The pretrained weights do not match the model architecture

[Runinthenight]

I followed the tutorial to initialize the model using geognn_l8.json. However, when loading the provided class.pdparams weights, I encountered a warning: the pretrained weights do not match the model architecture. I would like to know if the model architecture in the latest released code has been updated compared to what was described in the paper.

/usr/local/lib/python3.8/dist-packages/paddle/fluid/dygraph/layers.py:1492: UserWarning: Skip loading for init_atom_embedding.embed_list.3.weight. init_atom_embedding.embed_list.3.weight receives a shape [9, 32], but the expected shape is [14, 32].
warnings.warn(("Skip loading for {}. ".format(key) + str(err)))
/usr/local/lib/python3.8/dist-packages/paddle/fluid/dygraph/layers.py:1492: UserWarning: Skip loading for init_atom_embedding.embed_list.6.weight. init_atom_embedding.embed_list.6.weight receives a shape [13, 32], but the expected shape is [14, 32].
warnings.warn(("Skip loading for {}. ".format(key) + str(err)))