gpo.zugaina.org

Search Portage & Overlays:

sci-ml/onnx

Open Neural Network Exchange (ONNX)

Screenshots

  • onnx-1.23.2-r1
    ~amd64 ~arm ~arm64
    disableStaticReg python_targets_python3_12 python_targets_python3_13 python_targets_python3_14 debug

    View      Download      Browse     License: Apache-2.0   
    Overlay: stuff
  • onnx-1.23.1-r1
    ~amd64 ~arm ~arm64
    disableStaticReg python_targets_python3_12 python_targets_python3_13 python_targets_python3_14 debug

    View      Download      Browse     License: Apache-2.0   
    Overlay: stuff
  • onnx-1.23.0-r1
    ~amd64 ~arm ~arm64
    disableStaticReg python_targets_python3_12 python_targets_python3_13 python_targets_python3_14 debug

    View      Download      Browse     License: Apache-2.0   
    Overlay: stuff
  • onnx-1.20.1
    ~amd64 ~arm64
    disableStaticReg python_targets_python3_11 python_targets_python3_12 python_targets_python3_13 python_targets_python3_14 debug

    View      Download      Browse     License: Apache-2.0   
    Overlay: gentoo
  • onnx-1.18.0-r1
    ~amd64 ~arm64 ~riscv
    disableStaticReg python_targets_python3_11 python_targets_python3_12 python_targets_python3_13 python_targets_python3_14 debug

    View      Download      Browse     License: Apache-2.0   
    Overlay: gentoo

ChangeLog

commit 4a754cba2bf2fa73e272d901bca95a44342b3e78
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Mon Oct 19 08:03:30 2026 +0200

sci-ml/onnx: keep schema registration in the 1.23.0/1.23.1 module

Same fix as 1.23.2-r1. Both retained versions passed
USE=disableStaticReg, which onnxruntime requires, to the Python
extension too, so the installed module had an empty operator
registry and onnx.checker / shape inference failed on every op.

commit dcd5760a87873efd877f9d75d8855cee727f6054
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Mon Oct 19 07:46:52 2026 +0200

sci-ml/onnx: keep schema registration in the Python module

USE=disableStaticReg, which onnxruntime requires, was also passed to
the Python extension build. That extension links its own static copy
of onnx and never registers the operator schemas itself, so the
installed module had an empty registry: onnx.checker and shape
inference failed on every op ("No Op registered for MatMul").

Build the extension with static registration on, both through
CMAKE_ARGS and the environment variable pyproject.toml reads. The
shared libonnx keeps the flag as before.

commit d65143a214d2d0e2ff78a81f3512c74067afa7e0
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Tue Oct 6 11:43:26 2026 +0200

sci-ml/onnx: drop 1.22.0

commit 9eb578c81cbfe7d26f6681b5cfc3a87346478cf0
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Tue Oct 6 11:42:39 2026 +0200

sci-ml/onnx: add 1.23.2

commit 94c9b08eb48d812d4dc7b320d3f58b1cca170037
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat Oct 3 23:19:32 2026 +0200

*/*: indent the remaining metadata.xml files with spaces

Twenty-six files imported from ::gentoo in August and September kept
its tab indentation, while the other 705 use two spaces, as Gentoo's
skel.metadata.xml does. Only leading whitespace changes; each file
parses to the same content.

commit 30c649e2e18fad67bf52fe7f98e8968a96c6d124
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat Oct 3 20:06:29 2026 +0200

sci-ml/onnx: drop the duplicated export-shared-symbols patch

The 1.23.0 patch was a copy of the 1.22.0 one; use that file by name and
remove the copy.

commit c9c63920d88638bee813f7543cf6d1137c6c5f94
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Wed Sep 30 00:47:13 2026 +0200

sci-ml/onnx: add 1.23.1

The release cherry-picks a fix to external-data loading and the numpy
helpers; the C++ sources and build files are unchanged. The shared-
symbol export patch therefore applies as is, so the new ebuild points
at the 1.23.0 copy instead of adding a duplicate.

commit 44ba3881c0d6974987225db21ed5db0f1ca2f868
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Fri Sep 25 09:40:21 2026 +0200

sci-ml/onnx: Keyword 1.23.0 for ~arm

The C extension code has no architecture-specific handling in the
ebuild.

commit 67621bc7229b098a14ce630bc0c74295f4556874
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat Sep 19 15:50:12 2026 +0200

sci-ml/onnx: add 1.23.0

ONNX 1.23 requires Protobuf 31.1 or newer. Hidden visibility also
suppresses its shared C++ API, so preserve public symbols in shared
builds.

commit 0802cefb996ddc76912c1587bd94c99f1476bbda
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sun Aug 23 23:11:51 2026 +0200

sci-ml/onnx: add cmake BDEPEND

commit 669f045c7119a8516a3601ac142d9efada77ea3d
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Fri Aug 14 23:20:20 2026 +0200

sci-ml/onnx: export shared library symbols

ONNX Runtime links schema, checker, and shape-inference APIs from the system shared
library. ONNX 1.22 hides those definitions by default, leaving downstream links with
unresolved symbols.

commit ee1c7d9ffada1f3b62ec5b8e836dcdfbbbb81590
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Fri Aug 14 23:03:45 2026 +0200

sci-ml/onnx: add 1.22.0

ONNX Runtime 1.29 pins ONNX 1.22.0, which is newer than the system repository
package. Adapt the package to the upstream scikit-build-core transition and declare
the exact release metadata floors.