Use Keras_cv for object detection with GPU on Windows

Question

I am using keras_cv to create an object detection with yolov8. I've built the model with pretrained weights and a pretrained backbone directly from keras_cv. I also use methods like resizing from keras_cv. Overall with my current implementation I heavily rely on keras_cv inspired by this tutorial: https://keras.io/examples/vision/yolov8/#introduction

backbone = keras_cv.models.YOLOV8Backbone.from_preset(
    "yolo_v8_s_backbone_coco"  # We will use yolov8 small backbone with coco weights
)
yolo = keras_cv.models.YOLOV8Detector(
    num_classes=len(class_mapping),
    bounding_box_format="xyxy",
    backbone=backbone,
    fpn_depth=1,
)

resizing = keras_cv.layers.JitteredResize(
    target_size=(640, 640),
    scale_factor=(0.75, 1.3),
    bounding_box_format="xyxy",
)

Up until now I've only created the model and a data pipeline, logging and so on with the use of my CPU and now I want to start real trainings on my GPU. Since Tensorflow with GPUs works only on Windows when I use version 2.10 ("Caution: TensorFlow 2.10 was the last TensorFlow release that supported GPU on native-Windows." from Tensorflow website), I've downgraded to Tensorflow 2.10. Now I have the problem that keras_cv is not compatible with Tensorflow Versions lower than 2.11.

So my question now is: Is it not possible to use Keras_cv on Windows with GPU support. Because I can't find a way how to get GPU support for windows working with a Tensorflow higher than 2.10. but I also can't find a way how keras can run with Tensorflow version 2.10. If this is the case, is there any way how to evade the problem. Or is there a way to switch from keras_cv to another librabry with as less change neccessary as possible.