open-mmlab/mmdetection · error · ImportError

please run pip install seaborn

Error message

please run pip install seaborn

What it means

Raised by _cv2_show_tracks/imshow_mot_errors when seaborn is not importable. The MOT error visualization code uses seaborn color palettes to colorize ID switches/FN/FP tracks, so the module refuses to run without it.

Source

Thrown at mmdet/utils/mot_error_visualize.py:70

            Defaults to 2.
        font_scale (float, optional): Font scale to draw id and score.
            Defaults to 0.4.
        text_width (int, optional): Width to draw id and score.
            Defaults to 10.
        text_height (int, optional): Height to draw id and score.
            Defaults to 15.
        show (bool, optional): Whether to show the image on the fly.
            Defaults to False.
        wait_time (int, optional): Value of waitKey param.
            Defaults to 100.
        out_file (str, optional): The filename to write the image.
            Defaults to None.

    Returns:
        ndarray: Visualized image.
    """
    if sns is None:
        raise ImportError('please run pip install seaborn')
    assert bboxes.ndim == 2, \
        f' bboxes ndim should be 2, but its ndim is {bboxes.ndim}.'
    assert ids.ndim == 1, \
        f' ids ndim should be 1, but its ndim is {ids.ndim}.'
    assert error_types.ndim == 1, \
        f' error_types ndim should be 1, but its ndim is {error_types.ndim}.'
    assert bboxes.shape[0] == ids.shape[0], \
        'bboxes.shape[0] and ids.shape[0] should have the same length.'
    assert bboxes.shape[1] == 5, \
        f' bboxes.shape[1] should be 5, but its {bboxes.shape[1]}.'

    bbox_colors = sns.color_palette()
    # red, yellow, blue
    bbox_colors = [bbox_colors[3], bbox_colors[1], bbox_colors[0]]
    bbox_colors = [[int(255 * _c) for _c in bbox_color][::-1]
                   for bbox_color in bbox_colors]

    if isinstance(img, str):

View on GitHub (pinned to cfd5d3a985)

Solutions

1. pip install seaborn
2. Or skip MOT error visualization and inspect result dumps instead

Example fix

// before
imshow_mot_errors(img, bboxes, ids, error_types)  # ImportError
// after
# pip install seaborn
imshow_mot_errors(img, bboxes, ids, error_types)

import importlib.util
if importlib.util.find_spec('seaborn') is None:
    raise SystemExit('seaborn required for MOT error visualization: pip install seaborn')

try:
    imshow_mot_errors(...)
except ImportError as e:
    if 'seaborn' in str(e):
        print('skip MOT error vis:', e)

When it happens

Trigger: Calling imshow_mot_errors (or _cv2_show_wrong_tracks) in mmdet/utils/mot_error_visualize.py in an environment where `import seaborn` failed (sns is None).

Common situations: Using mmdet MOT visualization tools in a minimal install that only has mmdet core dependencies; seaborn is an extra, not a hard requirement.

Related errors

• motmetrics is not installed, please install it by: pip insta
• Invalid type for palette: {type(palette)}
• palette does not exist, random is used by default. You can a
• palette does not exist, random is used by default. You can a
• dataset_meta or class names are missed, use None by default.

AI-assisted analysis of open-mmlab/mmdetection@cfd5d3a985 (2026-08-27). Data as JSON: /api/errors/93c69bb4fbea8883. Report an issue: GitHub.