cuEquivariance is a math library that is a collective of low-level primitives and tensor ops to accelerate widely-used models, like DiffDock, MACE, Allegro and NEQUIP, based on equivariant neural networks. Also includes kernels for accelerated structure prediction.
active 2024-11-18 → 2026-08-09 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 630 days from 2024-11-18 to 2026-08-09. Pushes: 613 total, peak 19 in a day. Pull requests: 196 total, peak 8 in a day. Issues: 92 total, peak 4 in a day. Comments: 265 total, peak 14 in a day. Stars: 287 total, peak 33 in a day.
- Pushes
- Pull requests
- Issues
- Comments
- Stars
Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.
Top contributors
Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity
| Contributor | Contributions | Pushes | PRs | Comments |
|---|---|---|---|---|
| mariogeiger | 753 | 476 | 152 | 74 |
| hsadasivan | 109 | 36 | 16 | 29 |
| phiandark | 72 | 27 | 5 | 10 |
| borisfom | 65 | 10 | 14 | 22 |
| copy-pr-bot[bot] | 38 | 29 | 0 | 9 |
| v0i0 | 23 | 0 | 1 | 12 |
| guyujun | 17 | 0 | 0 | 9 |
| mitkotak | 13 | 0 | 1 | 5 |
| github-actions[bot] | 9 | 0 | 0 | 4 |
| chiang-yuan | 8 | 0 | 0 | 5 |
| rwkeane | 8 | 0 | 0 | 1 |
| sureshramani23 | 7 | 0 | 2 | 1 |
| YutackPark | 7 | 0 | 0 | 2 |
| jomitchellnv | 7 | 6 | 1 | 0 |
| LHJ1098826475 | 6 | 0 | 0 | 4 |
| vbharadwaj-bk | 6 | 0 | 0 | 5 |
| Copilot | 5 | 0 | 0 | 3 |
| stadlmax | 5 | 0 | 0 | 2 |
| moradza | 5 | 0 | 0 | 4 |
| kucukben | 4 | 0 | 0 | 4 |
Recent activity
Latest issues, pull requests and releases
- Pull request#283jomitchellnv2026-05-30 03:13
- Issue comment#283copy-pr-bot[bot]2026-05-30 00:42test: enable trace_autograd_ops for torch 2.11 compile/export
- Pull request#277changzhiai2026-05-02 01:24
- Issue comment#275achacond2026-04-22 03:07Release 0.10.0
- Issue comment#259kucukben2026-04-22 03:03JIT Compilation issues?
- Issue comment#265kucukben2026-04-22 02:58`SphericalHarmonics` naive backend does not support `torch.compile`
- Pull request#275mariogeiger2026-04-22 01:01
- Issue comment#268jchodera2026-04-10 00:06Torch inductor-compiled code does not support double-backward gradients with cuequivariance
- Issue comment#268mariogeiger2026-04-09 20:45Torch inductor-compiled code does not support double-backward gradients with cuequivariance
- Issue#268jchodera2026-04-08 19:32Torch inductor-compiled code does not support double-backward gradients with cuequivariance
- Issue comment#266copy-pr-bot[bot]2026-04-08 08:52add nvtx marker
- Issue#265rwkeane2026-04-08 01:32`SphericalHarmonics` naive backend does not support `torch.compile`
- Pull request#263phiandark2026-04-06 22:50
- Pull request#262hsadasivan2026-04-03 23:03
- Issue#253PrincessHakuryu2026-03-11 02:25Kernel launch failed when use DDP training in MACE
- Issue comment#253mariogeiger2026-03-10 16:35Kernel launch failed when use DDP training in MACE
- Issue comment#226phiandark2026-02-23 16:52cuequivariance-ops-torch-cu12 for Python 3.14
- Releasemariogeiger2026-02-17 22:56v0.9.0
- Pull request#252mariogeiger2026-02-17 13:06
- Issue comment#218sdvillal2026-02-13 00:03Conda packages for cuequivariance-ops-*
- Pull request#250hsadasivan2026-02-09 21:41
- Pull request#249hsadasivan2026-02-07 20:28
- Issue#218hsadasivan2026-02-07 00:18Conda packages for cuequivariance-ops-*
- Issue#227hsadasivan2026-02-07 00:10cuequivariance_torch.attention_pair_bias returns None when return_z_proj is False
- Issue comment#227hsadasivan2026-02-07 00:10cuequivariance_torch.attention_pair_bias returns None when return_z_proj is False
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 287 stars here means stars gained during the window, not the repo's star count.