Coargus's Computer Vision Inference and Server (cvias)

The cvias module is a computer vision inference and server package for Coargus CV project.

Installation

1. Development

pip install -e .["dev","test"]

2. Production

pip install -e .

MMdetection (Third Party)

1. FasterRCNN Available model list in the class script

from cvias.common.utils_mmdet import install_dependencies
install_dependencies()

from cvias.image.detection.object.open_vocabulary.faster_rcnn import FasterRCNN

model_list = FasterRCNN.available_models()

model = FasterRCNN(
        model_name="<select model from the list>",
        explicit_checkpoint_path=None,
        gpu_number=0,
    )

model.detect(image)

Yolo (Third Party)

from cvias.image.detection.object.yolo import Yolo

model_list = Yolo.available_models()

model = Yolo(
    model_name="<select model from the list>",
    explicit_checkpoint_path=None, 
    gpu_number=0
)

model.detect(image, ["person"])

YoloWorld (Third Party)

from cvias.image.detection.object.open_vocabulary.yolo_world import YoloWorld

model_list = YoloWorld.available_models()

model = YoloWorld(
        model_name="<select model from the list>",
        explicit_checkpoint_path=None,
        gpu_number=0,
    )

model.detect(image)