{"releases":[{"id":53425492,"tag":"v0.6","author":"ppwwyyxx","name":"v0.6","draft":false,"prerelease":false,"createdAt":"2021-10-26T02:11:12Z","publishedAt":"2021-11-15T22:08:26Z","markdown":"Pre-built Linux binaries are available for the following environment:\r\n\r\n\r\n
CUDA torch 1.10torch 1.9torch 1.8
11.3
install 
python -m pip install detectron2==0.6 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu113/torch1.10/index.html\r\n
11.1
install 
python -m pip install detectron2==0.6 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu111/torch1.10/index.html\r\n
install 
python -m pip install detectron2==0.6 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu111/torch1.9/index.html\r\n
install 
python -m pip install detectron2==0.6 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu111/torch1.8/index.html\r\n
10.2
install 
python -m pip install detectron2==0.6 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.10/index.html\r\n
install 
python -m pip install detectron2==0.6 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.9/index.html\r\n
install 
python -m pip install detectron2==0.6 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.8/index.html\r\n
10.1
install 
python -m pip install detectron2==0.6 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.8/index.html\r\n
cpu
install 
python -m pip install detectron2==0.6 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.10/index.html\r\n
install 
python -m pip install detectron2==0.6 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.9/index.html\r\n
install 
python -m pip install detectron2==0.6 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.8/index.html\r\n
\r\n","html":"

Pre-built Linux binaries are available for the following environment:

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
CUDA torch 1.10torch 1.9torch 1.8
11.3\n
install 
python -m pip install detectron2==0.6 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu113/torch1.10/index.html\n
11.1\n
install 
python -m pip install detectron2==0.6 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu111/torch1.10/index.html\n
 \n
install 
python -m pip install detectron2==0.6 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu111/torch1.9/index.html\n
 \n
install 
python -m pip install detectron2==0.6 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu111/torch1.8/index.html\n
10.2\n
install 
python -m pip install detectron2==0.6 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.10/index.html\n
 \n
install 
python -m pip install detectron2==0.6 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.9/index.html\n
 \n
install 
python -m pip install detectron2==0.6 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.8/index.html\n
10.1 \n
install 
python -m pip install detectron2==0.6 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.8/index.html\n
cpu\n
install 
python -m pip install detectron2==0.6 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.10/index.html\n
 \n
install 
python -m pip install detectron2==0.6 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.9/index.html\n
 \n
install 
python -m pip install detectron2==0.6 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.8/index.html\n
