Audio embeddings based on pruned Look, Listen, and Learn (L3) models for the Edge
Look, Listen, and Learn (L3) , a recently proposed state-of-the-art transfer learning technique, helps to train self-supervised deep audio embedding through binary Audio-Visual Correspondence. This embedding can be used to train a variety of downstream audio classification tasks which has limited data. However, with close to 4.7 million parameters, L3-Net is 18 MB in size making it infeasible for small edge devices, such as 'motes' that use a single microcontroller and limited memory to achieve long-lived self-powered operation.
In EdgeL3 , we comprehensively explore the feasibility of compressing the L3-Net for mote-scale inference. We used pruning, ablation, and knowledge distillation techniques to show that the originally proposed L3-Net architecture is substantially overparameterized, not only for AVC but for the target task of sound classification as evaluated on two popular downstream datasets, US8K and ESC50. EdgeL3, a 95.45% sparsified version of L3-Net, provides a useful reference model for approximating L3 audio embedding for transfer learning.
edgel3 is an open-source Python library for downloading the sparsified L3 models and computing deep audio embeddings from such models. The sparse audio embedding models provided have been re-trained using two different mechanisms as described in the paper. The code for the model and training implementation can be found here
Install Tensorflow (CPU-only/GPU) variant that best fits your usecase.
On most platforms, either of the following commands should properly install Tensorflow:
pip install tensorflow # CPU-only version pip install tensorflow-gpu # GPU version
For more detailed information, please consult the Tensorflow installation documentation.
EdgeL3 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
For more detailed information, please consult the
pysoundfile installation documentation.
The simplest way to install EdgeL3 is by using
pip, which will also install the additional required dependencies
if needed. To install EdgeL3 using
pip, simply run
pip install edgel3
To install the latest version of EdgeL3 from source:
Clone or pull the lastest version:
git clone https://github.com/ksangeeta2429/edgel3.git
Install using pip to handle python dependencies: cd edgel3 pip install -e .
Please cite the following papers when using EdgeL3 in your work:
 EdgeL3: Compressing L3-Net for Mote-Scale Urban Noise Monitoring
Sangeeta Kumari, Dhrubojyoti Roy, Mark Cartwright, Juan Pablo Bello, and Anish Arora.
Parallel AI and Systems for the Edge (PAISE), Rio de Janeiro, Brazil, May 2019.
 Look, Listen and Learn More: Design Choices for Deep Audio Embeddings
Jason 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.
 Look, Listen and Learn
Relja Arandjelović and Andrew Zisserman
IEEE International Conference on Computer Vision (ICCV), Venice, Italy, Oct. 2017.
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size||File type||Python version||Upload date||Hashes|
|Filename, size edgel3-0.1.0.tar.gz (17.9 kB)||File type Source||Python version None||Upload date||Hashes View hashes|