open-mmlab/mmdetection · info

dataset_meta or class names are missed, use None by default.

Error message

dataset_meta or class names are missed, use None by default.

What it means

Warning from init_track_model: the built tracking model has no dataset_meta attribute (checkpoints don't always carry it), so classes is set to None. MOT models don't need it, but video instance segmentation (VIS) does.

Source

Thrown at mmdet/apis/inference.py:365

                value = checkpoint_meta['dataset_meta'].pop('CLASSES')
                checkpoint_meta['dataset_meta']['classes'] = value
            model.dataset_meta = checkpoint_meta['dataset_meta']

    if detector is not None:
        assert not (checkpoint and detector), \
            'Error: checkpoint and detector checkpoint cannot both exist'
        load_checkpoint(model.detector, detector, map_location='cpu')

    if reid is not None:
        assert not (checkpoint and reid), \
            'Error: checkpoint and reid checkpoint cannot both exist'
        load_checkpoint(model.reid, reid, map_location='cpu')

    # Some methods don't load checkpoints or checkpoints don't contain
    # 'dataset_meta'
    # VIS need dataset_meta, MOT don't need dataset_meta
    if not hasattr(model, 'dataset_meta'):
        warnings.warn('dataset_meta or class names are missed, '
                      'use None by default.')
        model.dataset_meta = {'classes': None}

    model.cfg = config  # save the config in the model for convenience
    model.to(device)
    model.eval()
    return model

View on GitHub (pinned to cfd5d3a985)

Solutions

1. For MOT: safe to ignore
2. For VIS: load a checkpoint saved with dataset_meta, or set model.dataset_meta = {'classes': [...]} after init
3. Update the config/dataset metainfo so the checkpoint retains classes

Example fix

// before
model = init_track_model(mot_cfg, ckpt)
// after (VIS case)
model = init_track_model(vis_cfg, ckpt)
model.dataset_meta = {'classes': vis_classes}

model = init_track_model(cfg, ckpt)
if not hasattr(model, 'dataset_meta') or model.dataset_meta.get('classes') is None:
    if needs_classes:  # VIS-style usage
        model.dataset_meta = {'classes': my_classes}

When it happens

Trigger: Calling init_track_model(config, checkpoint, reid) where the detector or the loaded checkpoints lack 'dataset_meta' — common for MOT (DeepSORT etc.) configs.

Common situations: Running MOT demos; the warning is benign for pure tracking. If a VIS model hits it, downstream visualization needing class names will misbehave.

Related errors

• trackeval is not installed,please install it by: pip install
• metric must be a list or a str.
• metric {metric} is not supported.
• please run pip install seaborn
• dataset_meta or class names are not saved in the checkpoint'

AI-assisted analysis of open-mmlab/mmdetection@cfd5d3a985 (2026-08-27). Data as JSON: /api/errors/f2cbb701a0a9b43b. Report an issue: GitHub.