Bump torchvision from 0.15.1+metax3.7.2.0 to 0.28.0

[dependabot]

Bumps torchvision from 0.15.1+metax3.7.2.0 to 0.28.0.

Release notes

Sourced from torchvision's releases.

TorchVision 0.28.0 Release

TorchVision 0.28 is out with some small enhancement and bug-fixes:

Enhancements

• [transforms] Let wrap() preserve metadata for custom TVTensor subclasses (#9490)
• [transforms] Allow strings for interpolation param in resize transforms (#9461)

Bug fixes

• [transforms] Fix F.resize on tv_tensors.Mask to honor NEAREST_EXACT interpolation. Previously the interpolation argument was ignored for mask inputs (resize_mask hardcoded NEAREST), so NEAREST_EXACT silently produced plain NEAREST output (#9497)
• [io] Fix a GIF decoder bug on malformed GIFs that could write outside the allocated tensor's memory (#9520)

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:

Andrey Talman, Benson Ma, Jason Fried, Joanne Yun, Nicolas Hug

TorchVision 0.27.1 Release

This is a patch release, which is compatible with PyTorch 2.12.1. There are no new features added.

TorchVision 0.27 Release

TorchVision 0.27 is out! This is a small release where the main improvement is the addition of the popular lanczos interpolation mode for the v2.Resize transform on CPU. Results are equivalent to PIL's, but you can expect TorchVision to be faster as it leverages AVX2 (on x86) and NEON paths (on ARM).

Improvements

[transforms] Add support for lanczos interpolation mode (#9459) [transforms] Drastically speed-up Resize on NEON ARM (#9439) [ops] Vectorize masks_to_boxes for performance (#9358) [ops, transforms] Add direct XYWH-CXCYWH conversion for better performance (#9326) [datasets] torchvision.datasets.voc: update dataset and project site URLs (#9216) [ops] Add support for rotated boxes in box_iou (#9404, #9379) [ops][MPS] Improve runtime complexity of roi_align (#9100) [Code quality] #9359, #9364, #9359, #9317, #9409, #9408, #9416, #9411, #4463, #9475, #9427, #9448, #9443, #9396, #9316, #9286, #9324, #9338, #9381, #9386

[Documentation] #9339, #9351, #9323, #9374, #9412, #9378, #9428, #9431, #9474, #9472, #9463, #9440, #9385, #9327, #9334, #8879, #9350, #9392

Bug Fixes

[transforms] Fix incorrect normalization axis in v2.ElasticTransform (#9300) [transforms] Fix: add clamping to avoid v2.ElasticTransform IndexError when bbox equals canvas size (#9436) [transforms] Fix tv_tensors.wrap to preserve subclass types for BoundingBoxes and KeyPoints (#9332) [transforms] Fix CXCYWH to XYXY conversion for integer bounding boxes (#9322) [ops] Fix masks_to_boxes for empty masks (#9357)

Contributors

... (truncated)

Commits

• See full diff in compare view

[Dongxu-H]

torchvision 0.28.0 requires PyTorch 2.15+ but the project uses torch versions from 2.7.0 to 2.11.0 across different backends. This creates version incompatibilities:

metax backend: torch==2.8.0 incompatible with torchvision 0.28.0
kunlunxin backend: torch==2.9.0 incompatible with torchvision 0.28.0
enflame/nvidia-cuda128: torch==2.10.0 incompatible with torchvision 0.28.0
nvidia-cuda133/sunrise: torch==2.11.0 incompatible with torchvision 0.28.0
iluvatar: torch==2.7.1 incompatible with torchvision 0.28.0

Each backend should use torchvision version that matches its torch version. For reference: torch 2.8 → torchvision 0.21, torch 2.9 → torchvision 0.22, torch 2.10 → torchvision 0.25, torch 2.11 → torchvision 0.26.

This PR should be closed. Dependabot does not understand vendor-specific torch/torchvision version constraints.

[Dongxu-H]

torchvision 0.28.0 requires PyTorch 2.15+ but the project uses torch versions from 2.7.0 to 2.11.0 across different backends. This creates version incompatibilities.

metax backend: torch==2.8.0 incompatible with torchvision 0.28.0
kunlunxin backend: torch==2.9.0 incompatible with torchvision 0.28.0
enflame/nvidia-cuda128: torch==2.10.0 incompatible with torchvision 0.28.0
nvidia-cuda133/sunrise: torch==2.11.0 incompatible with torchvision 0.28.0
iluvatar/tsingmicro: torch==2.7.x incompatible with torchvision 0.28.0

Each backend should use torchvision version that matches its torch version. For reference: torch 2.8 → torchvision 0.21, torch 2.9 → torchvision 0.22, torch 2.10 → torchvision 0.25, torch 2.11 → torchvision 0.26.

This PR should be closed. Dependabot does not understand vendor-specific torch/torchvision version constraints.

[Dongxu-H]

LGTM

[Dongxu-H]

LGTM

[Dongxu-H]

CI code-style check failed. Please check the workflow logs for details. Otherwise the torchvision version bump across multiple backends looks reasonable.

[Dongxu-H]

CI code-style check failed. Updating torchvision versions across all backends to 0.28.0 will break vendor-specific SDK compatibility. Each backend requires specific torchvision versions to match their vendor torch and SDK versions (e.g., metax needs torchvision==0.15.1+metax3.7.2.0 to match torch==2.8.0+metax3.7.2.0, kunlunxin needs torchvision==0.24.0+cu129 to match torch==2.9.0+cu129). This update would break all backend-specific builds.

[Dongxu-H]

CI code-style check failed, need to check the failure details.

更新 torchvision 从厂商特定版本到通用版本 0.28.0 可能导致兼容性问题。metax 配置中 torch==2.8.0+metax3.7.2.0 搭配 torchvision==0.28.0 通用版本,iluvatar 配置中 torch==2.7.1+corex.4.4.0 搭配 torchvision==0.28.0 通用版本,这些组合可能无法正常工作。其他厂商如 kunlunxin、enflame、nvidia 等也存在类似问题。

移除厂商特定的 torchvision 构建版本可能丢失厂商优化功能,需要确认各厂商是否支持 torchvision 0.28.0 通用版本以及与当前 torch 版本的兼容性。

[Dongxu-H]

CI code-style check failed, please check the failure details.

The PR removes backend-specific torchvision versions and uses a unified 0.28.0, but this may cause version incompatibility. torchvision 0.28.0 requires PyTorch 2.13.x, however many backends configure lower torch versions (e.g., torch==2.7.1 for iluvatar, torch==2.8.0 for metax-maca3720). Using torchvision 0.28.0 with these torch versions may lead to runtime errors.

Consider keeping backend-specific torchvision versions that match their torch versions, or verify compatibility for each backend before merging.

[Dongxu-H]

torchvision 0.28.0 is for PyTorch 2.12, but many backends use older torch versions (e.g., nvidia-cuda128 uses torch==2.10.0, metax-maca3720 uses torch==2.8.0). This may cause compatibility issues. torchvision version should align with the torch version for each backend. Consider keeping backend-specific torchvision versions or verifying compatibility with each backend's torch version.

[dependabot]

OK, I won't notify you again about this release, but will get in touch when a new version is available. If you'd rather skip all updates until the next major or minor version, let me know by commenting @dependabot ignore this major version or @dependabot ignore this minor version. You can also ignore all major, minor, or patch releases for a dependency by adding an ignore condition with the desired update_types to your config file.

If you change your mind, just re-open this PR and I'll resolve any conflicts on it.