Add YOLOV8 instance segmentation model using YOLACT

[oarriaga]

What does this PR do?

Adds YOLACT fully convolutional instance segmentation model to YOLOV8.

Fixes # (issue)

#2007

Before submitting

• Did you read the contributor guideline,
  Pull Request section?
• Was this discussed/approved via a Github issue? Please add a link
  to it if that's the case.
  YOLOv8 instance segmentation #2007
• [ ] Did you write any new necessary tests?
• [ ] If this adds a new model, can you run a few training steps on TPU in Colab to ensure that no XLA incompatible OP are used?

Who can review?

@divyashreepathihalli

[divyashreepathihalli]

Thanks for the PR! left som initial comments. Can we please add a yolo_v8_backbone file into the folder, even if it is just wrapping the original yolov8 model?

Also, please add a demo colab to verify outputs and numerics. A colob for weights conversion or training is also needed.

[oarriaga]

Hi @divyashreepathihalli where would you like me to add the colab?

[divyashreepathihalli]

Hi @divyashreepathihalli where would you like me to add the colab?

Please add a Colab link here is the comments, so that I can verify the output, it doesn't have to be a part of the code added in the PR.

[divyashreepathihalli]

can you please add the model folder path here -

keras-cv/.kokoro/github/ubuntu/gpu/build.sh

Line 77 in ba2556c

keras_cv/src/models/stable_diffusion

to trigger the GPU tests. Thanks!

[oarriaga]

Hi @divyashreepathihalli, here is a colab with a training instance.

[divyashreepathihalli]

LGTM! a few small edits.

[divyashreepathihalli]

remove todo

[divyashreepathihalli]

please move this to doc-string description block