\n","assets":[]},{"id":46691509,"tag":"v0.5","author":"ppwwyyxx","name":"v0.5","draft":false,"prerelease":false,"createdAt":"2021-07-23T16:28:18Z","publishedAt":"2021-07-23T18:55:45Z","markdown":"* Add [“LazyConfig” system](https://detectron2.readthedocs.io/en/latest/tutorials/lazyconfigs.html), an alternative python-based config system with more flexible syntax and no predefined structures.\r\n* Add improved Mask R-CNN baselines to model zoo using large scale jittering and longer training. See [blog](https://ai.facebook.com/blog/advancing-computer-vision-research-with-new-detectron2-mask-r-cnn-baselines) and [configs](https://github.com/facebookresearch/detectron2/tree/master/configs/new_baselines).\r\n* Add [backbone wrapper](https://detectron2.readthedocs.io/en/latest/modules/modeling.html#detectron2.modeling.MMDetBackbone) and [detector wrapper](https://detectron2.readthedocs.io/en/latest/modules/modeling.html#detectron2.modeling.MMDetDetector) for models in [mmdetection](https://github.com/open-mmlab/mmdetection/). This allows training and evaluation of these models in detectron2. See [an example config](https://github.com/facebookresearch/detectron2/blob/master/configs/Misc/mmdet_mask_rcnn_R_50_FPN_1x.py) that trains a mmdet’s Mask R-CNN\r\n* Code release for [Implicit PointRend](https://github.com/facebookresearch/detectron2/tree/master/projects/PointRend) & [PointSup](https://github.com/facebookresearch/detectron2/tree/master/projects/PointSup).\r\n* Code release for [Rethinking Batch in BatchNorm](https://github.com/facebookresearch/detectron2/tree/master/projects/Rethinking-BatchNorm).\r\n* Add support for RegNet backbones. [Example](https://github.com/facebookresearch/detectron2/blob/master/configs/COCO-InstanceSegmentation/mask_rcnn_regnety_4gf_dds_fpn_1x.py).\r\n* [Integration](https://detectron2.readthedocs.io/en/latest/modules/utils.html#detectron2.utils.analysis.FlopCountAnalysis) of fvcore’s tracing-based advanced [flop counter](https://github.com/facebookresearch/fvcore/blob/master/docs/flop_count.md).\r\n* New features in DensePose CSE: see its [release notes](https://github.com/facebookresearch/detectron2/blob/master/projects/DensePose/doc/RELEASE_2021_06.md) for more details.\r\n\r\nPre-built Linux binaries are available for the following environment:\r\n\r\n
CUDA torch 1.9torch 1.8torch 1.7
11.1
install 
python -m pip install detectron2==0.5 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu111/torch1.9/index.html\r\n
install 
python -m pip install detectron2==0.5 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu111/torch1.8/index.html\r\n
11.0
install 
python -m pip install detectron2==0.5 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu110/torch1.7/index.html\r\n
10.2
install 
python -m pip install detectron2==0.5 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.9/index.html\r\n
install 
python -m pip install detectron2==0.5 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.8/index.html\r\n
install 
python -m pip install detectron2==0.5 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.7/index.html\r\n
10.1
install 
python -m pip install detectron2==0.5 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.8/index.html\r\n
install 
python -m pip install detectron2==0.5 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.7/index.html\r\n
9.2
install 
python -m pip install detectron2==0.5 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.7/index.html\r\n
cpu
install 
python -m pip install detectron2==0.5 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.9/index.html\r\n
install 
python -m pip install detectron2==0.5 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.8/index.html\r\n
install 
python -m pip install detectron2==0.5 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.7/index.html\r\n
\r\n","html":"\n

