Revert "Merge main to v2 (#2464)"

This reverts commit 818fa5a.

Author: samet-akcay

docs/source/markdown/guides/how_to/pipelines/index.md

@@ -6,6 +6,13 @@ This section contains tutorials on how to use different pipelines of Anomalib an
 :margin: 1 1 0 0
 :gutter: 1
 
+:::{grid-item-card} {octicon}`stack` Tiled Ensemble
+:link: ./tiled_ensemble
+:link-type: doc
+
+Learn more about how to use the tiled ensemble pipelines.
+:::
+
 :::{grid-item-card} {octicon}`gear` Custom Pipeline
 :link: ./custom_pipeline
 :link-type: doc

docs/source/markdown/guides/how_to/pipelines/tiled_ensemble.md

@@ -0,0 +1,157 @@
+# Tiled ensemble
+
+This guide will show you how to use **The Tiled Ensemble** method for anomaly detection. For more details, refer to the official [Paper](https://openaccess.thecvf.com/content/CVPR2024W/VAND/html/Rolih_Divide_and_Conquer_High-Resolution_Industrial_Anomaly_Detection_via_Memory_Efficient_CVPRW_2024_paper.html).
+
+The tiled ensemble approach reduces memory consumption by dividing input images into a grid of tiles and training a dedicated model for each tile location.
+It is compatible with any existing image anomaly detection model without the need for any modification of the underlying architecture.
+
+![Tiled ensemble flow](../../../../images/tiled_ensemble/ensemble_flow.png)
+
+```{note}
+This feature is experimental and may not work as expected.
+For any problems refer to [Issues](https://github.com/openvinotoolkit/anomalib/issues) and feel free to ask any question in [Discussions](https://github.com/openvinotoolkit/anomalib/discussions).
+```
+
+## Training
+
+You can train a tiled ensemble using the training script located inside `tools/tiled_ensemble` directory:
+
+```{code-block} bash
+
+python tools/tiled_ensemble/train_ensemble.py \
+    --config tools/tiled_ensemble/ens_config.yaml
+```
+
+By default, the Padim model is trained on **MVTec AD bottle** category using image size of 256x256, divided into non-overlapping 128x128 tiles.
+You can modify these parameters in the [config file](#ensemble-configuration).
+
+## Evaluation
+
+After training, you can evaluate the tiled ensemble on test data using:
+
+```{code-block} bash
+
+python tools/tiled_ensemble/eval.py \
+    --config tools/tiled_ensemble/ens_config.yaml \
+    --root path_to_results_dir
+
+```
+
+Ensure that `root` points to the directory containing the training results, typically `results/padim/mvtec/bottle/runX`.
+
+## Ensemble configuration
+
+Tiled ensemble is configured using `ens_config.yaml` file in the `tools/tiled_ensemble` directory.
+It contains general settings and tiled ensemble specific settings.
+
+### General
+
+General settings at the top of the config file are used to set up the random `seed`, `accelerator` (device) and the path to where results will be saved `default_root_dir`.
+
+```{code-block} yaml
+seed: 42
+accelerator: "gpu"
+default_root_dir: "results"
+```
+
+### Tiling
+
+This section contains the following settings, used for image tiling:
+
+```{code-block} yaml
+
+tiling:
+    tile_size: 256
+    stride: 256
+```
+
+These settings determine the tile size and stride. Another important parameter is image_size from `data` section later in the config. It determines the original size of the image.
+
+Input image is split into tiles, where each tile is of shape set by `tile_size` and tiles are taken with step set by `stride`.
+For example: having image_size: 512, tile_size: 256, and stride: 256, results in 4 non-overlapping tile locations.
+
+### Normalization and thresholding
+
+Next up are the normalization and thresholding settings:
+
+```{code-block} yaml
+normalization_stage: image
+thresholding:
+  method: F1AdaptiveThreshold
+  stage: image
+```
+
+- **Normalization**: Can be applied per each tile location separately (`tile` option), after combining prediction (`image` option), or skipped (`none` option).
+
+- **Thresholding**: Can also be applied at different stages, but it is limited to `tile` and `image`. Another setting for thresholding is the method used. It can be specified as a string or by the class path.
+
+### Data
+
+The `data` section is used to configure the input `image_size` and other parameters for the dataset used.
+
+```{code-block} yaml
+data:
+  class_path: anomalib.data.MVTec
+  init_args:
+    root: ./datasets/MVTec
+    category: bottle
+    train_batch_size: 32
+    eval_batch_size: 32
+    num_workers: 8
+    task: segmentation
+    transform: null
+    train_transform: null
+    eval_transform: null
+    test_split_mode: from_dir
+    test_split_ratio: 0.2
+    val_split_mode: same_as_test
+    val_split_ratio: 0.5
+    image_size: [256, 256]
+```
+
+Refer to [Data](../../reference/data/image/index.md) for more details on parameters.
+
+### SeamSmoothing
+
+This section contains settings for `SeamSmoothing` block of pipeline:
+
+```{code-block} yaml
+SeamSmoothing:
+  apply: True
+  sigma: 2
+  width: 0.1
+
+```
+
+SeamSmoothing job is responsible for smoothing of regions where tiles meet - called tile seams.
+
+- **apply**: If True, smoothing will be applied.
+- **sigma**: Controls the sigma of Gaussian filter used for smoothing.
+- **width**: Sets the percentage of the region around the seam to be smoothed.
+
+### TrainModels
+
+The last section `TrainModels` contains the setup for model training:
+
+```{code-block} yaml
+TrainModels:
+  model:
+    class_path: Fastflow
+
+  metrics:
+    pixel: AUROC
+    image: AUROC
+
+  trainer:
+    max_epochs: 500
+    callbacks:
+      - class_path: lightning.pytorch.callbacks.EarlyStopping
+        init_args:
+          patience: 42
+          monitor: pixel_AUROC
+          mode: max
+```
+
+- **Model**: Specifies the model used. Refer to [Models](../../reference/models/image/index.md) for more details on the model parameters.
+- **Metrics**: Defines evaluation metrics for pixel and image level.
+- **Trainer**: _optional_ parameters, used to control the training process. Refer to [Engine](../../reference/engine/index.md) for more details.

