Deeplabcut

Final release before 2.2 -- stay tuned & already update!

Improvements:
- Epipolar lines and related changes 834 (from 2.1.10.3)
- Unbundled GUI code 1155 (from 2.1.10.3)
- Storing evaluation results from all shuffles/trainingsplits etc. in one summary table. So far they were separate per shuffle and split but combined iterations. 85126659534496055a626719d645a0d9cf9880fd
- Contrast augmentation conveniently included in imgaug, see: 1166 84259d70baa6f192def16ffbe5f8f45b9fbcba24

Minor changes:
- Rollback of ResNet_50 naming convention. 1195 -- this was accidentally introduced when adding the EfficientNets in January and caused videos to be reanalyzed as the `DLCsccorername` changed. This is a fix for backwards compatibility, but of course all new projects from January 21 - now suffer the same problem... One can of course write a script to re-name the output files & or simply re-run DLC. Sorry for the inconvenience.

How to upgrade:

with GUI support:
`pip install --upgrade 'deeplabcut[gui]'`
otherwise:
`pip install --upgrade deeplabcut`

The release includes some new multi-animal features, and important bug fixes. We strongly recommend using this version, which fixes several bugs 1105 1116.

Major update:
-- EfficientNet backbones integrated + Cosine Decay Learning rate schedule -- see WACV paper: https://openaccess.thecvf.com/content/WACV2021/html/Mathis_Pretraining_Boosts_Out-of-Domain_Robustness_for_Pose_Estimation_WACV_2021_paper.html

Various minor fixes, including:
-- getting video duration from metadata by def. https://github.com/DeepLabCut/DeepLabCut/commit/53b44a02262934c2c2ca898e9ff6f88a5c0cc6a0
-- fix infinite loop in testscript on some windows systems 1070

- Releasing the DeepLabCut model zoo -- http://www.mousemotorlab.org/dlc-modelzoo -- we provide model weights that are already trained on specific animals & scenarios. You can use these models for video analysis (inference) without the need to train it yourself. Contributing models and data is welcome and we are actively working on making the models more robust. Currently contains: `full_dog`, `full_cat`, `full_human` & `primate_face`. Just use `model_export` to share a model!
- Colab Notebook for using the ModelZoo: https://colab.research.google.com/github/AlexEMG/DeepLabCut/blob/master/examples/COLAB_DLC_ModelZoo.ipynb
- Updated auxiliary function to auto-update the config.yaml path, if one moves the project https://github.com/AlexEMG/DeepLabCut/pull/680
- Faster image loading/writing https://github.com/AlexEMG/DeepLabCut/pull/688
- new CropVideo function in dlc.`CropVideo`
- added 'mouse_pupil_vclose' to model zoo!


GUI updates:
- new video editor panel
- new crop video function (faster/ more robust)
- add new video button

Bugfixes:
- Hotkey for deleting labels during refinement step https://github.com/AlexEMG/DeepLabCut/issues/681
- Various bugs in wxPython 4.1.x > we recommend using 4.0.7. https://github.com/AlexEMG/DeepLabCut/issues/682
- Bodypart ordering / when changing the bodypart order in config.yaml and adding bodyparts there could be some misallignment, that should be solved now. https://github.com/AlexEMG/DeepLabCut/pull/689
- wxPython Black Box Glitch (692)

Contributors:
https://github.com/jeylau
https://github.com/jesusdpa1
https://github.com/sneakers-the-rat
https://github.com/MMathisLab
https://github.com/AlexEMG
https://github.com/ehsainit

funded in part by https://chanzuckerberg.com/eoss/ (thank you!)

- Most importantly model export functionality: https://github.com/AlexEMG/DeepLabCut/pull/649
Export DLC models. Saves the pose configuration, snapshot files, and frozen graph of the model to a directory named exported-models within the project directory.

Example: deeplabcut.export_model('/analysis/project/randomDLC-task/config.yaml')
- ability to label data that are not pngs! https://github.com/AlexEMG/DeepLabCut/commit/2e9003ce1304eccb5776a51b432ac047e012aee8
E.g. Label all jpg and jpeg in a labeled-data folder:
label_frames(config ,imtypes=['*.jpg','*.jpeg'])

- better assesment of frame extraction errors (642)
- Numpy/Moviepy conflict in dependencies (645 )
- GUI shuffle value 647
- update to Demo notebook 657
- note that https://github.com/AlexEMG/DeepLabCut/commit/46ff7f667a6dfecc1b4a300df35a5f32b1dcdd55 allows TF backwards compatibility (earlier than 1.12)
- Fix for bug in repeated refinement 670, 581 ... 671 Thanks https://github.com/hosuk88 for documenting this well!

Thanks to all the contributors!

*Surmmary for all releases since DLC2.1.6:*
- Colab Updates so they work with TF 2.0 default
- DLC 2.1.6.3; fixes pandas upg. 629, 611, 602 & Sarimax 540 and 603
- updates to docs, readme & conda installation files
- CCW rotation ability during downsampling https://github.com/AlexEMG/DeepLabCut/commit/615ad5e2d822c6f204ce8e13210785cfbf46df74
- DLC 2.16.2; cropping in GUI for frame extraction https://github.com/AlexEMG/DeepLabCut/commit/615ad5e2d822c6f204ce8e13210785cfbf46df74
- DLC 2.16.1 p-cutoff default change https://github.com/AlexEMG/DeepLabCut/commit/6b4f78b0b67d36496e14c92f76f8ac9e1aa56635