MMYOLO releases v0.2.0

v0.2.0（1/12/2022)

Highlights

1. Support YOLOv7 P5 and P6 model
2. Support YOLOv6 ML model
3. Support Grad-Based CAM and Grad-Free CAM
4. Support large image inference based on sahi
5. Add easydeploy project under the projects folder
6. Add custom dataset guide

New Features

1. browse_dataset.py script supports visualization of original image, data augmentation and intermediate results (#304)
2. Add flag to output labelme label file in image_demo.py (#288, #314)
3. Add labelme2coco script (#308, #313)
4. Add split COCO dataset script (#311)
5. Add two examples of backbone replacement in how-to.md and update plugin.md (#291)
6. Add contributing.md and code_style.md (#322)
7. Add docs about how to use mim to run scripts across libraries (#321)
8. Support YOLOv5 deployment at RV1126 device (#262)

Bug Fixes

1. Fix MixUp padding error (#319)
2. Fix scale factor order error of LetterResize and YOLOv5KeepRatioResize (#305)
3. Fix training errors of YOLOX Nano model (#285)
4. Fix RTMDet deploy error (#287)
5. Fix int8 deploy config (#315)
6. Fix make_stage_plugins doc in basebackbone (#296)
7. Enable switch to deploy when create pytorch model in deployment (#324)
8. Fix some errors in RTMDet model graph (#317)

Improvements

1. Add option of json output in test.py (#316)
2. Add area condition in extract_subcoco.py script (#286)
3. Deployment doc translation (#289)
4. Add YOLOv6 description overview doc (#252)
5. Improve config.md (#297, #303)
   6Add mosaic9 graph in docstring (#307)
6. Improve browse_coco_json.py script args (#309)
7. Refactor some functions in dataset_analysis.py to be more general (#294)

Contributors

A total of 14 developers contributed to this release.

Thank @fcakyon, @matrixgame2018, @MambaWong, @imAzhou, @triple-Mu, @RangeKing, @PeterH0323, @xin-li-67, @kitecats, @hanrui1sensetime, @AllentDan, @Zheng-LinXiao, @hhaAndroid, @wanghonglie