Conversation
…are 33190782) and make it the default force field matpes_r2scan: alignn_atomwise_pure 2/2/128 smooth, MATPES-r2SCAN ep100, test MAE 48.7 meV/atom (E)/0.163 eV/A (F), NVE-stable, chipsff vacancy MAE 0.775 eV with r2SCAN-consistent chempots. NOTE raw r2SCAN energy scale (not PBE) - absolute energies/formation differ from matpes_pbe.
…S-safe) - use_zbl/zbl_cutoff config (default off, backward-compatible); analytic ZBL screened-Coulomb term added after the network (no retrain), smooth DimeNet cutoff so bonding/bulk are untouched. - Wired into forward_tensors (+forward_tensors_z -> TorchScript/LAMMPS) and forward(g) (ASE); autograd gives ZBL forces + virial. - Fix: pure_torch graph now carries real atomic numbers as ndata['Z'] (was absent; DGL builder's 'Z' was mistakenly the cgcnn feature matrix). - Fix: forward(g) folds ZBL into per-atom 'out' so the returned ASE energy includes it (penalty/ZBL previously reached forces only). Verified: fixes diatomic short-range collapse (92/92) in ASE + scripted (LAMMPS) model; bulk Si/Al/Fe energy shift = 0.000000 eV; default off = numbers identical.
…OpenMM exporter - forward_tensors/forward_tensors_z now build the angle (line) graph with the same three_body_cutoff (=3.5) filter the graph builder uses, so exported TorchScript/LAMMPS/OpenMM energies+forces match the ASE forward(g) path to ~1e-4 (was diverging: full line graph vs builder's cutoff-restricted one). - Apply the smooth cutoff envelope in forward_tensors edge embedding too. - Add three_body_cutoff config field (+ instance attr for TorchScript). - New alignn/scripts/export_openmm.py: scriptable positions(+box)->energy TorchForce wrapper (nm/kJ-mol units, self-built periodic neighbor list).
…lies to force-field (energy/force) models, never property/multi-output ones that share the class
… unset (Optional[bool]=None); explicit True/False still overrides. FF example config: use_zbl=True
…/lambda/wlog) + a2F Eliashberg; now 119 figshare-hosted models
…ical); move legacy DGL helpers to alignn/deprecated/pretrained.py with back-compat shim; update unified_calculator import
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.