Messages

1

General Questions and Answers / Re: Publication reference for the ML-DFT Density Predictor

« on: Yesterday at 17:19 »

There doesn’t appear to be a published paper specifically describing the QuantumATK Y-2026.03 DensityPredictor yet.

For now, in a paper, the safest approach is to:

Cite the standard QuantumATK software/reference paper.
Cite the Y-2026.03 DensityPredictor documentation.
State that you used the pretrained ML DensityPredictor to generate the initial electron density.

Don’t cite a generic GNN/ML-density paper as the methodology reference unless QuantumATK explicitly identifies it as the basis of their implementation.

2

General Questions and Answers / Re: Different phase colors for different transmission eigenchannels

« on: Yesterday at 17:17 »

Yes, it’s normal. Different transmission eigenchannels can have different phase/colour patterns even in the same molecular region.

BlueRed colour = phase → useful for seeing phase/interference, but not for comparing size.
Isosurface = magnitude/amplitude → this is what you want for comparing how spatially spread out the channels are.
Use the same isovalue for both channels when comparing their spatial extent.
For a quantitative comparison, use \(|\psi|^2\) or integrate the magnitude/density over particular regions/atoms.

So: same location + different colours ≠ something wrong.

3

Scripts, Tutorials and Applications / Re: Exporting COSMO solvent surface

« on: Yesterday at 17:15 »

Yes. The traceback indicates that the visualization step, rather than necessarily the COSMO calculation itself, which is what is exhausting memory. QuantumATK is trying to construct an intermediate array of roughly 740 GiB.

There may be a way to work around this by accessing the COSMO surface/tessera data directly from the QuantumATK Python API and writing it to a more suitable format, but the exact approach depends on your QuantumATK version and on which COSMO object/calculator produced the surface.