gnu: Remove xnnpack-for-r-torch. * gnu/packages/machine-learning.scm (xnnpack-for-r-torch): Delete variable. Change-Id: I3e68165e8e150c5901e74cce189b60e7d85739d0
gnu: Remove qnnpack-pytorch-for-r-torch. * gnu/packages/machine-learning.scm (qnnpack-pytorch-for-r-torch): Delete variable. Change-Id: Ide167461e36c640e88e66b4ccd786e37a2e57687
gnu: Remove ideep-pytorch-for-r-torch. * gnu/packages/machine-learning.scm (ideep-pytorch-for-r-torch): Remove variable. Change-Id: Ia930ab53550688ae6ca806af4f60fb02aa139ece
gnu: Remove oneapi-dnnl-for-r-torch. * gnu/packages/machine-learning.scm (oneapi-dnnl-for-r-torch): Remove variable. Change-Id: Ie8cacc9566410422390f9bf00dc0bfbc135e502f
gnu: liblantern: Update to 0.16.3. * gnu/packages/machine-learning.scm (liblantern): Update to 0.16.3. Change-Id: Idbf7bcd135ff63eb538f08e79cdfbf1c4b5dd3fc
gnu: python-pytorch-for-r-torch: Update to 2.7.1. * gnu/packages/patches/python-pytorch-for-r-torch-fix-codegen.patch, gnu/packages/patches/python-pytorch-for-r-torch-system-libraries.patch: Update. * gnu/packages/patches/python-pytorch-for-r-torch-without-kineto.patch: New file. * gnu/local.mk (dist_patch_DATA): Record it. * gnu/packages/machine-learning.scm (python-pytorch-for-r-torch): Update to 2.7.1. [source]: Use new patch. [arguments]: Remove phase 'fix-aten-vec; copy and adjust 'use-system-libraries phase from python-pytorch. [inputs]: Inherit all from python-pytorch; replace gloo with gloo-for-r-torch. [native-inputs]: Inherit all from python-pytorch. [propagated-inputs]: Inherit all from python-pytorch. Change-Id: Ib2cf511fc34f609bbc7e92971720b00c4523419f
gnu: Add gloo-for-r-torch. * gnu/packages/machine-learning.scm (gloo-for-r-torch): New variable. Change-Id: Ibaf55d20d42acf6379464f7f7484c9ce67197d40
gnu: python-torchmetrics: Update to 1.8.2. * gnu/packages/machine-learning.scm (python-torchmetrics): Update to 1.8.2. [source]: Update hash. Change-Id: I95df5598ae30528cae7c2c10d6695decb06647e7 Signed-off-by: Sharlatan Hellseher <sharlatanus@gmail.com>
gnu: python-botorch: Update to 0.16.0. * gnu/packages/machine-learning.scm (python-botorch): Update to 0.16.0. [source]: Switch upstream URL to meta-pytorch; update hash. [arguments]: In #:test-flags, skip tests that mock fmin_l_bfgs_b_batched (not imported with SciPy<1.13): test_emsemble_map_saas, test_negative_fixed_features. Also ignore test_community/ (requires optional 'pfns'). [native-inputs]: Drop python-wheel. Change-Id: I8249ef0730bd6f6849dc515fbafa8068bf0bfe9a Signed-off-by: Sharlatan Hellseher <sharlatanus@gmail.com>
gnu: python-gpytorch: Update to 1.14.2. * gnu/packages/machine-learning.scm (python-gpytorch): Update to 1.14.2. [source]: Update hash. [native-inputs]: Drop python-wheel. Change-Id: Icd6ef2b964e17b918bbe92179e0803d4e001ec13 Signed-off-by: Sharlatan Hellseher <sharlatanus@gmail.com>
gnu: python-pot: Update to 0.9.6. * gnu/packages/machine-learning.scm (python-pot): Update to 0.9.6. [source]: Switch to git-fetch (upstream tag). Update hash. Remove snippet for deleting 'ot/lp/emd_wrap.cpp'—not present in VCS source. [arguments]: Add test-flags to skip 'test_entropic_semirelaxed_gromov' (TypeError). [native-inputs]: Drop python-wheel; add python-pytest and python-pytest-cov. Change-Id: I8932a58ad9b83fd9c82c21fbb7db23fb06336a64 Signed-off-by: Sharlatan Hellseher <sharlatanus@gmail.com>
