Bump the torch group in /requirements with 2 updates

[dependabot]

Bumps the torch group in /requirements with 2 updates: torch and torchvision.

Updates torch from 2.5.1 to 2.6.0

Release notes

Sourced from torch's releases.

PyTorch 2.6.0 Release

• Highlights
• Tracked Regressions
• Backwards Incompatible Change
• Deprecations
• New Features
• Improvements
• Bug fixes
• Performance
• Documentation
• Developers

Highlights

We are excited to announce the release of PyTorch® 2.6 (release notes)! This release features multiple improvements for PT2: torch.compile can now be used with Python 3.13; new performance-related knob torch.compiler.set_stance; several AOTInductor enhancements. Besides the PT2 improvements, another highlight is FP16 support on X86 CPUs.

NOTE: Starting with this release we are not going to publish on Conda, please see [Announcement] Deprecating PyTorch’s official Anaconda channel for the details.

For this release the experimental Linux binaries shipped with CUDA 12.6.3 (as well as Linux Aarch64, Linux ROCm 6.2.4, and Linux XPU binaries) are built with CXX11_ABI=1 and are using the Manylinux 2.28 build platform. If you build PyTorch extensions with custom C++ or CUDA extensions, please update these builds to use CXX_ABI=1 as well and report any issues you are seeing. For the next PyTorch 2.7 release we plan to switch all Linux builds to Manylinux 2.28 and CXX11_ABI=1, please see [RFC] PyTorch next wheel build platform: manylinux-2.28 for the details and discussion.

Also in this release as an important security improvement measure we have changed the default value for weights_only parameter of torch.load. This is a backward compatibility-breaking change, please see this forum post for more details.

This release is composed of 3892 commits from 520 contributors since PyTorch 2.5. We want to sincerely thank our dedicated community for your contributions. As always, we encourage you to try these out and report any issues as we improve PyTorch. More information about how to get started with the PyTorch 2-series can be found at our Getting Started page.

... (truncated)

Commits

• 1eba9b3 change the test wheel to release wheel when release wheel available (#145884)
• 2236df1 [CUDA] Change slim-wheel libraries load order (#145662)
• 3207040 [CD] Fix slim-wheel cuda_nvrtc import problem (#145614)
• ca3c3a6 [Release-Only] Remove ptx from Linux CUDA 12.6 binary builds (#145616)
• 7be6b5d Fix IdentationError of code example (#145525)
• dcb8ad0 update get start xpu (#145286)
• 8d4b8a9 Prevent legacy_load when weights_only=True (correctly) (#145111)
• 9c34a20 Revert "Prevent _legacy_load with weights_only=True (#144993)"
• cd15d7b Prevent _legacy_load with weights_only=True (#144993)
• a2639bc [Release/2.6] Enable python-3.13t aarch64 builds (#144878)
• Additional commits viewable in compare view

Updates torchvision from 0.20.1 to 0.21.0

Release notes

Sourced from torchvision's releases.

Torchvision 0.21 release

Highlights

Detailed changes

Image decoding

Torchvision continues to improve its image decoding capabilities. For this version, we added support for HEIC and AVIF image formats. Things are a bit different this time: to enable it, you'll need to pip install torchvision-extra-decoders, and the decoders are available in torchvision as torchvision.io.decode_heic() and torchvision.io.decode_avif(). This is still experimental / BETA, so let us know if you encounter any issue.

Read more in our docs!

New Features

[io] Add support for decoding AVIF and HEIC image formats (#8671)

Improvements

[datasets] Don't error when dataset is already downloaded (#8691) [datasets] Don't print when dataset is already downloaded (#8681) [datasets] remove printing info in datasets (#8683) [utils] Add label_colors argument to draw_bounding_boxes (#8578) [models] Add __deepcopy__ support for DualGraphModule (#8708) [Docs] Various documentation improvements (#8798, #8709, #8576, #8620, #8846, #8758) [Code quality] Various code quality improvements (#8757, #8755, #8754, #8689, #8719, #8772, #8774, #8791, #8705)

