Mismatch between provided configs and checkpoints

[DriessenA]

Hi,

I'm trying to load a pretrained model for which the weights are provided in: checkpoint weights. For this I wanted to use the load_model function from notebooks/utils.py, which requires the model config. However, I inferred from the load_model function that the config_hash and the model_hash should be the same. However, the provided best configs and the checkpoints don't have matching hashes. Could you point me to the config for the chemCPA sciplex rdkit model (not pretrained on bulk data), which I believe is hash 59bdaefb1c1adfaf2976e3fdf62afa21. Thank you!