Pre-built Linux binaries are available for the following environment:

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
CUDA torch 1.9torch 1.8torch 1.7
11.1\n
install 
python -m pip install detectron2==0.5 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu111/torch1.9/index.html\n
 \n
install 
python -m pip install detectron2==0.5 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu111/torch1.8/index.html\n
11.0 \n
install 
python -m pip install detectron2==0.5 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu110/torch1.7/index.html\n
10.2\n
install 
python -m pip install detectron2==0.5 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.9/index.html\n
 \n
install 
python -m pip install detectron2==0.5 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.8/index.html\n
 \n
install 
python -m pip install detectron2==0.5 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.7/index.html\n
10.1 \n
install 
python -m pip install detectron2==0.5 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.8/index.html\n
 \n
install 
python -m pip install detectron2==0.5 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.7/index.html\n
9.2 \n
install 
python -m pip install detectron2==0.5 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.7/index.html\n
cpu\n
install 
python -m pip install detectron2==0.5 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.9/index.html\n
 \n
install 
python -m pip install detectron2==0.5 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.8/index.html\n
 \n
install 
python -m pip install detectron2==0.5 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.7/index.html\n
\n","assets":[]},{"id":39752554,"tag":"v0.4","author":"ppwwyyxx","name":"v0.4","draft":false,"prerelease":false,"createdAt":"2021-03-13T01:21:35Z","publishedAt":"2021-03-13T01:36:49Z","markdown":"## New Features\r\n* All common models can be converted to TorchScript format by tracing or scripting ([tutorial](https://detectron2.readthedocs.io/tutorials/deployment.html)). Requires pytorch≥1.8.\r\n* Support fvcore parameter schedulers (originally from ClassyVision) that are composable, scale-invariant, and can be used on parameters other than learning rate.\r\n* Refactor PointRend as a mask head (instead of an ROIHead).\r\n* New export and C++ deployment [examples](https://github.com/facebookresearch/detectron2/tree/master/tools/deploy).\r\n* Release [d2go](https://ai.facebook.com/blog/d2go-brings-detectron2-to-mobile/) which provides end-to-end production pipeline.\r\n## New Features in DensePose:\r\nRelease __DensePose CSE__ (a framework to extend DensePose to various categories using 3D models) and __DensePose Evolution__ (a framework to bootstrap DensePose on unlabeled data). See [here](https://github.com/facebookresearch/detectron2/blob/master/projects/DensePose/doc/RELEASE_2021_03.md) for more details.\r\n\r\n## Deprecations:\r\n* Deprecate cfg argument from COCO/LVIS evaluator; Deprecate num_classes and ignore_label argument from `SemSegEvaluator`\r\n* Deprecate `WarmupMultiStepLR`, `WarmupCosineLR` in favor of fvcore schedulers\r\n* Deprecated features will be removed in future releases\r\n\r\nPre-built Linux binaries are available for the following environment:\r\n\r\n
CUDA torch 1.8torch 1.7torch 1.6
11.1
install 
python -m pip install detectron2==0.4 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu111/torch1.8/index.html\r\n
11.0
install 
python -m pip install detectron2==0.4 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu110/torch1.7/index.html\r\n
10.2
install 
python -m pip install detectron2==0.4 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.8/index.html\r\n
install 
python -m pip install detectron2==0.4 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.7/index.html\r\n
install 
python -m pip install detectron2==0.4 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.6/index.html\r\n
10.1
install 
python -m pip install detectron2==0.4 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.8/index.html\r\n
install 
python -m pip install detectron2==0.4 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.7/index.html\r\n
install 
python -m pip install detectron2==0.4 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.6/index.html\r\n
9.2
install 
python -m pip install detectron2==0.4 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.7/index.html\r\n
install 
python -m pip install detectron2==0.4 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.6/index.html\r\n
cpu
install 
python -m pip install detectron2==0.4 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.8/index.html\r\n
install 
python -m pip install detectron2==0.4 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.7/index.html\r\n
install 
python -m pip install detectron2==0.4 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.6/index.html\r\n
\r\n\r\n","html":"

New Features

\n\n

New Features in DensePose:

\n

Release DensePose CSE (a framework to extend DensePose to various categories using 3D models) and DensePose Evolution (a framework to bootstrap DensePose on unlabeled data). See here for more details.

\n

Deprecations:

\n\n

Pre-built Linux binaries are available for the following environment:

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
CUDA torch 1.8torch 1.7torch 1.6
11.1\n
install 
python -m pip install detectron2==0.4 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu111/torch1.8/index.html\n
11.0 \n
install 
python -m pip install detectron2==0.4 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu110/torch1.7/index.html\n
10.2\n
install 
python -m pip install detectron2==0.4 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.8/index.html\n
 \n
install 
python -m pip install detectron2==0.4 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.7/index.html\n
 \n
install 
python -m pip install detectron2==0.4 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.6/index.html\n
10.1\n
install 
python -m pip install detectron2==0.4 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.8/index.html\n
 \n
install 
python -m pip install detectron2==0.4 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.7/index.html\n
 \n
install 
python -m pip install detectron2==0.4 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.6/index.html\n
9.2 \n
install 
python -m pip install detectron2==0.4 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.7/index.html\n
 \n
install 
python -m pip install detectron2==0.4 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.6/index.html\n
cpu\n
install 
python -m pip install detectron2==0.4 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.8/index.html\n
 \n
install 
python -m pip install detectron2==0.4 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.7/index.html\n
 \n
install 
python -m pip install detectron2==0.4 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.6/index.html\n
\n","assets":[]},{"id":33537147,"tag":"v0.3","author":"ppwwyyxx","name":"v0.3","draft":false,"prerelease":false,"createdAt":"2020-11-06T02:23:48Z","publishedAt":"2020-11-06T03:20:56Z","markdown":"## Features & Improvements:\r\n* Support constructing RetinaNet, data loader, optimizer, COCOEvaluator without configs, in addition to Mask R-CNN.\r\n* Add [DeepLab](https://github.com/facebookresearch/detectron2/tree/master/projects/DeepLab) & [PanopticDeepLab](https://github.com/facebookresearch/detectron2/tree/master/projects/Panoptic-DeepLab) in `projects/`.\r\n* Support importing 3 projects (`point_rend`, `deeplab`, `panoptic_deeplab`) directly with `import detectron2.projects.xxx`.\r\n* Support mixed precision in training (using `cfg.SOLVER.AMP.ENABLED`) and inference.\r\n* Support ADE20k semantic segmentation dataset (named `ade20k_sem_seg_train`, `ade20k_sem_seg_val`).\r\n* Continuous build on Windows.\r\n\r\nPre-built Linux binaries are provided for the following environment:\r\n\r\n
CUDA torch 1.7torch 1.6torch 1.5
11.0
install 
python -m pip install detectron2==0.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu110/torch1.7/index.html\r\n
10.2
install 
python -m pip install detectron2==0.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.7/index.html\r\n
install 
python -m pip install detectron2==0.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.6/index.html\r\n
install 
python -m pip install detectron2==0.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.5/index.html\r\n
10.1
install 
python -m pip install detectron2==0.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.7/index.html\r\n
install 
python -m pip install detectron2==0.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.6/index.html\r\n
install 
python -m pip install detectron2==0.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.5/index.html\r\n
9.2
install 
python -m pip install detectron2==0.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.7/index.html\r\n
install 
python -m pip install detectron2==0.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.6/index.html\r\n
install 
python -m pip install detectron2==0.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.5/index.html\r\n
cpu
install 
python -m pip install detectron2==0.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.7/index.html\r\n
install 
python -m pip install detectron2==0.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.6/index.html\r\n
install 
python -m pip install detectron2==0.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.5/index.html\r\n
\r\n","html":"

Features & Improvements:

\n\n

Pre-built Linux binaries are provided for the following environment:

