remove subset-specific transforms in preprocessor

Author: djdameln

configs/model/cfa.yaml

@@ -8,9 +8,6 @@ model:
     num_hard_negative_features: 3
     radius: 1.0e-05
 
-metrics:
-  pixel: AUROC
-
 trainer:
   max_epochs: 30
   callbacks:

configs/model/cflow.yaml

@@ -15,10 +15,6 @@ model:
     permute_soft: false
     lr: 0.0001
 
-metrics:
-  pixel:
-    - AUROC
-
 trainer:
   max_epochs: 50
   callbacks:

configs/model/csflow.yaml

@@ -6,10 +6,6 @@ model:
     clamp: 3
     num_channels: 3
 
-metrics:
-  pixel:
-    - AUROC
-
 trainer:
   max_epochs: 240
   callbacks:

configs/model/draem.yaml

@@ -6,10 +6,6 @@ model:
     sspcab_lambda: 0.1
     anomaly_source_path: null
 
-metrics:
-  pixel:
-    - AUROC
-
 trainer:
   max_epochs: 700
   callbacks:

configs/model/dsr.yaml

@@ -4,10 +4,6 @@ model:
     latent_anomaly_strength: 0.2
     upsampling_train_ratio: 0.7
 
-metrics:
-  pixel:
-    - AUROC
-
 # PL Trainer Args. Don't add extra parameter here.
 trainer:
   max_epochs: 700

configs/model/efficient_ad.yaml

@@ -8,10 +8,6 @@ model:
     padding: false
     pad_maps: true
 
-metrics:
-  pixel:
-    - AUROC
-
 trainer:
   max_epochs: 1000
   max_steps: 70000

configs/model/fastflow.yaml

@@ -7,10 +7,6 @@ model:
     conv3x3_only: false
     hidden_ratio: 1.0
 
-metrics:
-  pixel:
-    - AUROC
-
 trainer:
   max_epochs: 500
   callbacks:

configs/model/padim.yaml

@@ -8,6 +8,3 @@ model:
     backbone: resnet18
     pre_trained: true
     n_features: null
-
-metrics:
-  pixel: AUROC

configs/model/reverse_distillation.yaml

@@ -9,10 +9,6 @@ model:
     anomaly_map_mode: ADD
     pre_trained: true
 
-metrics:
-  pixel:
-    - AUROC
-
 trainer:
   callbacks:
     - class_path: lightning.pytorch.callbacks.EarlyStopping

configs/model/stfpm.yaml

@@ -7,10 +7,6 @@ model:
       - layer2
       - layer3
 
-metrics:
-  pixel:
-    - AUROC
-
 trainer:
   max_epochs: 100
   callbacks:

configs/model/uflow.yaml

@@ -7,10 +7,6 @@ model:
     affine_subnet_channels_ratio: 1.0
     backbone: mcait # official: mcait, other extractors tested: resnet18, wide_resnet50_2. Could use others...
 
-metrics:
-  pixel:
-    - AUROC
-
 # PL Trainer Args. Don't add extra parameter here.
 trainer:
   max_epochs: 200

src/anomalib/data/datamodules/base/image.py

@@ -120,7 +120,7 @@ def _update_augmentations(self) -> None:
         for subset_name in ["train", "val", "test"]:
             subset = getattr(self, f"{subset_name}_data", None)
             augmentations = getattr(self, f"{subset_name}_augmentations", None)
-            model_transform = self.get_nested_attr(self, f"trainer.model.pre_processor.{subset_name}_transform")
+            model_transform = self.get_nested_attr(self, "trainer.model.pre_processor.transform")
             if subset and augmentations:
                 self._update_subset_augmentations(subset, augmentations, model_transform)
 

