Merge branch 'main' into fix/utils/windows-path-support

Author: ashwinvaidya17

docs/source/markdown/get_started/anomalib.md

@@ -52,7 +52,7 @@ interfaces, and might be easier for those who would like to use anomalib off-the
 
 ```{literalinclude} ../../../../examples/api/01_getting_started/basic_training.py
 :language: python
-:lines: 10-34
+:lines: 10-53
 ```
 
 :::
@@ -61,6 +61,7 @@ interfaces, and might be easier for those who would like to use anomalib off-the
 
 ```{literalinclude} ../../../../examples/cli/01_getting_started/basic_training.sh
 :language: bash
+:lines: 10-33
 ```
 
 :::
@@ -81,6 +82,7 @@ Anomalib includes multiple inferencing scripts, including Torch, Lightning, Grad
 
 ```{literalinclude} ../../../../examples/api/01_getting_started/basic_inference.py
 :language: python
+:lines: 10-40
 ```
 
 :::
@@ -104,8 +106,9 @@ Anomalib includes multiple inferencing scripts, including Torch, Lightning, Grad
 :::{tab-item} API
 :sync: label-1
 
-```{code-block} python
-
+```{literalinclude} ../../../../examples/api/01_getting_started/basic_torch_inference.py
+:language: python
+:lines: 10-11
 ```
 
 :::
@@ -129,8 +132,9 @@ Anomalib includes multiple inferencing scripts, including Torch, Lightning, Grad
 :::{tab-item} API
 :sync: label-1
 
-```{code-block} python
-
+```{literalinclude} ../../../../examples/api/01_getting_started/basic_openvino_inference.py
+:language: python
+:lines: 10-28
 ```
 
 :::

examples/api/01_getting_started/basic_inference.py

@@ -8,20 +8,18 @@
 """
 
 # 1. Import required modules
-from pathlib import Path
-
 from anomalib.data import PredictDataset
 from anomalib.engine import Engine
-from anomalib.models import EfficientAd
+from anomalib.models import Patchcore
 
 # 2. Initialize the model and load weights
-model = EfficientAd()
+model = Patchcore()
 engine = Engine()
 
 # 3. Prepare test data
 # You can use a single image or a folder of images
 dataset = PredictDataset(
-    path=Path("path/to/test/images"),
+    path="path/to/test/images",
     image_size=(256, 256),
 )
 

examples/api/01_getting_started/basic_openvino_inference.py

@@ -0,0 +1,28 @@
+# Copyright (C) 2025 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+
+"""Getting Started with Anomalib Inference using the Python API.
+
+This example shows how to perform inference on a trained model
+using the Anomalib Python API.
+"""
+
+# 1. Import required modules
+from anomalib.deploy import OpenVINOInferencer
+
+# 2. Initialize the inferencer
+inferencer = OpenVINOInferencer(
+    path="/path/to/openvino/model.bin",
+)
+
+# 4. Get predictions
+predictions = inferencer.predict(
+    image="/path/to/image.png",
+)
+
+# 5. Access the results
+if predictions is not None:
+    for prediction in predictions:
+        anomaly_map = prediction.anomaly_map  # Pixel-level anomaly heatmap
+        pred_label = prediction.pred_label  # Image-level label (0: normal, 1: anomalous)
+        pred_score = prediction.pred_score  # Image-level anomaly score

examples/api/01_getting_started/basic_torch_inference.py

@@ -0,0 +1,11 @@
+# Copyright (C) 2025 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+
+"""Getting Started with Anomalib Inference using the Python API.
+
+This example shows how to perform inference on a trained model
+using the Anomalib Python API.
+"""
+
+# TorchInferencer is a legacy inferencer. Consider using Engine.predict() instead,
+# which provides a more modern and feature-rich interface for model inference.

examples/api/01_getting_started/basic_training.py

@@ -9,8 +9,9 @@
 
 # 1. Import required modules
 from anomalib.data import MVTecAD
+from anomalib.deploy import ExportType
 from anomalib.engine import Engine
-from anomalib.models import EfficientAd
+from anomalib.models import Patchcore
 
 # 2. Create a dataset
 # MVTecAD is a popular dataset for anomaly detection
@@ -19,15 +20,29 @@
     category="bottle",  # MVTec category to use
     train_batch_size=32,  # Number of images per training batch
     eval_batch_size=32,  # Number of images per validation/test batch
-    num_workers=8,  # Number of parallel processes for data loading
 )
 
 # 3. Initialize the model
-# EfficientAd is a good default choice for beginners
-model = EfficientAd()
+# Patchcore is a good choice for beginners
+model = Patchcore(
+    num_neighbors=6,  # Override default model settings
+)
 
