Awesome

OpenL3

OpenL3 is an open-source Python library for computing deep audio and image embeddings.

Please refer to the documentation for detailed instructions and examples.

UPDATE: Openl3 now has Tensorflow 2 support!

NOTE: Whoops! A bug was reported in the training code, with the effect that positive audio-image pairs that come from the same video do not necessarily overlap in time. Nonetheless, the embedding still seems to capture useful semantic information.

The audio and image embedding models provided here are published as part of [1], and are based on the Look, Listen and Learn approach [2]. For details about the embedding models and how they were trained, please see:

Look, Listen and Learn More: Design Choices for Deep Audio Embeddings Aurora Cramer, Ho-Hsiang Wu, Justin Salamon, and Juan Pablo Bello. IEEE Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP), pages 3852–3856, Brighton, UK, May 2019.

Installing OpenL3

Dependencies

libsndfile

OpenL3 depends on the pysoundfile module to load audio files, which depends on the non-Python library libsndfile. On Windows and macOS, these will be installed via pip and you can therefore skip this step. However, on Linux this must be installed manually via your platform's package manager. For Debian-based distributions (such as Ubuntu), this can be done by simply running

apt-get install libsndfile1

Alternatively, if you are using conda, you can install libsndfile simply by running

conda install -c conda-forge libsndfile

For more detailed information, please consult the pysoundfile installation documentation.

Tensorflow

Starting with openl3>=0.4.0, Openl3 has been upgraded to use Tensorflow 2. Because Tensorflow 2 and higher now includes GPU support, tensorflow>=2.0.0 is included as a dependency and no longer needs to be installed separately.

If you are interested in using Tensorflow 1.x, please install using pip install 'openl3<=0.3.1'.

Tensorflow 1x & OpenL3 <= v0.3.1

Because Tensorflow 1.x comes in CPU-only and GPU variants, we leave it up to the user to install the version that best fits their usecase.

On most platforms, either of the following commands should properly install Tensorflow:

pip install "tensorflow<1.14" # CPU-only version
pip install "tensorflow-gpu<1.14" # GPU version

For more detailed information, please consult the Tensorflow installation documentation.

Installing OpenL3

The simplest way to install OpenL3 is by using pip, which will also install the additional required dependencies if needed. To install OpenL3 using pip, simply run

pip install openl3

To install the latest version of OpenL3 from source:

Clone or pull the latest version, only retrieving the main branch to avoid downloading the branch where we store the model weight files (these will be properly downloaded during installation).
```
 git clone git@github.com:marl/openl3.git --branch main --single-branch
```
Install using pip to handle python dependencies. The installation also downloads model files, which requires a stable network connection.
```
 cd openl3
 pip install -e .
```

Using OpenL3

To help you get started with OpenL3 please see the tutorial.

Acknowledging OpenL3

Please cite the following papers when using OpenL3 in your work:

[1] Look, Listen and Learn More: Design Choices for Deep Audio Embeddings Aurora Cramer, Ho-Hsiang Wu, Justin Salamon, and Juan Pablo Bello. IEEE Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP), pages 3852–3856, Brighton, UK, May 2019.

[2] Look, Listen and Learn Relja Arandjelović and Andrew Zisserman IEEE International Conference on Computer Vision (ICCV), Venice, Italy, Oct. 2017.

Model Weights License

The model weights are made available under a Creative Commons Attribution 4.0 International (CC BY 4.0) License.