TorchMetrics is a collection of 80+ PyTorch metrics implementations and an easy-to-use API to create custom metrics. It offers:
- A standardized interface to increase reproducibility
- Reduces Boilerplate
- Distributed-training compatible
- Rigorously tested
- Automatic accumulation over batches
- Automatic synchronization between multiple devices
You can use TorchMetrics in any PyTorch model, or within PyTorch Lightning to enjoy the following additional benefits:
- Your data will always be placed on the same device as your metrics
- You can log :class:`~torchmetrics.Metric` objects directly in Lightning to reduce even more boilerplate
For pip users
pip install torchmetricsOr directly from conda
conda install -c conda-forge torchmetrics.. customcalloutitem:: :description: Use this quickstart guide to learn key concepts. :header: New to TorchMetrics? :button_link: pages/quickstart.html
.. customcalloutitem:: :description: Easily use TorchMetrics in your PyTorch Lightning code. :header: TorchMetrics with PyTorch Lightning :button_link: pages/lightning.html
.. customcalloutitem:: :description: View the full list of metrics and filter by task and data type. :header: Metrics :button_link: metrics.html
.. customcalloutitem:: :description: A detailed overview of the TorchMetrics API and concepts. :header: Overview :button_link: pages/overview.html
.. customcalloutitem:: :description: Learn how to implement a custom metric with TorchMetrics. :header: Custom Metrics :button_link: pages/implement.html
.. customcalloutitem:: :description: Detailed descriptions of each API package. :header: API Reference :button_link: references/metric.html
.. toctree:: :maxdepth: 2 :name: guide :caption: User Guide pages/quickstart pages/all-metrics pages/overview pages/implement pages/lightning pages/classification pages/retrieval
.. toctree:: :maxdepth: 2 :name: audio :caption: Audio :glob: audio/*
.. toctree:: :maxdepth: 2 :name: classification :caption: Classification :glob: classification/*
.. toctree:: :maxdepth: 2 :name: image :caption: Image :glob: image/*
.. toctree:: :maxdepth: 2 :name: detection :caption: Detection :glob: detection/*
.. toctree:: :maxdepth: 2 :name: pairwise :caption: Pairwise :glob: pairwise/*
.. toctree:: :maxdepth: 2 :name: regression :caption: Regression :glob: regression/*
.. toctree:: :maxdepth: 2 :name: retrieval :caption: Retrieval :glob: retrieval/*
.. toctree:: :maxdepth: 2 :name: text :caption: Text :glob: text/*
.. toctree:: :maxdepth: 2 :name: aggregation :caption: Aggregation :glob: aggregation/*
.. toctree:: :maxdepth: 2 :name: wrappers :caption: Wrappers :glob: wrappers/*
.. toctree:: :maxdepth: 3 :name: metrics :caption: API Reference references/metric references/utilities
.. toctree:: :maxdepth: 1 :name: community :caption: Community governance generated/CODE_OF_CONDUCT.md generated/CONTRIBUTING.md generated/CHANGELOG.md