OpenVINO™ Toolkit Test Data Repository

This repository contains the data you need to pass all tests successfully.

Tests That Require Using This Repository

• MklDnnFunctionalTests
• MyriadFunctionalTests
• InferenceEngineCAPITests
• InferenceEnginePythonAPITests

Repository Components

• Test models are required for low-precision transformations and the OpenVINO™ toolkit Inference Engine bridges validation.
• Image data is required for low-precision transformations and the OpenVINO™ toolkit Inference Engine bridges validation.
• IE Class test XML-files are required for the Inference Engine Core validation.
• VPU test weights are required for some VPU layers validation.

License

Deep Learning Deployment toolkit is licensed under the Apache License Version 2.0. By contributing to the project, you agree to the license and copyright terms therein and release your contribution under these terms.

Usage

Choose one of the ways to download this repository and follow the steps to start using the tests:

• Download the repository automatically

  1. Add the additional CMake -DENABLE_DATA=ON option when building the project. Use the -DENABLE_FUNCTIONAL_TESTS=ON option to download the repository automatically to the inference-engine\temp\data\src\data directory.
  2. Go to the inference-engine\temp\models\src\data directory, open a terminal, and enter the command to download the large file storage:

  git lfs pull

  3. Run tests

• Clone the repository manually

  1. Open a terminal and clone the repository:

  git clone https://github.com/openvinotoolkit/testdata.git

  2. In the same terminal, go to the repository directory:

  cd testdata

  3. Download the large file storage:

  git lfs pull

  4. Set environment variables. Replace {PATH_TO_REPOSITORY} with the directory you cloned the repository to:

  DATA_PATH={PATH_TO_REPOSITORY}/testdata

  MODELS_PATH={PATH_TO_REPOSITORY}/testdata

  5. Run tests

How to Contribute

Your contributions are welcome. Follow these steps:

1. Make sure you can build the product and run all tests and samples with your patch.
2. Add the license to third-party-programs.txt. If the license is not yet in the repository, add a file with it as well.
3. Submit a pull request and explain its purpose in the description.

We assign your pull request to the maintainers. Once you obtain an approval, you can add new test data. The maintainers review your contribution, request additional fixes or modifications if necessary, and merge the pull request into the GitHub* repository.

Support

To report questions, issues, and suggestions use GitHub Issues.