Breaking up a big molecule into two parts; making predictions from graphs of the two separate "molecular" parts

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

Breaking up a big molecule into two parts; making predictions from graphs of the two separate "molecular" parts #368

Closed yudhajitp closed 1 year ago

yudhajitp commented 1 year ago

Hi,

I was wondering if this is possible to do within the framework of MegNET. My molecules are bulky and hence, I assume MegNET needs a lot of epochs/blocks to capture the variance in the chemical makeup of the whole of the training molecules.

Yudhajit