benchmarks.hota.models.segments_to_prediction_with_features
def segments_to_prediction_with_features(segmentation: torch.Tensor, segments: list[dict], emb_keep: torch.Tensor, thing_ids: Iterable[int]) -> FramePredictionBuild a thing-only ``FramePrediction`` with embeddings and centroids.
Uses the same thing-filter logic as :func:`segments_to_prediction`.
Source: unitrack/benchmarks/hota/models.py:117