nexuml_library.scenarios.model.resnet_classifier¶
ResNet classification model scenario fragments.
resnet_classifier(
resnet_type: str = "resnet18",
pretrained: bool = False,
cifar_stem: bool | None = None,
label_key: str = "class_labels",
) -> PipelineSpec
Create a PipelineSpec for a ResNet image classifier.
Pipeline stages: - Encoder: ResNet backbone producing embeddings - Head: LatentClassificationHead producing logits - Loss: ClassificationLoss + ClassificationMetrics
Returns:
| Name | Type | Description |
|---|---|---|
PipelineSpec |
PipelineSpec
|
Pipeline with ResNet encoder, classification head and loss layers. |