I am currently importing the H2O-Phase-DIagram from the AISsquare website.
In the documentation, you state that:
The accumulated number of snapshots in the training dataset is 31058, a tiny fraction (∼0.05%) of the configurations visited by DeePMD. At this point, the relative energies of configurations within each phase are well described, but deviations from DFT still affect the averages. To reduce these deviations below a small threshold, 3519 additional training configurations are added.
Yet, if I run the count.sh script that comes with the dataset, the script returns a count of:
check number of frames in count.out total number of frames: 48419 0.6861793712032503
Could you kindly elaborate where the difference in frames is coming from?
Hello together,
I am currently importing the H2O-Phase-DIagram from the AISsquare website. In the documentation, you state that:
The accumulated number of snapshots in the training dataset is 31058, a tiny fraction (∼0.05%) of the configurations visited by DeePMD. At this point, the relative energies of configurations within each phase are well described, but deviations from DFT still affect the averages. To reduce these deviations below a small threshold, 3519 additional training configurations are added.Yet, if I run the
count.shscript that comes with the dataset, the script returns a count of:check number of frames in count.out total number of frames: 48419 0.6861793712032503Could you kindly elaborate where the difference in frames is coming from?
Best regards