why no 3D convolution?

[waral]

Hi, first of all, thank you so much for sharing the repo.

It's not really an issue, but rather a question, but I didn't know where to ask this.

I noticed that in your architecture there is a ResNet feature extractor, but there is no 3D convolutional network after that as in the original paper. Is there any particular reason for that? How well does the model perform compared to the results in the paper? Have you maybe compared the performance with and without it?

Thank you!