notebooks/700_metrics/701b_aupimo_advanced_i.ipynb

@@ -125,7 +125,7 @@ def to_onnx(
         dynamic_axes = (
             {"input": {0: "batch_size"}, "output": {0: "batch_size"}}
             if input_size
-            else {"input": {0: "batch_size", 2: "height", 3: "weight"}, "output": {0: "batch_size"}}
+            else {"input": {0: "batch_size", 2: "height", 3: "width"}, "output": {0: "batch_size"}}
         )
         onnx_path = export_root / "model.onnx"
         # apply pass through the model to get the output names

notebooks/700_metrics/701c_aupimo_advanced_ii.ipynb

@@ -0,0 +1,43 @@
+seed: 42
+accelerator: "gpu"
+default_root_dir: "results"
+
+tiling:
+  tile_size: [128, 128]
+  stride: 128
+
+normalization_stage: image # on what level we normalize, options: [tile, image, none]
+thresholding:
+  method: F1AdaptiveThreshold # refer to documentation for thresholding methods
+  stage: image # stage at which we apply threshold, options: [tile, image]
+
+data:
+  class_path: anomalib.data.MVTec
+  init_args:
+    root: ./datasets/MVTec
+    category: bottle
+    train_batch_size: 32
+    eval_batch_size: 32
+    num_workers: 8
+    task: segmentation
+    transform: null
+    train_transform: null
+    eval_transform: null
+    test_split_mode: from_dir
+    test_split_ratio: 0.2
+    val_split_mode: same_as_test
+    val_split_ratio: 0.5
+    image_size: [256, 256]
+
+SeamSmoothing:
+  apply: True # if this is applied, area around tile seams are is smoothed
+  sigma: 2 # sigma of gaussian filter used to smooth this area
+  width: 0.1 # width factor, multiplied by tile dimension gives the region width around seam which will be smoothed
+
+TrainModels:
+  model:
+    class_path: Padim
+
+  metrics:
+    pixel: AUROC
+    image: AUROC

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

@@ -0,0 +1,28 @@
+"""Run tiled ensemble prediction."""
+
+# Copyright (C) 2024 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+
+from pathlib import Path
+
+from jsonargparse import ArgumentParser
+
+from anomalib.pipelines.tiled_ensemble import EvalTiledEnsemble
+
+
+def get_parser() -> ArgumentParser:
+    """Create a new parser if none is provided."""
+    parser = ArgumentParser()
+    parser.add_argument("--config", type=str | Path, help="Configuration file path.", required=True)
+    parser.add_argument("--root", type=str | Path, help="Weights file path.", required=True)
+
+    return parser
+
+
+if __name__ == "__main__":
+    args = get_parser().parse_args()
+
+    print("Running tiled ensemble test pipeline.")
+    # pass the path to root dir with checkpoints
+    test_pipeline = EvalTiledEnsemble(args.root)
+    test_pipeline.run(args)

tools/tiled_ensemble/ens_config.yaml

@@ -0,0 +1,17 @@
+"""Run tiled ensemble training."""
+
+# Copyright (C) 2024 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+
+from anomalib.pipelines.tiled_ensemble import EvalTiledEnsemble, TrainTiledEnsemble
+
+if __name__ == "__main__":
+    print("Running tiled ensemble train pipeline")
+    train_pipeline = TrainTiledEnsemble()
+    # run training
+    train_pipeline.run()
+
+    print("Running tiled ensemble test pipeline.")
+    # pass the root dir from train run to load checkpoints
+    test_pipeline = EvalTiledEnsemble(train_pipeline.root_dir)
+    test_pipeline.run()