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
CUDA torch 1.7torch 1.6torch 1.5
11.0\n
install 
python -m pip install detectron2==0.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu110/torch1.7/index.html\n
10.2\n
install 
python -m pip install detectron2==0.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.7/index.html\n
 \n
install 
python -m pip install detectron2==0.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.6/index.html\n
 \n
install 
python -m pip install detectron2==0.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.5/index.html\n
10.1\n
install 
python -m pip install detectron2==0.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.7/index.html\n
 \n
install 
python -m pip install detectron2==0.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.6/index.html\n
 \n
install 
python -m pip install detectron2==0.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.5/index.html\n
9.2\n
install 
python -m pip install detectron2==0.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.7/index.html\n
 \n
install 
python -m pip install detectron2==0.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.6/index.html\n
 \n
install 
python -m pip install detectron2==0.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.5/index.html\n
cpu\n
install 
python -m pip install detectron2==0.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.7/index.html\n
 \n
install 
python -m pip install detectron2==0.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.6/index.html\n
 \n
install 
python -m pip install detectron2==0.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.5/index.html\n
\n","assets":[]},{"id":29338922,"tag":"v0.2.1","author":"ppwwyyxx","name":"v0.2.1","draft":false,"prerelease":false,"createdAt":"2020-07-30T07:23:23Z","publishedAt":"2020-08-04T23:11:10Z","markdown":"* Added pre-built binary for PyTorch 1.6\r\n\r\n
CUDA torch 1.6torch 1.5torch 1.4
10.2
install 
python -m pip install detectron2==0.2.1 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.6/index.html\r\n
install 
python -m pip install detectron2==0.2.1 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.5/index.html\r\n
10.1
install 
python -m pip install detectron2==0.2.1 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.6/index.html\r\n
install 
python -m pip install detectron2==0.2.1 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.5/index.html\r\n
install 
python -m pip install detectron2==0.2.1 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.4/index.html\r\n
10.0
install 
python -m pip install detectron2==0.2.1 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu100/torch1.4/index.html\r\n
9.2
install 
python -m pip install detectron2==0.2.1 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.6/index.html\r\n
install 
python -m pip install detectron2==0.2.1 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.5/index.html\r\n
install 
python -m pip install detectron2==0.2.1 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.4/index.html\r\n
cpu
install 
python -m pip install detectron2==0.2.1 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.6/index.html\r\n
install 
python -m pip install detectron2==0.2.1 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.5/index.html\r\n
install 
python -m pip install detectron2==0.2.1 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.4/index.html\r\n
","html":"\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
CUDA torch 1.6torch 1.5torch 1.4
10.2\n
install 
python -m pip install detectron2==0.2.1 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.6/index.html\n
 \n
install 
python -m pip install detectron2==0.2.1 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.5/index.html\n
10.1\n
install 
python -m pip install detectron2==0.2.1 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.6/index.html\n
 \n
install 
python -m pip install detectron2==0.2.1 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.5/index.html\n
 \n
install 
python -m pip install detectron2==0.2.1 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.4/index.html\n
10.0 \n
install 
python -m pip install detectron2==0.2.1 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu100/torch1.4/index.html\n
9.2\n
install 
python -m pip install detectron2==0.2.1 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.6/index.html\n
 \n
install 
python -m pip install detectron2==0.2.1 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.5/index.html\n
 \n
install 
python -m pip install detectron2==0.2.1 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.4/index.html\n
cpu\n
install 
python -m pip install detectron2==0.2.1 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.6/index.html\n
 \n
install 
python -m pip install detectron2==0.2.1 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.5/index.html\n
 \n
install 
python -m pip install detectron2==0.2.1 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.4/index.html\n
\n","assets":[]},{"id":28378032,"tag":"v0.2","author":"ppwwyyxx","name":"v0.2","draft":false,"prerelease":false,"createdAt":"2020-07-08T21:04:49Z","publishedAt":"2020-07-08T21:48:55Z","markdown":"### Features & Improvements:\r\n* Support constructing objects with either configs or explicit arguments. As [an example](https://detectron2.readthedocs.io/tutorials/extend.html), the entire Mask R-CNN can be built without using configs\r\n* Rename `TransformGen` to `Augmentation` and keep `TransformGen` as an alias. Design the interface of `Augmentation` so that it can access arbitrary custom data types. See [augmentation tutorial](https://detectron2.readthedocs.io/tutorials/augmentation.html) for details.\r\n* Improve speed of `COCOEvaluator` by about 3x\r\n* Support LVIS v1 dataset\r\n* Support GIoU loss in RPN and R-CNN\r\n* Support auto-scaling of batch size and learning rate in `DefaultTrainer`. See [cfg.SOLVER.REFERENCE_WORLD_SIZE](https://detectron2.readthedocs.io/modules/config.html#config-references)\r\n\r\nPre-built Linux binaries are provided for the following environment:\r\n
CUDA torch 1.5torch 1.4
10.2
install 
python -m pip install detectron2==0.2 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.5/index.html\r\n
10.1
install 
python -m pip install detectron2==0.2 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.5/index.html\r\n
install 
python -m pip install detectron2==0.2 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.4/index.html\r\n
10.0
install 
python -m pip install detectron2==0.2 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu100/torch1.4/index.html\r\n
9.2
install 
python -m pip install detectron2==0.2 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.5/index.html\r\n
install 
python -m pip install detectron2==0.2 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.4/index.html\r\n
cpu
install 
python -m pip install detectron2==0.2 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.5/index.html\r\n
install 
python -m pip install detectron2==0.2 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.4/index.html\r\n
\r\n\r\n","html":"

Features & Improvements:

\n\n

Pre-built Linux binaries are provided for the following environment:

