____ ____ _ ____ _____ _ _
| _ \ / ___| / \ / ___|| ____| _ __ ___ ___ __| | ___| |___
| | | | | / _ \ \___ \| _| _____| '_ ` _ \ / _ \ / _` |/ _ \ / __|
| |_| | |___ / ___ \ ___) | |__|_____| | | | | | (_) | (_| | __/ \__ \
|____/ \____/_/ \_\____/|_____| |_| |_| |_|\___/ \__,_|\___|_|___/
DCASE-models is an open-source Python library for rapid prototyping of environmental sound analysis systems, with an emphasis on deep–learning models. The library has a flat and light design that allows easy extension and integration with other existing tools.
See https://dcase-models.readthedocs.io for a complete reference manual and introductory tutorials.
We recommend to install DCASE-models in a dedicated virtual environment. For instance, using anaconda:
conda create -n dcase python=3.6
conda activate dcase
For GPU support:
conda install cudatoolkit cudnn
DCASE-models uses SoX for functions related to the datasets. You can install it in your conda environment by:
conda install -c conda-forge sox
When installing the library, you must select the tensorflow variant: version 1 (CPU-only or GPU) or version 2.
pip install DCASE-models[keras_tf] # for tensorflow 1 CPU-only version
pip install DCASE-models[keras_tf_gpu] # for tensorflow 1 GPU version
pip install DCASE-models[tf2] # for tensorflow 2
To include visualization related dependencies, run the following instead:
pip install DCASE-models[visualization]
There are several ways to use this library. In this repository, we accompany the library with three types of examples.
Note that the default parameters for each model, dataset and feature representation, are stored in
parameters.jsonon the root directory.
The folder scripts includes python scripts for data downloading, feature extraction, model training and testing, and fine-tuning. These examples show how to use DCASE-models within a python script.
The folder notebooks includes a list of notebooks that replicate scientific experiments using DCASE-models.
The folder visualization includes a user interface to define, train and visualize the models defined in this library.
Go to DCASE-models folder and run:
python -m visualization.index
Then, open your browser and navigate to:
http://localhost:8050/