 # 4. Create the training engine
-engine = Engine(max_epochs=10)  # Train for 10 epochs
+engine = Engine(
+    max_epochs=1,  # Override default trainer settings
+)
 
 # 5. Train the model
+# This produces a lightning model (.ckpt)
 engine.fit(datamodule=datamodule, model=model)
+
+# 6. Test the model performance
+test_results = engine.test(datamodule=datamodule, model=model)
+
+# 7. Export the model
+# Different formats are available: Torch, OpenVINO, ONNX
+engine.export(
+    model=model,
+    export_type=ExportType.OPENVINO,
+)

examples/cli/01_getting_started/basic_training.sh

@@ -8,19 +8,26 @@
 # This example shows the basic steps to train an anomaly detection model.
 
 # 1. Basic Training
-# Train a model using default configuration (recommended for beginners)
+# Train and test a model using default configuration on MVTecAD bottle (default category)
 echo "Training with default configuration..."
-anomalib train --model efficient_ad
+anomalib train --model Patchcore --data anomalib.data.MVTecAD
 
 # 2. Training with Basic Customization
-# Customize basic parameters like batch size and epochs
+# Customize basic parameters like batch size and epochs.
+# For example `EfficientAd` requires a train batch size of 1
 echo -e "\nTraining with custom parameters..."
-anomalib train --model efficient_ad \
-    --data.train_batch_size 32 \
-    --trainer.max_epochs 10
+anomalib train --model EfficientAd --data anomalib.data.MVTecAD \
+    --data.category hazelnut \
+    --data.train_batch_size 1 \
+    --trainer.max_epochs 200
 
-# 3. Using a Different Dataset
-# Train on a specific category of MVTecAD dataset
-echo -e "\nTraining on MVTecAD bottle category..."
-anomalib train --model efficient_ad \
-    --data.category bottle
+# 3. Train with config file
+# Train with a custom config file
+echo -e "\nTraining with config file..."
+anomalib train --config path/to/config.yaml
+
+# 4. Export a trained model into OpenVINO
+echo .e "\nExporting model into OpenVINO..."
+anomalib export --model Patchcore \
+        --ckpt_path /path/to/model.ckpt \
+        --export_type OPENVINO

src/anomalib/metrics/__init__.py

@@ -30,13 +30,21 @@
 
 Example:
     >>> from anomalib.metrics import AUROC, F1Score
-    >>> auroc = AUROC()
-    >>> f1 = F1Score()
-    >>> labels = torch.tensor([0, 1, 0, 1])
-    >>> scores = torch.tensor([0.1, 0.9, 0.2, 0.8])
-    >>> auroc(scores, labels)
+    >>> from anomalib.data import ImageBatch
+    >>> import torch
+    >>> # Initialize metrics with required fields
+    >>> auroc = AUROC(fields=["pred_score", "gt_label"])
+    >>> f1 = F1Score(fields=["pred_score", "gt_label"], threshold=0.5)
+    >>> # Create sample batch
+    >>> batch = ImageBatch(
+    ...     image=torch.rand(4, 3, 32, 32),
+    ...     pred_score=torch.tensor([0.1, 0.9, 0.2, 0.8]),
+    ...     gt_label=torch.tensor([0, 1, 0, 1])
+    ... )
+    >>> # Calculate metrics
+    >>> auroc(batch)
     tensor(1.)
-    >>> f1(scores, labels, threshold=0.5)
+    >>> f1(batch)
     tensor(1.)
 """
 

src/anomalib/metrics/aupr.py

@@ -12,21 +12,24 @@
 
 Example:
     >>> from anomalib.metrics import AUPR
+    >>> from anomalib.data import ImageBatch
     >>> import torch
-    >>> # Create sample data
-    >>> labels = torch.tensor([0, 0, 1, 1])  # Binary labels
-    >>> scores = torch.tensor([0.1, 0.2, 0.8, 0.9])  # Anomaly scores
+    >>> # Create sample batch
+    >>> batch = ImageBatch(
+    ...     image=torch.rand(4, 3, 32, 32),
+    ...     pred_score=torch.tensor([0.1, 0.2, 0.8, 0.9]),
+    ...     gt_label=torch.tensor([0, 0, 1, 1])
+    ... )
     >>> # Initialize and compute AUPR
-    >>> metric = AUPR()
-    >>> aupr_score = metric(scores, labels)
-    >>> aupr_score
-    tensor(1.0)
+    >>> aupr = AUPR(fields=["pred_score", "gt_label"])
+    >>> aupr(batch)
+    tensor(0.8750)
 
