pytorch/vision issues and pull requests

#7704 - Add masks to boundaries

Pull Request - State: open - Opened by bhack over 1 year ago - 9 comments
Labels: cla signed

#7697 - Unable to load keypointrcnn_resnet50_fpn on C++

Issue - State: open - Opened by roboticswithamir over 1 year ago - 1 comment

#7690 - ViTDet object detection + segmentation implementation

Pull Request - State: open - Opened by hgaiser over 1 year ago - 13 comments
Labels: cla signed

#7676 - Add softmax_focal_loss() to allow multi-class focal loss

Pull Request - State: closed - Opened by dhruvbird over 1 year ago - 2 comments
Labels: cla signed

#7672 - Use libjpeg-turbo

Pull Request - State: closed - Opened by NicolasHug over 1 year ago - 6 comments
Labels: bug, dependency issue, cla signed

#7660 - [Conda] pillow 9.5.0 incompatible with any torchvision version due to switch from jpeg to libjpeg-turbo

Issue - State: closed - Opened by junpuf over 1 year ago - 7 comments

#7643 - Add MPS kernels

Pull Request - State: closed - Opened by qqaatw over 1 year ago - 8 comments
Labels: module: ops, new feature, cla signed

#7637 - Torchvision dataset mirrors

Issue - State: open - Opened by dhruvbird over 1 year ago - 6 comments

#7636 - Improve function reduce dict in reference

Pull Request - State: closed - Opened by kafmws over 1 year ago - 3 comments
Labels: cla signed

#7635 - Use the same CUDNN version on Windows as PyTorch

Pull Request - State: closed - Opened by huydhn over 1 year ago - 3 comments
Labels: cla signed

#7624 - add PermuteChannels transform

Pull Request - State: closed - Opened by pmeier over 1 year ago - 1 comment
Labels: enhancement, module: transforms, cla signed

#7623 - fix lint

Pull Request - State: closed - Opened by pmeier over 1 year ago - 2 comments
Labels: cla signed, code quality

#7622 - fix system image library uninstallation on macos x86

Pull Request - State: closed - Opened by pmeier over 1 year ago - 8 comments
Labels: module: ci, cla signed

#7621 - Improve "Illustration of Transform" docs

Pull Request - State: closed - Opened by minhlong94 over 1 year ago - 3 comments
Labels: enhancement, module: documentation, cla signed

#7620 - ChannelShuffle transforms

Issue - State: closed - Opened by dhruvbird over 1 year ago - 5 comments
Labels: enhancement, module: transforms

#7619 - ConvertImageDtype does not work properly in the range (0. 1.)

Issue - State: closed - Opened by mjack3 over 1 year ago - 4 comments

#7618 - "CUDA error, AssertionError: Tensor-likes are not close!" on model test for cuda-resnet101

Issue - State: open - Opened by atalman over 1 year ago
Labels: module: tests, cuda

#7617 - test cuda118 turn off

Pull Request - State: closed - Opened by atalman over 1 year ago - 2 comments
Labels: cla signed

#7616 - Remove CUDA 11.7 builds; add 11.8

Pull Request - State: closed - Opened by ptrblck over 1 year ago - 7 comments
Labels: module: ci, cla signed

#7615 - Feature Request: enhancing "Illustration of Transforms" docs to have separate sections for geometric-based and photometric-based methods

Issue - State: closed - Opened by minhlong94 over 1 year ago - 4 comments
Labels: enhancement, module: documentation

#7614 - [torchvision] replace if statement with assert for TRT compilation

Pull Request - State: closed - Opened by NicolasHug over 1 year ago - 2 comments
Labels: fb-exported, cla signed

#7613 - remove unused kwargs from make_grid

Pull Request - State: closed - Opened by pmeier over 1 year ago - 1 comment
Labels: module: utils, cla signed, code quality

#7612 - make_grid doesn't use kwargs

Issue - State: closed - Opened by DavidFM43 over 1 year ago - 1 comment

#7611 - kill CircleCI

Pull Request - State: closed - Opened by pmeier over 1 year ago - 3 comments
Labels: module: ci, cla signed

#7610 - Do we need the Docker CI job?

Issue - State: closed - Opened by pmeier over 1 year ago - 1 comment
Labels: module: ci

#7609 - Questions about prototype builtin datasets using `torchdata`

Issue - State: closed - Opened by ain-soph over 1 year ago - 1 comment

#7608 - CXX_STANDARD 14 -> 17

Pull Request - State: closed - Opened by pmeier over 1 year ago - 2 comments
Labels: cla signed, core issue

#7607 - Division by zero -> INF values -> NaN values

Issue - State: closed - Opened by alexkoltun over 1 year ago - 3 comments

#7606 - Remove addressed workaround in ResizeV2

Pull Request - State: closed - Opened by NicolasHug over 1 year ago - 4 comments
Labels: module: transforms, cla signed, code quality

#7605 - Cannot reproduce the result of ConvNeXt pretraining

Issue - State: open - Opened by iShohei220 over 1 year ago - 5 comments

#7604 - Cmake Windows integration

Issue - State: closed - Opened by atalman over 1 year ago - 5 comments

#7603 - Perf of classification models

Issue - State: closed - Opened by NicolasHug over 1 year ago

#7602 - Fix version strings in doc

Pull Request - State: closed - Opened by pmeier over 1 year ago - 1 comment
Labels: module: documentation, cla signed

