NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI, #907 most downloaded on PyPI
image and video datasets and models for torch deep learning
Latest stable
0.29.2
Last release 22 days ago
28 Oct 2026
Ships fairly regularly
a new release about every 2 months
Nearly every release is documented
notes for 50 of 52 stable releases
11 versions withdrawn
withdrawn after publishing
10 years old
63 releases · first in 2017
Drag across the ruler to compare, or tap a release to open it.
SqueezeNet 1.0 and 1.1 models added, along with pre-trained weights
torchvision.models now expose num_classes as a constructor argumenttransforms.Pad now allows fill colors of either number tuples, or named colors like "white"make_grid and save_imageToTensor now supports more input typesToTensor now copes with all PIL Image types, not just RGB images.DS_Store is now no longer a blocking hindranceYour coding agent can read these notes before it upgrades. Set up the MCP server →
A small release, just needed a version bump because of PyPI.
A small release, just needed a version bump because of PyPI.
Add torchvision.models: Definitions and pre-trained models for common vision models
torchvision.models: Definitions and pre-trained models for common vision models
transforms.PadOne column per quarter.