materialsvirtuallab / megnet

Graph Networks as a Universal Machine Learning Framework for Molecules and Crystals
BSD 3-Clause "New" or "Revised" License
497 stars 155 forks source link

Unable to understand the shape of the target/output tensor for example in "Customized Graph Network Models" #369

Open yudhajitp opened 1 year ago

yudhajitp commented 1 year ago

Hi,

I was looking at the "Customized Graph Network Model" section of the documentation. I am trying to rework it into a custom model for my problem. I am getting stuck on the correct shape of the target/output tensor "Y". What would be your suggestion for this?

To give further context, suppose I write a line of code as follows:

model.fit( X, Y)

where, X corresponds to a graph-based input consistent with graph_to_input(). What shape should Y have? I would appreciate any help on this.

Best regards, Yudhajit