Closed KirillKulaev closed 7 months ago
Hi @KirillKulaev, thanks for reporting the error. I would like to clarify a bit on your regression problem. I think you are referring to the multiple values per target for each structure/graph but not another way around i.e. single value but multiple targets for each structure/graph. Regarding the multiple values per target, the framework of collate_fn needs to be modified so that it works for your purpose. I will push the fix soon. Thanks!
Hi @KirillKulaev, thanks for reporting the error. I would like to clarify a bit on your regression problem. I think you are referring to the multiple values per target for each structure/graph but not another way around i.e. single value but multiple targets for each structure/graph. Regarding the multiple values per target, the framework of collate_fn needs to be modified so that it works for your purpose. I will push the fix soon. Thanks!
Thank you very much for your answer!
Hi @KirillKulaev, I just pushed the fix for the collate_fn and now it should work. Please pull the latest version of MatGL and take the unit test of model training (tests/utils/test_training.py) for multiple values per target with M3GNet as a reference to modify your script. I would like to stress that this is an experimental feature and I am not sure how good is the accuracy of the M3GNet model for your purpose.
Hi @KirillKulaev, I just pushed the fix for the collate_fn and now it should work. Please pull the latest version of MatGL and take the unit test of model training (tests/utils/test_training.py) for multiple values per target with M3GNet as a reference to modify your script. I would like to stress that this is an experimental feature and I am not sure how good is the accuracy of the M3GNet model for your purpose.
Thank you very much, everything works well
Hi, I tried to repeat the example “Training a M3GNet Formation Energy Model with PyTorch Lightning.ipynb”, but I want to train this model to predict spectra as a vector, and when I try to train m3gnet model, I get the error, although I put the ntarget parameter.
https://colab.research.google.com/drive/1L05611HYB6UMb380xYWXp9nBZL51iHYc#scrollTo=6crRrc29Dawl