Repository files navigation

A library for benchmarking, developing and deploying deep learning anomaly detection algorithms

---

Key Features •Docs •Notebooks •License

🌟Announcing v2.0.0 Release! 🌟

We're excited to announce the release of Anomalib v2.0.0! This version introduces significant improvements and customization options to enhance your anomaly detection workflows. Please be aware that there are several API changes betweenv1.2.0 andv2.0.0, so please be careful when updating your existing pipelines. Key features include:

• Multi-GPU support
• Newdataclasses for model in- and outputs.
• Flexible configuration ofmodel transforms and data augmentations.
• Configurable modules for pre- and post-processing operations viaPreprocessor andPostprocessor
• Customizable model evaluation workflow with newMetrics API andEvaluator module.
• Configurable module for visualization viaVisualizer (docs guide: coming soon)

We value your input! Please share feedback viaGitHub Issues or ourDiscussions

Anomalib is a deep learning library that aims to collect state-of-the-art anomaly detection algorithms for benchmarking on both public and private datasets. Anomalib provides several ready-to-use implementations of anomaly detection algorithms described in the recent literature, as well as a set of tools that facilitate the development and implementation of custom models. The library has a strong focus on visual anomaly detection, where the goal of the algorithm is to detect and/or localize anomalies within images or videos in a dataset. Anomalib is constantly updated with new algorithms and training/inference extensions, so keep checking!

• Simple and modular API and CLI for training, inference, benchmarking, and hyperparameter optimization.
• The largest public collection of ready-to-use deep learning anomaly detection algorithms and benchmark datasets.
• Lightning based model implementations to reduce boilerplate code and limit the implementation efforts to the bare essentials.
• The majority of models can be exported toOpenVINO Intermediate Representation (IR) for accelerated inference on Intel hardware.
• A set ofinference tools for quick and easy deployment of the standard or custom anomaly detection models.

Anomalib can be installed from PyPI. We recommend using a virtual environment and a modern package installer likeuv orpip.

For a standard installation, you can useuv orpip. This will install the latest version of Anomalib with its core dependencies. PyTorch will be installed based on its default behavior, which usually works for CPU and standard CUDA setups.

# With uvuv pip install anomalib# Or with pippip install anomalib

For more control over the installation, such as specifying the PyTorch backend (e.g., XPU, CUDA and ROCm) or installing extra dependencies for specific models, see the advanced options below.

# CPU support (default, works on all platforms)uv pip install"anomalib[cpu]"# CUDA 12.4 support (Linux/Windows with NVIDIA GPU)uv pip install"anomalib[cu124]"# CUDA 12.1 support (Linux/Windows with NVIDIA GPU)uv pip install"anomalib[cu121]"# CUDA 11.8 support (Linux/Windows with NVIDIA GPU)uv pip install"anomalib[cu118]"# ROCm support (Linux with AMD GPU)uv pip install"anomalib[rocm]"# Intel XPU support (Linux with Intel GPU)uv pip install"anomalib[xpu]"

pip install"anomalib[cu124]"

# Example: Install with OpenVINO support and CUDA 12.4uv pip install"anomalib[openvino,cu124]"# Example: Install all optional dependencies for a CPU-only setupuv pip install"anomalib[full,cpu]"

git clone https://github.com/open-edge-platform/anomalib.gitcd anomalib# Create the virtual environmentuv venv# Sync with the lockfile for a specific backend (e.g., CPU)uv sync --extra cpu# Or for a different backend like CUDA 12.4uv sync --extra cu124# To set up a full development environmentuv sync --extra dev --extra cpu

git clone https://github.com/open-edge-platform/anomalib.gitcd anomalib# Install in editable mode with a specific backendpip install -e".[cpu]"# Install with development dependenciespip install -e".[dev,cpu]"

Anomalib supports both API and CLI-based training approaches:

fromanomalib.dataimportMVTecADfromanomalib.modelsimportPatchcorefromanomalib.engineimportEngine# Initialize componentsdatamodule=MVTecAD()model=Patchcore()engine=Engine()# Train the modelengine.fit(datamodule=datamodule,model=model)

# Train with default settingsanomalib train --model Patchcore --data anomalib.data.MVTecAD# Train with custom categoryanomalib train --model Patchcore --data anomalib.data.MVTecAD --data.category transistor# Train with config fileanomalib train --config path/to/config.yaml

Anomalib provides multiple inference options including Torch, Lightning, Gradio, and OpenVINO. Here's how to get started:

# Load model and make predictionspredictions=engine.predict(datamodule=datamodule,model=model,ckpt_path="path/to/checkpoint.ckpt",)

# Basic predictionanomalib predict --model anomalib.models.Patchcore \                 --data anomalib.data.MVTecAD \                 --ckpt_path path/to/model.ckpt# Prediction with resultsanomalib predict --model anomalib.models.Patchcore \                 --data anomalib.data.MVTecAD \                 --ckpt_path path/to/model.ckpt \                 --return_predictions

📘Note: For advanced inference options including Gradio and OpenVINO, check ourInference Documentation.

Ensure that you have PyTorch with XPU support installed. For more information, please refer to thePyTorch XPU documentation

fromanomalib.dataimportMVTecADfromanomalib.engineimportEngine,SingleXPUStrategy,XPUAcceleratorfromanomalib.modelsimportStfpmengine=Engine(strategy=SingleXPUStrategy(),accelerator=XPUAccelerator(),)engine.train(Stfpm(),datamodule=MVTecAD())

anomalib train --model Padim --data MVTecAD --trainer.accelerator xpu --trainer.strategy xpu_single

Anomalib supports hyperparameter optimization (HPO) usingWeights & Biases andComet.ml.

# Run HPO with Weights & Biasesanomalib hpo --backend WANDB --sweep_config tools/hpo/configs/wandb.yaml

📘Note: For detailed HPO configuration, check ourHPO Documentation.

Track your experiments with popular logging platforms throughPyTorch Lightning loggers:

• 📊 Weights & Biases
• 📈 Comet.ml
• 📉 TensorBoard

Enable logging in your config file to track:

• Hyperparameters
• Metrics
• Model graphs
• Test predictions

📘Note: For logging setup, see ourLogging Documentation.

Evaluate and compare model performance across different datasets:

# Run benchmarking with default configurationanomalib benchmark --config tools/experimental/benchmarking/sample.yaml

💡Tip: Check individual model performance in their respective README files:

• Patchcore Results
• Other Models

If you find Anomalib useful in your research or work, please cite:

@inproceedings{akcay2022anomalib,  title={Anomalib: A deep learning library for anomaly detection},  author={Akcay, Samet and Ameln, Dick and Vaidya, Ashwin and Lakshmanan, Barath and Ahuja, Nilesh and Genc, Utku},  booktitle={2022 IEEE International Conference on Image Processing (ICIP)},  pages={1706--1710},  year={2022},  organization={IEEE}}

We welcome contributions! Check out ourContributing Guide to get started.

Thank you to all our contributors!