src/anomalib/models/components/base/anomalib_module.py

@@ -22,7 +22,6 @@
 from anomalib.data import Batch, InferenceBatch
 from anomalib.metrics import AUROC, F1Score
 from anomalib.metrics.evaluator import Evaluator
-from anomalib.metrics.threshold import Threshold
 from anomalib.post_processing import OneClassPostProcessor, PostProcessor
 from anomalib.pre_processing import PreProcessor
 from anomalib.visualization import ImageVisualizer, Visualizer
@@ -368,7 +367,7 @@ def input_size(self) -> tuple[int, int] | None:
         The effective input size is the size of the input tensor after the transform has been applied. If the transform
         is not set, or if the transform does not change the shape of the input tensor, this method will return None.
         """
-        transform = self.pre_processor.predict_transform if self.pre_processor else None
+        transform = self.pre_processor.transform if self.pre_processor else None
         if transform is None:
             return None
         dummy_input = torch.zeros(1, 3, 1, 1)
@@ -418,9 +417,6 @@ def from_config(
             help="Path to a configuration file in json or yaml format.",
         )
         model_parser.add_subclass_arguments(AnomalibModule, "model", required=False, fail_untyped=False)
-        model_parser.add_argument("--metrics.image", type=list[str] | str | None, default=["F1Score", "AUROC"])
-        model_parser.add_argument("--metrics.pixel", type=list[str] | str | None, default=None, required=False)
-        model_parser.add_argument("--metrics.threshold", type=Threshold | str, default="F1AdaptiveThreshold")
         model_parser.add_class_arguments(Trainer, "trainer", fail_untyped=False, instantiate=False, sub_configs=True)
         args = ["--config", str(config_path)]
         for key, value in kwargs.items():

src/anomalib/models/image/winclip/lightning_model.py

@@ -199,7 +199,7 @@ def configure_pre_processor(cls, image_size: tuple[int, int] | None = None) -> P
             Resize((240, 240), antialias=True, interpolation=InterpolationMode.BICUBIC),
             Normalize(mean=(0.48145466, 0.4578275, 0.40821073), std=(0.26862954, 0.26130258, 0.27577711)),
         ])
-        return PreProcessor(val_transform=transform, test_transform=transform)
+        return PreProcessor(transform=transform)
 
     @staticmethod
     def configure_post_processor() -> OneClassPostProcessor:

src/anomalib/pre_processing/pre_processing.py

@@ -77,33 +77,18 @@ class PreProcessor(nn.Module, Callback):
 
     def __init__(
         self,
-        train_transform: Transform | None = None,
-        val_transform: Transform | None = None,
-        test_transform: Transform | None = None,
         transform: Transform | None = None,
     ) -> None:
         super().__init__()
 
-        if transform and any([train_transform, val_transform, test_transform]):
-            msg = (
-                "`transforms` cannot be used together with `train_transform`, `val_transform`, `test_transform`.\n"
-                "If you want to apply the same transform to the training, validation and test data, "
-                "use only `transforms`. \n"
-                "Otherwise, specify transforms for training, validation and test individually."
-            )
-            raise ValueError(msg)
-
-        self.train_transform = train_transform or transform
-        self.val_transform = val_transform or transform
-        self.test_transform = test_transform or transform
-        self.predict_transform = self.test_transform
-        self.export_transform = get_exportable_transform(self.test_transform)
+        self.transform = transform
+        self.export_transform = get_exportable_transform(self.transform)
 
     def on_train_batch_start(self, trainer: Trainer, pl_module: LightningModule, batch: Batch, batch_idx: int) -> None:
         """Apply transforms to the batch of tensors during training."""
         del trainer, pl_module, batch_idx  # Unused
-        if self.train_transform:
-            batch.image, batch.gt_mask = self.train_transform(batch.image, batch.gt_mask)
+        if self.transform:
+            batch.image, batch.gt_mask = self.transform(batch.image, batch.gt_mask)
 
     def on_validation_batch_start(
         self,
@@ -114,8 +99,8 @@ def on_validation_batch_start(
     ) -> None:
         """Apply transforms to the batch of tensors during validation."""
         del trainer, pl_module, batch_idx  # Unused
-        if self.val_transform:
-            batch.image, batch.gt_mask = self.val_transform(batch.image, batch.gt_mask)
+        if self.transform:
+            batch.image, batch.gt_mask = self.transform(batch.image, batch.gt_mask)
 
     def on_test_batch_start(
         self,
@@ -127,8 +112,8 @@ def on_test_batch_start(
     ) -> None:
         """Apply transforms to the batch of tensors during testing."""
         del trainer, pl_module, batch_idx, dataloader_idx  # Unused
-        if self.test_transform:
-            batch.image, batch.gt_mask = self.test_transform(batch.image, batch.gt_mask)
+        if self.transform:
+            batch.image, batch.gt_mask = self.transform(batch.image, batch.gt_mask)
 
