Training time for Geom Drug & Incorrect mask generation?

[Accelerator-thu]

How long will it take for training geom drug dataset on 8 GPUs? It seems that it would take about 4 hours to finish the first epoch?

[Accelerator-thu]

I just wanted to double check if the pretrained model was produced using the same hyper parameters. If so, I'll skip this training process and directly use the pretrained version you provided. Thanks in advance!

[Accelerator-thu]

Also, would you please check if the edge_mask in the geom_drug dataset is properly generated? I think in the current version of implementation, the edge_masks are generated in a reversed way, making the equivariant update block malfunctioning (the same as an Identity module now I guess.) Could you please check if there is something wrong on my side?