Bug Fixes

[io] Fix memory leak in decode_webp (#8712) [io] Fix pyav 14 compatibility error (#8776) [models] Fix order of auxiliary networks in googlenet.py (#8743) [transforms] Fix adjust_hue on ARM (#8618) [reference scripts] Fix error when loading the cached dataset in video classification reference(#8727) [build] fix CUDA build with NVCC_FLAGS in env (#8692)

Tracked Regressions

[build] aarch64 builds are build with manylinux_2_34_aarch64 tag according to auditwheel check (#8883)

Contributors

We're grateful for our community, which helps us improve torchvision by submitting issues and PRs, and providing feedback and suggestions. The following persons have contributed patches for this release:

amdfaa Andreas Floros, Andrey Talman , Beh Chuen Yang, David Miguel Susano Pinto, GdoongMathew, Jason Chou, Li-Huai (Allan) Lin, Maohua Li, Nicolas Hug , pblwk, R. Yao, sclarkson, vfdev, Ștefan Talpalaru

Commits

• 7af6987 [Cherry-Pick] Update Requires-Python: >=3.9 (#8886)
• eafc293 [Cherry-pick for 0.21] Migrate avif and heic decoders to torchvision-extra-de...
• 2f3a43f [Cherry-pick for 0.21] Fix doc of masks_to_boxes (#8798) (#8799)
• b411ffc [0.21 release] Use release/2.6 branch of test-infra (#8795)
• 8217401 [release/0.21] set version to 0.21.0 (#8797)
• a1c7f8e [release 0.21] remove prototype stuff (#8794)
• a9a726a Make v2 transforms authoring public (#8787)
• 48f01de Ignore mypy for maxvit (#8791)
• 6279faa Fix pyav 14 error (#8776)
• 36e219b mypy again (#8774)
• Additional commits viewable in compare view

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

---

Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

• @dependabot rebase will rebase this PR
• @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
• @dependabot merge will merge this PR after your CI passes on it
• @dependabot squash and merge will squash and merge this PR after your CI passes on it
• @dependabot cancel merge will cancel a previously requested merge and block automerging
• @dependabot reopen will reopen this PR if it is closed
• @dependabot close will close this PR and stop Dependabot recreating it. You can achieve the same result by closing it manually
• @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
• @dependabot ignore <dependency name> major version will close this group update PR and stop Dependabot creating any more for the specific dependency's major version (unless you unignore this specific dependency's major version or upgrade to it yourself)
• @dependabot ignore <dependency name> minor version will close this group update PR and stop Dependabot creating any more for the specific dependency's minor version (unless you unignore this specific dependency's minor version or upgrade to it yourself)
• @dependabot ignore <dependency name> will close this group update PR and stop Dependabot creating any more for the specific dependency (unless you unignore this specific dependency or upgrade to it yourself)
• @dependabot unignore <dependency name> will remove all of the ignore conditions of the specified dependency
• @dependabot unignore <dependency name> <ignore condition> will remove the ignore condition of the specified dependency and ignore conditions

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Files with missing lines	Coverage Δ
segmentation_models_pytorch/encoders/dpn.py	93.87% <100.00%> (ø)

[adamjstewart]

Should this have an if-statement based on PyTorch version?

[qubvel]

hmm.. this way, we should test other encoders for all the previous torch versions as well. torch.export is still in active development and it's recommended to use the latest torch, so I suppose a comment is enough unless we see an issue

[adamjstewart]

That's why I designed the tests to run on the newest and oldest version of all of our dependencies.

[qubvel]

Yes, I understand, but some features are introduced with the latest version of torch, while the main functionality works with the previous versions. I'm not sure we should claim we are only compatible with PyTorch 2.6+ if only a very small subset of features fails. WDYT?

[qubvel]

I think it should be clearly stated in the docs/README that export/scripting/compilation has been tested and is expected to work with the latest versions and one may encounter issues with previous versions.

[adamjstewart]

We can support a wider range of PyTorch versions by using an if-statement on this one line of code. Alternatively, we can skip the export test for this specific model for certain versions of PyTorch.