9
Step 3: Testing
python test_2D_fully.py
--root_path ../data/Synapse
--exp Synapse/mamba
--model mamba_unet
--cfg ../code/configs/vmamba_tiny.yaml
--num_classes 9
The code runs successfully and outputs metric values, but I’m not entirely sure if this is the intended way to run Synapse dataset with Mamba-UNet, especially regarding the correct config file and parameters.

Could you please confirm if these steps are correct, or suggest any modifications to ensure proper training/testing?
Any help or guidance from the authors or the community would be greatly appreciated.

Thank you in advance!