Skip to content

Visualize CLIP Image embedding#

Use with CLI#

Warning

This sample command is written with the assumption that it will be executed in the root directory of the amber package.

amber visualize image_embedding tests/visualization/clip_image_visualization.yaml tests/rosbag/ford_with_annotation/read_images_and_bounding_box.yaml tests/rosbag/ford_with_annotation/bounding_box.mcap
tensorboard --host 0.0.0.0 --port 6006 --logdir runs

After executing tensorboard, access http://0.0.0.0:6006#projector by your browser.

In tensorboard, you can see the embedding space of the CLIP.

Use with Python API#

current_path = Path(os.path.dirname(os.path.realpath(__file__)))
visualization = ClipEmbeddingsVisualization(
    str(current_path / "visualization" / "clip_image_visualization.yaml")
)
dataset = ImagesAndAnnotationsDataset(
    str(current_path / "rosbag" / "ford_with_annotation" / "bounding_box.mcap"),
    str(
        current_path
        / "rosbag"
        / "ford_with_annotation"
        / "read_images_and_bounding_box.yaml"
    ),
)
visualization.visualize(dataset)