Cannot repreduce the results of the paper in the HumanML3D

[Giat995]

Hello Dear Authors, thanks for sharing the code of MoMask! you're so great !!! but i have a question...

I trained three models RVQ VAE, T2M transformer, and Res Transformer according to the instructions you provided, but the test results differ significantly from those presented in the paper. What could be the reasons for this?

[Giat995]

Can you provide the evaluation results of the RVQ-VAE, Masked Transformer and Residual Transformer from the training stage? I can't determine which part the problem is coming from. Thank you！