gnu: onnxruntime: Skip exporter-dependent tests needing onnxscript. * gnu/packages/machine-learning.scm (onnxruntime): Skip exporter-dependent tests needing onnxscript. [arguments]: Extend pytest ignore list to exclude exporter tests that require python-onnxscript (test_pytorch_export_contrib_ops.py and transformers parity/phi_vision tests). Change-Id: I7fe48815fcbd53121049a3b61d43af636df2bffe Signed-off-by: Sharlatan Hellseher <sharlatanus@gmail.com>
gnu: python-torchaudio: Skip additional TorchScript consistency tests. * gnu/packages/machine-learning.scm (python-torchaudio): Extend pytest -k filter to exclude FrequencyMasking/TimeMasking/deemphasis/RNNTLoss TorchScript tests that fail with PyTorch 2.9.0. [native-inputs]: Drop python-wheel. Change-Id: I87e4360cd3ebe433f5e5e508028872b8fdc294b7 Signed-off-by: Sharlatan Hellseher <sharlatanus@gmail.com>
gnu: python-pytorch: Update to 2.9.0. * gnu/packages/machine-learning.scm (python-pytorch): Update to 2.9.0. [source]: Update hash. [arguments]: Add 'skip-pip-redirect' phase to avoid pip redirect in setup.py. Extend concurrentqueue include fix to cover nativert. [native-inputs]: Add python-setuptools. * gnu/packages/patches/python-pytorch-system-libraries.patch: Update for 2.9.0. Change-Id: I26606949da3c36fe8d7b6933ec54cb2c347a1bb9 Signed-off-by: Sharlatan Hellseher <sharlatanus@gmail.com>
gnu: gloo: Update to 0.0.0-4.54cbae0. * gnu/packages/machine-learning.scm (gloo): Update to commit 54cbae0d3a67fa890b4c3d9ee162b7860315e341. [source]: Switch upstream URL to https://github.com/pytorch/gloo and update hash. Change-Id: Ibbea0a6739dc658f9ae147641fe128650f1370b9 Signed-off-by: Sharlatan Hellseher <sharlatanus@gmail.com>
gnu: python-torchvision: Build with FFMPEG 6. * gnu/packages/machine-learning.scm (python-torchvision)[inputs]: Replace ffmpeg with ffmpeg-6.
gnu: tensorflow-lite: Update to 2.15.1. * gnu/packages/machine-learning.scm (tensorflow-lite): Update to 2.15.1. Change-Id: Ideb7605cb2cb6ddae7f53f5f6841522969ffbe83
gnu: tensorflow-lite: Adjust style. * gnu/packages/machine-learning.scm (tensorflow-lite): Tweak whitespaces. [native-inputs]: Remove labels. Replace auto-generated git repository tarballs to git-fetch. Change-Id: I5993c5accffff5b5831643ac2e221f26c899e0eb
gnu: tensorflow-lite: Fix package. Fixes some breakages since the last python-team merge. * gnu/packages/machine-learning.scm (tensorflow-lite) [arguments] <#:configure-flags>: Access OpenCL headers through mesa-headers. <#:phases>: Adapt phase 'build-python to process <#:configure-flags> properly, in particular -DCMAKE_C_FLAGS and -DCMAKE_CXX_FLAGS. [native-inputs]: Add python-setuptools and python-wrapper. Remove python-wheel. [inputs]: Replace python-wrapper by python. Remove opencl-clhpp and opencl-headers, their content is already provided through mesa-headers. Change-Id: I77c287141238081e4df0c8df48f374d8d6aa5a82 Signed-off-by: Hilton Chain <hako@ultrarare.space>
gnu: python-ml-collections: Fix build. * gnu/packages/machine-learning.scm (python-ml-collections) [propagated-inputs]: Add python-six. Change-Id: I171afef3f31d792f2016cf6a8dee0a6add3308e0