Tool Solutions: ML Frameworks r25.12

PyTorch build

• The pytorch-aarch64 r25.12 README.md is available here.
• A Docker image is available on Docker Hub under armlimited/pytorch-arm-neoverse.
• Attached as artefacts are:
  • A full Software Bill of Materials (SBOM) as Tool-Solutions_r25.12_SPDX_SBOM.json, and;
  • License texts as Tool-Solutions_r25.12_licenses.zip which cover all packages included in the built image, in addition to the base Ubuntu image.

Added

• Adds oneDNN PR optimize brgemm.

Changed

• Updates hashes for:
  • PYTORCH_HASH to 93fef4bd1dd265588863929e35d9ac89328d5695, 2.10.0.dev20251124 from viable/strict, Nov 24th.
  • IDEEP_HASH to 3724bec97a77ce990e8c6dc5e595bb3beee75257, from ideep_pytorch, Nov 24th.
  • ONEDNN_HASH to 0b8a866c009b03f322e6526d7c33cfec84a4a97a, from main, Nov 25th.
  • TORCH_AO_HASH to ab6bc89512d912c17a79ed8d4d709612d3e32884, from main, Nov 25th.
  • KLEIDIAI_HASH to 94d6cc40689f44d308dbd57cb842e335fdd958f1, v1.17.0 from main, Nov 17th.

TensorFlow build

Added

Changed

• Updates TensorFlow hash to 5d46b65af45d5694cb1676bc872d24a4a64a6b57, from nightly, Nov 25th

Note

Running tensorflow-aarch64/examples/run_unit_tests.sh with ONEDNN_DEFAULT_FPMATH_MODE=BF16 can lead to some failing tests. Addressing these failures is a WIP. However, all testing is expected to pass with the default setting ONEDNN_DEFAULT_FPMATH_MODE=FP32.

What's Changed

• Bump hashes for r25.12 and patch PyTorch build by @puneetmatharu in #408
• Bump actions/upload-artifact version by @puneetmatharu in #411
• add conv optimizations by @Ryo-not-rio in #410
• Updates CHANGELOGs for 25.12 release by @puneetmatharu in #412

Full diff: r25.11...r25.12