You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: CHANGELOG.md
+1Lines changed: 1 addition & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -17,6 +17,7 @@ Most recent change on the top.
17
17
- optional `--constant-fold` acceleration argument for `nequip-compile --mode aotinductor` that can provide small speed ups for PyTorch >= 2.8 (may fail with some models, please open issues if that such instances are encountered)
18
18
- OpenEquivariance is now compatible with AOTInductor compilation for ASE inference with PyTorch >= 2.9
19
19
- better handling of datasets where stresses are partially populated, e.g. `AddNaNStressTransform` added to add NaN stress tensors for structures without stress data
20
+
- automatic caching for nequip.net models
20
21
21
22
### Removed
22
23
-[Breaking]`NequIPGNNEnergyModel` has been removed. Energy-only models can be constructed by using `NequIPGNNModel` with a new argument `do_derivatives=False`
0 commit comments