Model Efficiency in the v2 paper

[YunduanLou]

Thanks for your great work! I have a question regarding model efficiency mentioned in your paper. In Table 5 of the MedSegDiff-V2 paper, it states that the model has 46M parameters, while MedSegDiff-V1 (v1) has 23M. However, when I inspect the .pt file from the sample code, I found that the model's weights contain around 129M parameters.

Could you please clarify how the parameter count in the paper was calculated? Is there a specific version of the model that corresponds to the 46M parameter count? Or are there any additional steps or modifications applied to the model that I should be aware of?

Thanks again for your great work and looking forward to your response!