     def on_predict_batch_start(
         self,
@@ -140,8 +125,8 @@ def on_predict_batch_start(
     ) -> None:
         """Apply transforms to the batch of tensors during prediction."""
         del trainer, pl_module, batch_idx, dataloader_idx  # Unused
-        if self.predict_transform:
-            batch.image, batch.gt_mask = self.predict_transform(batch.image, batch.gt_mask)
+        if self.transform:
+            batch.image, batch.gt_mask = self.transform(batch.image, batch.gt_mask)
 
     def forward(self, batch: torch.Tensor) -> torch.Tensor:
         """Apply transforms to the batch of tensors for inference.

tests/unit/pre_processing/test_pre_processing.py

@@ -23,26 +23,6 @@ def setup(self) -> None:
         self.dummy_batch = ImageBatch(image=image, gt_mask=gt_mask)
         self.common_transform = Compose([Resize((224, 224)), ToImage(), ToDtype(torch.float32, scale=True)])
 
-    def test_init(self) -> None:
-        """Test the initialization of the PreProcessor class."""
-        # Test with stage-specific transforms
-        train_transform = Compose([Resize((224, 224)), ToImage(), ToDtype(torch.float32, scale=True)])
-        val_transform = Compose([Resize((256, 256)), ToImage(), ToDtype(torch.float32, scale=True)])
-        pre_processor = PreProcessor(train_transform=train_transform, val_transform=val_transform)
-        assert pre_processor.train_transform == train_transform
-        assert pre_processor.val_transform == val_transform
-        assert pre_processor.test_transform is None
-
-        # Test with single transform for all stages
-        pre_processor = PreProcessor(transform=self.common_transform)
-        assert pre_processor.train_transform == self.common_transform
-        assert pre_processor.val_transform == self.common_transform
-        assert pre_processor.test_transform == self.common_transform
-
-        # Test error case: both transform and stage-specific transform
-        with pytest.raises(ValueError, match="`transforms` cannot be used together with"):
-            PreProcessor(transform=self.common_transform, train_transform=train_transform)
-
     def test_forward(self) -> None:
         """Test the forward method of the PreProcessor class."""
         pre_processor = PreProcessor(transform=self.common_transform)
@@ -56,34 +36,3 @@ def test_no_transform(self) -> None:
         processed_batch = pre_processor(self.dummy_batch.image)
         assert isinstance(processed_batch, torch.Tensor)
         assert processed_batch.shape == (1, 3, 256, 256)
-
-    @staticmethod
-    def test_different_stage_transforms() -> None:
-        """Test different stage transforms."""
-        train_transform = Compose([Resize((224, 224)), ToImage(), ToDtype(torch.float32, scale=True)])
-        val_transform = Compose([Resize((256, 256)), ToImage(), ToDtype(torch.float32, scale=True)])
-        test_transform = Compose([Resize((288, 288)), ToImage(), ToDtype(torch.float32, scale=True)])
-
-        pre_processor = PreProcessor(
-            train_transform=train_transform,
-            val_transform=val_transform,
-            test_transform=test_transform,
-        )
-
-        # Test train transform
-        test_batch = ImageBatch(image=Image(torch.rand(3, 256, 256)), gt_mask=Mask(torch.zeros(256, 256)))
-        processed_batch = pre_processor.train_transform(test_batch.image)
-        assert isinstance(processed_batch, torch.Tensor)
-        assert processed_batch.shape == (1, 3, 224, 224)
-
-        # Test validation transform
-        test_batch = ImageBatch(image=Image(torch.rand(3, 256, 256)), gt_mask=Mask(torch.zeros(256, 256)))
-        processed_batch = pre_processor.val_transform(test_batch.image)
-        assert isinstance(processed_batch, torch.Tensor)
-        assert processed_batch.shape == (1, 3, 256, 256)
-
-        # Test test transform
-        test_batch = ImageBatch(image=Image(torch.rand(3, 256, 256)), gt_mask=Mask(torch.zeros(256, 256)))
-        processed_batch = pre_processor.test_transform(test_batch.image)
-        assert isinstance(processed_batch, torch.Tensor)
-        assert processed_batch.shape == (1, 3, 288, 288)