-The metric can also be updated incrementally with batches:
+The metric can also be updated incrementally:
 
-    >>> for batch_scores, batch_labels in dataloader:
-    ...     metric.update(batch_scores, batch_labels)
-    >>> final_score = metric.compute()
+    >>> for batch in dataloader:
+    ...     aupr.update(batch)
+    >>> final_score = aupr.compute()
 
 Note:
     The AUPR score ranges from 0 to 1, with 1 indicating perfect ranking of
@@ -53,17 +56,18 @@ class _AUPR(BinaryPrecisionRecallCurve):
 
     Examples:
         To compute the metric for a set of predictions and ground truth targets:
-
+        >>> from anomalib.metrics.aupr import _AUPR
+        >>> import torch
         >>> true = torch.tensor([0, 1, 1, 1, 0, 0, 0, 0, 1, 1])
         >>> pred = torch.tensor([0.59, 0.35, 0.72, 0.33, 0.73, 0.81, 0.30, 0.05, 0.04, 0.48])
 
-        >>> metric = AUPR()
+        >>> metric = _AUPR()
         >>> metric(pred, true)
         tensor(0.4899)
 
         It is also possible to update the metric state incrementally within batches:
 
-        >>> for batch in dataloader:
+        >>> for pred, true in dataloader:
         ...     # Compute prediction and target tensors
         ...     metric.update(pred, true)
         >>> metric.compute()

src/anomalib/metrics/auroc.py

@@ -12,21 +12,24 @@
 
 Example:
     >>> from anomalib.metrics import AUROC
+    >>> from anomalib.data import ImageBatch
     >>> import torch
-    >>> # Create sample data
-    >>> labels = torch.tensor([0, 0, 1, 1])  # Binary labels
-    >>> scores = torch.tensor([0.1, 0.2, 0.8, 0.9])  # Anomaly scores
+    >>> # Create sample batch
+    >>> batch = ImageBatch(
+    ...     image=torch.rand(4, 3, 32, 32),
+    ...     pred_score=torch.tensor([0.1, 0.2, 0.8, 0.9]),
+    ...     gt_label=torch.tensor([0, 0, 1, 1])
+    ... )
     >>> # Initialize and compute AUROC
-    >>> metric = AUROC()
-    >>> auroc_score = metric(scores, labels)
-    >>> auroc_score
+    >>> auroc = AUROC(fields=["pred_score", "gt_label"])
+    >>> auroc(batch)
     tensor(1.0)
 
-The metric can also be updated incrementally with batches:
+The metric can also be updated incrementally:
 
-    >>> for batch_scores, batch_labels in dataloader:
-    ...     metric.update(batch_scores, batch_labels)
-    >>> final_score = metric.compute()
+    >>> for batch in dataloader:
+    ...     auroc.update(batch)
+    >>> final_score = auroc.compute()
 
 Once computed, the ROC curve can be visualized:
 
@@ -56,12 +59,11 @@ class _AUROC(BinaryROC):
 
     Examples:
         To compute the metric for a set of predictions and ground truth targets:
-
+        >>> from anomalib.metrics.auroc import _AUROC
         >>> import torch
-        >>> from anomalib.metrics import AUROC
         >>> preds = torch.tensor([0.13, 0.26, 0.08, 0.92, 0.03])
         >>> target = torch.tensor([0, 0, 1, 1, 0])
-        >>> auroc = AUROC()
+        >>> auroc = _AUROC()
         >>> auroc(preds, target)
         tensor(0.6667)
 
@@ -148,4 +150,23 @@ def generate_figure(self) -> tuple[Figure, str]:
 
 
 class AUROC(AnomalibMetric, _AUROC):  # type: ignore[misc]
-    """Wrapper to add AnomalibMetric functionality to AUROC metric."""
+    """Wrapper to add AnomalibMetric functionality to AUROC metric.
+
+    This class wraps the internal ``_AUROC`` metric to make it compatible with
+    Anomalib's batch processing capabilities.
+
+    Example:
+        >>> from anomalib.metrics import AUROC
+        >>> from anomalib.data import ImageBatch
+        >>> import torch
+        >>> # Create sample batch
+        >>> batch = ImageBatch(
+        ...     image=torch.rand(4, 3, 32, 32),
+        ...     pred_score=torch.tensor([0.1, 0.2, 0.8, 0.9]),
+        ...     gt_label=torch.tensor([0, 0, 1, 1])
+        ... )
+        >>> # Initialize and compute AUROC
+        >>> auroc = AUROC(fields=["pred_score", "gt_label"])
+        >>> auroc(batch)
+        tensor(1.0)
+    """