# Inference for Segmentations
Semantic segmentation models label each sample with a [`Segmentation`](../../api/fiftyone.core.labels.md#fiftyone.core.labels.Segmentation) field,
while instance segmentation output is stored as a [`Detections`](../../api/fiftyone.core.labels.md#fiftyone.core.labels.Detections) field whose
[`Detection`](../../api/fiftyone.core.labels.md#fiftyone.core.labels.Detection) instances carry per-object masks. Both are populated by passing
a model to
[`apply_model()`](../../api/fiftyone.core.collections.md#fiftyone.core.collections.SampleCollection.apply_model).
The [Segmentation Guide](../../getting_started/segmentation/index.md#segmentation-guide) walks through applying a
zoo model — Segment Anything 2 — to generate segmentations from prompts.
For a custom model, configure a
[`TorchImageModel`](../../api/fiftyone.utils.torch.md#fiftyone.utils.torch.TorchImageModel) with an
[`InstanceSegmenterOutputProcessor`](../../api/fiftyone.utils.torch.md#fiftyone.utils.torch.InstanceSegmenterOutputProcessor)
or
[`SemanticSegmenterOutputProcessor`](../../api/fiftyone.utils.torch.md#fiftyone.utils.torch.SemanticSegmenterOutputProcessor)
so it works with `apply_model()` directly, as described in
[Inference with custom models](index.md#running-inference-custom).