Closed ZhuYun97 closed 2 years ago
There are two main things which should be noted. (1) The class_num
should be set as 128
for the ogbg-molpcba dataset. Because this dataset has multiple tasks and can contain nan
that indicates the corresponding label is not assigned to the molecule. (2) And if you use the pre-trained model trained on other datasets(e.g. PCQM4MV1), the shape will mismatch for the encoder.embed_out
layer while loading pre-trained model. In such a situation, you need some extra operations.
Thanks for your fabulous codes. When I use such command below to run Graphormer V2 on MolPCBA dataset, some errors happen.
The error shows the mask shape is incorrect, I check the
targets
shape which is [128, 128] (should be [128, 1]) and the content oftargets
is also weird which isAre there any mistakes while I using Graphormer V2, could you help me figure them out. Thanks a lot.