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
CUDA torch 1.5torch 1.4
10.2\n
install 
python -m pip install detectron2==0.2 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.5/index.html\n
10.1\n
install 
python -m pip install detectron2==0.2 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.5/index.html\n
 \n
install 
python -m pip install detectron2==0.2 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.4/index.html\n
10.0 \n
install 
python -m pip install detectron2==0.2 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu100/torch1.4/index.html\n
9.2\n
install 
python -m pip install detectron2==0.2 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.5/index.html\n
 \n
install 
python -m pip install detectron2==0.2 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.4/index.html\n
cpu\n
install 
python -m pip install detectron2==0.2 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.5/index.html\n
 \n
install 
python -m pip install detectron2==0.2 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.4/index.html\n
\n","assets":[]},{"id":26594468,"tag":"v0.1.3","author":"ppwwyyxx","name":"v0.1.3","draft":false,"prerelease":false,"createdAt":"2020-05-16T08:17:41Z","publishedAt":"2020-05-17T00:47:58Z","markdown":"Bugfix version.\r\n\r\nWe started to release pre-built wheels for multiple PyTorch versions:\r\n\r\n
CUDA torch 1.5torch 1.4
10.2
install 
python -m pip install detectron2==0.1.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.5/index.html\r\n
10.1
install 
python -m pip install detectron2==0.1.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.5/index.html\r\n
install 
python -m pip install detectron2==0.1.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.4/index.html\r\n
10.0
install 
python -m pip install detectron2==0.1.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu100/torch1.4/index.html\r\n
9.2
install 
python -m pip install detectron2==0.1.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.5/index.html\r\n
install 
python -m pip install detectron2==0.1.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.4/index.html\r\n
cpu
install 
python -m pip install detectron2==0.1.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.5/index.html\r\n
install 
python -m pip install detectron2==0.1.3 -f \\\r\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.4/index.html\r\n
\r\n\r\n## Incompatible changes about internal interface:\r\n* `_init_{box,mask,keypoint}_head` of `StandardROIHeads` was changed from instance method to class method.","html":"

Bugfix version.

\n

We started to release pre-built wheels for multiple PyTorch versions:

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
CUDA torch 1.5torch 1.4
10.2\n
install 
python -m pip install detectron2==0.1.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.5/index.html\n
10.1\n
install 
python -m pip install detectron2==0.1.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.5/index.html\n
 \n
install 
python -m pip install detectron2==0.1.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.4/index.html\n
10.0 \n
install 
python -m pip install detectron2==0.1.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu100/torch1.4/index.html\n
9.2\n
install 
python -m pip install detectron2==0.1.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.5/index.html\n
 \n
install 
python -m pip install detectron2==0.1.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cu92/torch1.4/index.html\n
cpu\n
install 
python -m pip install detectron2==0.1.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.5/index.html\n
 \n
install 
python -m pip install detectron2==0.1.3 -f \\\n  https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.4/index.html\n
\n

Incompatible changes about internal interface:

\n\n","assets":[]},{"id":26153501,"tag":"v0.1.2","author":"ppwwyyxx","name":"v0.1.2","draft":false,"prerelease":false,"createdAt":"2020-05-04T17:48:25Z","publishedAt":"2020-05-04T17:52:59Z","markdown":"The pre-built wheels for this version have to be used with an official binary release of PyTorch 1.5.\r\n\r\n## Improvements:\r\n* Add semantic segmentation models to [PointRend](https://github.com/facebookresearch/detectron2/tree/master/projects/PointRend/)\r\n* Add examples to load a detectron2 model in C++\r\n* New models that reproduce [Rethinking ImageNet Pre-training](https://arxiv.org/abs/1811.08883)\r\n* Lots of new features in [DensePose](https://github.com/facebookresearch/detectron2/tree/master/projects/DensePose), see below\r\n* Fix a few bugs in rotated box computation\r\n\r\n## Incompatible changes:\r\n* When loading a checkpoint with [`resume_or_load()`](https://detectron2.readthedocs.io/modules/engine.html#detectron2.engine.defaults.DefaultTrainer.resume_or_load), training states like `optimizer`, `start_iter` will only be loaded when `resume` is True and the last checkpoint is found. This matches users’ expectations better\r\n* `.output_size` in custom box head is renamed to `.output_shape`\r\n* anchor_generator no longer duplicates the anchors for each image\r\n* `feature_strides` and `feature_channels` attributes are removed from `ROIHeads`. Use the input argument `input_shape` instead.\r\n## New in DensePose:\r\n* New evaluation metric (GPSm) that yields more reliable scores\r\n* Panoptic FPN head implementation following [Panoptic Feature Pyramid Networks](https://arxiv.org/pdf/1901.02446.pdf)\r\n* DeepLabV3 head implementation following [Rethinking Atrous Convolution for Semantic Image Segmentation](https://arxiv.org/pdf/1706.05587.pdf)\r\n* Models with confidence estimation: implementation of the paper [Correlated Uncertainty for Learning Dense Correspondences from Noisy Labels](https://papers.nips.cc/paper/8378-correlated-uncertainty-for-learning-dense-correspondences-from-noisy-labels.pdf)\r\n* Contributions to the Model Zoo (for details please see the [Model Zoo page](https://github.com/facebookresearch/detectron2/blob/master/projects/DensePose/doc/MODEL_ZOO.md)):\r\n * Panoptic FPN and tuned hyperparameters for the original fully convolutional DensePose head: the performance of the existing R50 and R101 baselines improved by +10 AP\r\n * Panoptic FPN with DeepLabV3 head for DensePose: gives additional improvement of +2 AP\r\n * Models with confidence estimation\r\n* Test time augmentations for DensePose (additional improvement of about +0.5 AP)\r\n","html":"

