Skip to content

2.6.5

Choose a tag to compare

@ternaus ternaus released this 16 Nov 09:51
· 8 commits to main since this release
536841d

Summary

  • Exact grid symmetries, crops, and pads process vector fields directly in their input dtype. #676 · D4
  • In the single-threaded M4 Max test, a batch of 12 finite FP16 CPU gradients of shape 2×1200×1200 took 24.89 ms for D4 and collation, down from 341.47 ms. The measurement covers in-memory CPU preparation. Benchmark and workload
  • CPU Tensor fallbacks use Albucore 0.3.10 so PyTorch can detect overlapping storage when a result shares its input buffer. #674

Breaking changes

Exact field geometry preserves partially non-finite vectors component by component and keeps the original length of normals. If a model requires unit normals or all-NaN sentinels for invalid vectors, prepare the fields at load time. Interpolation excludes a vector from sampling when any component is non-finite. Vector-field contract

Bug fixes

  • PyTorch now detects overlapping writes when a CPU Tensor fallback returns a view into the original tensor. Use result.clone() as the copy source when the view overlaps the destination. #674
  • Dense 2D field warps reuse their sampled buffers and validity masks. Dense 3D displacements reuse packed pull maps across inverse iterations. #676

Misc

  • Updated GitHub Actions dependencies and the CycloneDX SBOM generator. #671 · #669

Commits

Commit PR Description
b3b816b #669 Update CycloneDX SBOM tooling to 8.0.1
93ea3d1 #671 Update GitHub Actions dependencies
8003ae1 #674 Detect hidden storage aliases in CPU Tensor fallback
46e2d27 #676 Remove redundant vector-field geometry work
536841d #677 Prepare 2.6.5