https://github.com/idiap/exvo-2022
Extracting pre-trained self-supervised embeddings for ICML ExVO 2022 challenge
https://github.com/idiap/exvo-2022
Last synced: 4 months ago
JSON representation
Extracting pre-trained self-supervised embeddings for ICML ExVO 2022 challenge
- Host: GitHub
- URL: https://github.com/idiap/exvo-2022
- Owner: idiap
- Created: 2022-07-19T11:44:40.000Z (almost 4 years ago)
- Default Branch: main
- Last Pushed: 2022-07-19T11:45:21.000Z (almost 4 years ago)
- Last Synced: 2025-06-04T22:46:46.870Z (about 1 year ago)
- Language: Python
- Size: 12.7 KB
- Stars: 5
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSES/GPL-3.0-only.txt
Awesome Lists containing this project
README
# Extracting pre-trained self-supervised embeddings for ICML ExVO 2022 challenge
This python code generates Generation SSL (WavLM) feature representation:
- Estimate frame-level SSL embeddings from pre-trained models
- Generates fixed-length feature representation by using mean and standard deviation functionals
The extracted feature representations were used during ICML ExVo 2022 challenge, for more details see:
T. Purohit, I. B. Mahmoud, B. Vlasenko, M. Magimai.-Doss. Comparing supervised and self-supervised embedding for ExVo Multi-Task
learning track, workshop ICML ExVo 2022
## Install
Create a Conda environment with python 3.10 and activate it:
```bash
conda create -n env python=3.10
conda activate env
```
Install all required python packages
```bash
pip install -r requirements.txt
```