#7601 - Change default value in ConvdNormActivation Module

Pull Request - State: open - Opened by barrh over 1 year ago - 5 comments

#7600 - fix version wrangling in docs

Pull Request - State: closed - Opened by pmeier over 1 year ago - 2 comments
Labels: module: documentation, cla signed

#7599 - Mask R-CNN error when changing aspect_ratios

Issue - State: closed - Opened by PandhaB over 1 year ago - 2 comments

#7598 - Test cmake changes

Pull Request - State: closed - Opened by atalman over 1 year ago - 2 comments
Labels: cla signed

#7597 - convert torch.split return to list in RAFT

Pull Request - State: closed - Opened by pmeier over 1 year ago - 5 comments
Labels: cla signed, other

#7596 - Replaced gpuAtomicAdd by fastAtomicAdd

Pull Request - State: closed - Opened by vfdev-5 over 1 year ago - 2 comments
Labels: enhancement, module: ops, cla signed

#7595 - Max mode support for roi_align

Issue - State: closed - Opened by rit7897 over 1 year ago - 1 comment

#7594 - Unknown builtin op: torchvision::roi_align.

Issue - State: closed - Opened by taqta over 1 year ago - 10 comments

#7593 - add comments for ops LastLevelMaxPool to avoid potential confusion

Pull Request - State: closed - Opened by iaoqian over 1 year ago - 7 comments
Labels: module: models, cla signed, code quality

#7592 - Add non-TS'able _resize_image_and_masks variant with less tensor ops

Pull Request - State: closed - Opened by ezyang over 1 year ago - 4 comments
Labels: enhancement, module: models, cla signed

#7591 - Cleanup circleci binary generation

Pull Request - State: closed - Opened by atalman over 1 year ago - 2 comments
Labels: cla signed

#7590 - Include dlls required for wheel building

Pull Request - State: closed - Opened by atalman over 1 year ago - 2 comments
Labels: cla signed

#7589 - Include dlls required for wheel building

Pull Request - State: closed - Opened by atalman over 1 year ago - 2 comments
Labels: cla signed

#7588 - Fix lint

Pull Request - State: closed - Opened by NicolasHug over 1 year ago - 2 comments
Labels: cla signed, code quality, other

#7587 - Add deterministic, pure-Python roi_align implementation

Pull Request - State: closed - Opened by ezyang over 1 year ago - 2 comments
Labels: cla signed

#7586 - ADD: ApplySomeOf transforms to Random Choice

Pull Request - State: open - Opened by ambujpawar over 1 year ago - 2 comments
Labels: cla signed

#7585 - PyTorch and torchvision versions are incompatible

Issue - State: open - Opened by Donalcon over 1 year ago - 2 comments
Labels: awaiting response

#7584 - Converting RAFT models to ONNX with torch.onnx.export() results in mismatch

Issue - State: open - Opened by sramshetty over 1 year ago - 2 comments

#7583 - Add cpu smoke_test_torchvision_decode_jpeg test

Pull Request - State: closed - Opened by atalman over 1 year ago - 2 comments
Labels: cla signed

#7582 - Alert non-deterministic on kernels that use gpuAtomicAdd

Pull Request - State: closed - Opened by ezyang over 1 year ago - 3 comments
Labels: cla signed

#7581 - Use fastAtomicAdd instead of gpuAtomicAdd

Issue - State: closed - Opened by ezyang over 1 year ago - 1 comment

#7580 - reinstate doc upload

Pull Request - State: closed - Opened by pmeier over 1 year ago - 3 comments
Labels: module: documentation, module: ci, cla signed

#7579 - Respect torch.use_deterministic_algorithms

Issue - State: closed - Opened by ezyang over 1 year ago - 2 comments

#7578 - Take 2, Turn off windows circleci nightly builds. Enable GHA

Pull Request - State: closed - Opened by atalman over 1 year ago - 2 comments
Labels: cla signed

#7577 - live docs are out of date

Issue - State: closed - Opened by pmeier over 1 year ago
Labels: module: documentation, module: ci

#7576 - add 0.15.2 release to readme

Pull Request - State: closed - Opened by pmeier over 1 year ago - 2 comments
Labels: module: documentation, cla signed

#7575 - fix code format

Pull Request - State: closed - Opened by pmeier over 1 year ago - 2 comments
Labels: cla signed, code quality, other

#7574 - consolidate test workflows

Pull Request - State: closed - Opened by pmeier over 1 year ago - 1 comment
Labels: module: ci, cla signed, code quality

#7573 - setup.py: fix ROCm build

Pull Request - State: closed - Opened by justinkb over 1 year ago - 8 comments
Labels: bug, topic: binaries, module: rocm, cla signed

#7572 - Revert "Turn off windows circleci nightly builds. Enable GHA"

Pull Request - State: closed - Opened by atalman over 1 year ago - 2 comments
Labels: cla signed

#7571 - Turn off windows circleci nightly builds. Enable GHA

Pull Request - State: closed - Opened by atalman over 1 year ago - 2 comments
Labels: cla signed

#7570 - Significant Differences in the Confidence scores across IMAGENET1K_V1 and IMAGENET1K_V2 pre-trained models during Inference

Issue - State: closed - Opened by Kunaldawn7 over 1 year ago - 5 comments

GitHub / pytorch/vision issues and pull requests