The pre-built wheels for this version have to be used with an official binary release of PyTorch 1.5.

\n

Improvements:

\n\n

Incompatible changes:

\n\n

New in DensePose:

\n\n","assets":[]},{"id":24263591,"tag":"v0.1.1","author":"ppwwyyxx","name":"v0.1.1","draft":false,"prerelease":false,"createdAt":"2020-02-25T19:11:31Z","publishedAt":"2020-03-05T18:32:24Z","markdown":"## Incompatible changes about head design:\r\n* Mask head and keypoint head now include logic for losses & inference. Custom heads should overwrite the feature computation by `layers()` method. \r\n* `_forward_{box,mask,keypoint}` methods of `StandardROIHeads` now accept dict of features.\r\n\r\nThis release is made to be compatible with such changes in projects ([Mesh R-CNN](https://github.com/facebookresearch/meshrcnn), [PointRend](https://github.com/facebookresearch/detectron2/tree/master/projects/PointRend), etc)\r\n\r\n## Other additional features:\r\n* [flops & parameters counting](https://detectron2.readthedocs.io/modules/utils.html#module-detectron2.utils.analysis)\r\n* Improve speed & RAM consumption of data loader\r\n\r\nThe pre-built wheels for this version have to be used with an official binary release of PyTorch 1.4.","html":"

Incompatible changes about head design:

\n\n

This release is made to be compatible with such changes in projects (Mesh R-CNN, PointRend, etc)

\n

Other additional features:

\n\n

The pre-built wheels for this version have to be used with an official binary release of PyTorch 1.4.

\n","assets":[]},{"id":23469666,"tag":"v0.1","author":"ppwwyyxx","name":"Detectron2 v0.1 Release","draft":false,"prerelease":false,"createdAt":"2020-02-07T01:22:13Z","publishedAt":"2020-02-07T05:30:13Z","markdown":"Some major additional features since open source:\r\n* Add TensorMask and PointRend in [`projects/`](https://github.com/facebookresearch/detectron2/tree/master/projects/).\r\n* Support [exporting models to caffe2 format](https://detectron2.readthedocs.io/tutorials/deployment.html#caffe2-deployment).\r\n* [`detectron2.model_zoo`](https://detectron2.readthedocs.io/modules/model_zoo.html) APIs.\r\n* Support training with segmentation annotations in binary mask format. \r\n* Support evaluating COCO-style AP for datasets in custom format.\r\n\r\nWe start to provide pre-built binary wheels at https://dl.fbaipublicfiles.com/detectron2/wheels/index.html. \r\nThe pre-built wheels for this version have to be used with an official binary release of PyTorch 1.4.","html":"

Some major additional features since open source:

\n\n

We start to provide pre-built binary wheels at https://dl.fbaipublicfiles.com/detectron2/wheels/index.html.\nThe pre-built wheels for this version have to be used with an official binary release of PyTorch 1.4.

\n","assets":[]}]}