Training results on each domain

[ieeedre]

Hello, thank you for sharing Compositionally Equivariant Representation Learning paper, this is really an excellent work. I see that your paper only shows the overall training results. Do you have specific training results of vmfweak and vmfpseudo in different domains on the M&MS and SCGM data sets? For example, the table listed in your previous vmfnet.

I would like to use your excellent work as a baseline for comparison. Can you share it? Thank you so much!