Releases: supervisely-ecosystem/mmsegmentation
Releases · supervisely-ecosystem/mmsegmentation
v1.2.6
sly version upgrade
v1.2.5 sly version upgrade, change images urls (#23)
README.md updated
v1.2.4 Update README.md
Inference Video ID method added
Merge pull request #22 from supervisely-ecosystem/inf-video-id-method Inference Video ID method added
v1.2.2
v1.2.1
Support mmsegmentation 0.23.0
Merge pull request #21 from supervisely-ecosystem/dev-mmseg0.23.0 Features: - Augmentation configs updated - Supported 3 new models from mmsegmentation 0.22.0 and 0.23.0 (K-Net, BEiT, ConvNeXt) - You can change parameters from any steps if training failed and you can start new training - Many little fixes
support ROI inference defined by object bbox
support ROI inference defined by object bbox (#19) * supervisely version upgrade imports upgrade add requirements.txt add inference roi decorator * add keyword args to inference funcs * fix import * fix import * fix import * supervisely version upgrade * Update README.md
Bugfix: no augs
v1.0.2 bugfix when no augs
Filter empty masks
- empty Bitmaps filtered