{"full_name":"pytorch/vision","owner":"pytorch","name":"vision","shard":"1","url":"https://github.com/pytorch/vision","description":"Datasets, Transforms and Models specific to Computer Vision","homepage":"https://pytorch.org/vision","topics":["computer-vision","machine-learning"],"tag_pages":["computer-vision","machine-learning"],"language":"Python","license":"BSD-3-Clause","avatar":"https://avatars.githubusercontent.com/u/21003710?v=4","created_at":"2016-11-09T23:11:43Z","pushed_at":"2026-11-15T11:34:51Z","metrics":{"stars":18010,"forks":7295,"watchers":436,"issues":3634,"prs":6090,"releases":59,"issues_open":911,"issues_closed":2723,"prs_open":362,"prs_closed":1211,"prs_merged":4517,"discussions":0},"growth":{"day":{"stars":3,"forks":-1,"base_day":"2026-11-15"}},"history":{"dates":["2026-11-10","2026-11-11","2026-11-12","2026-11-13","2026-11-14","2026-11-15","2026-11-16"],"stars":[17995,18000,18002,18002,18006,18007,18010],"forks":[7287,7293,7294,7294,7294,7296,7295],"watchers":[null,436,436,436,436,436,436],"issues_open":[null,909,909,909,910,911,911],"prs_open":[null,359,359,359,360,362,362]},"boards":{},"same_owner":[{"full_name":"pytorch/pytorch","name":"pytorch","shard":"1","boards":{"total":132}},{"full_name":"pytorch/torchtitan","name":"torchtitan","shard":"1","boards":{}},{"full_name":"pytorch/tutorials","name":"tutorials","shard":"1","boards":{}}],"issues":[{"number":6323,"title":"[RFC] Batteries Included - Phase 3","comments":52,"created_at":"2022-07-27T14:43:03Z"},{"number":645,"title":"[Master Issue] Add more models to torchvision","comments":46,"created_at":"2018-10-30T10:27:49Z"},{"number":5088,"title":"[FEEDBACK] Multi-weight support prototype API","comments":42,"created_at":"2021-12-11T09:53:51Z"},{"number":6341,"title":"[RFC] Support YOLOX detection model","comments":34,"created_at":"2022-08-01T13:59:32Z"},{"number":4156,"title":"HTTP Error 403: rate limit exceeded when loading model","comments":32,"created_at":"2021-07-06T15:41:36Z"},{"number":3058,"title":"How to solve this error? RuntimeError: Could not run 'torchvision::nms' with arguments from the 'CUDA' backend","comments":31,"created_at":"2020-11-30T08:17:04Z"},{"number":2980,"title":"[RFC] Loss Functions in Torchvision","comments":29,"created_at":"2020-11-10T09:22:47Z"},{"number":1706,"title":"FasterRCNN to ONNX model","comments":29,"created_at":"2019-12-30T09:10:41Z"}],"discussions":[],"latest_release":{"name":"TorchVision 0.29.2","tag":"v0.29.2","published_at":"2026-10-28T16:02:36Z","body":"This is a minor release with no new feature. It's compatible with PyTorch 2.14 and later.\r\n\r\n### Improvements\r\n\r\n[packaging] Build a single `py3-none` wheel per platform, covering all supported CPython versions including free-threaded builds (#9643)\r\n\r\n### Bug Fixes\r\n\r\n[datasets] Use `filter='data'` when extracting tar archives, to prevent path traversal outside the target directory (#9694)\r\n[ops] Add more input checks to the CPU `roi_align`, `roi_pool`, `ps_roi_align` and `ps_roi_pool` kernels. Out-of-range batch indices, bad rois/grad shapes and oversized sampling grids now raise an error instead of accessing memory out of bounds (#9684)\r\n\r\nVarious code-quality improvements: #9660, #9700, #9579, #9663, #9667, #9671, #9669, #9685\r\n\r\n\r\n### Contributors\r\n\r\nThanks to Adrian Abeyta, Andrey Talman, Dmitry Nikolaev, Huy Do and Nicolas Hug.","is_prerelease":false},"readme":true,"readme_zh":false,"news":[],"stale":false,"snapshot_at":"2026-11-16T09:23:14Z","history_start":"2026-09-06"}