Rare find
deepmd-kit. A deep learning package for many-body potential energy representation and molecular dynamics
docs.deepmodeling.com/projects/deepmd
Vaya's read on this project
Problem, audience, market, and the verdict — sign in to see it.
Updates
June 2026
- feat(pt): add custom save behaviors (#5589)
- feat(pt/dpa4): Support property fitting in DPA4/SeZM models (#5587)
- feat(pt): add atomic charge population fitting and model (#5494)
- feat: enable PyTorch custom ops by default (#5578)
- feat(tf): support training stat_file (#5551)
- [pre-commit.ci] pre-commit autoupdate (#5576)
- feat: dp test stress (#5588)
- fix(finetune): warn on descriptor config mismatches (#5549)
- build(deps): bump actions/checkout from 6 to 7 (#5585)
- build(deps): bump actions/cache from 5 to 6 (#5586)
- feat(pt/dpa4): support multi-rank lammps inference (#5580)
- feat(dpmodel): NeighborGraph foundation — contract, segment toolkit, …
- test(api): add universal DeepPot C/C++ API tests (#5538)
- docs: add DPA3 cyclohexane distillation tutorial (#5564)
- feat(pt): add embedding sub cli for pt models (#5571)
- fix(gpu): avoid neighbor-list prefix scan deadlock (#5575)
- build: bypass Torch CUDA discovery without nvcc (#5558)
- fix(pt-expt): detach tensors before numpy serialization (#5566)
- refactor(pt/dpa4): use edge schema for dpa4 and ignore sel (#5562)
- fix(pt/train): keep epoch-derived num_steps consistent across ranks (…