https://github.com/bagustris/a-vb2022_ccc

Another version of https://github.com/bagustris/A-VB2022 using CCC loss function

https://github.com/bagustris/a-vb2022_ccc

Science Score: 49.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • CITATION.cff file
  • codemeta.json file
    Found codemeta.json file
  • .zenodo.json file
    Found .zenodo.json file
  • DOI references
    Found 1 DOI reference(s) in README
  • Academic publication links
    Links to: ieee.org
  • Committers with academic emails
  • Institutional organization owner
  • JOSS paper metadata
  • Scientific vocabulary similarity
    Low similarity (11.8%) to scientific vocabulary
Last synced: 7 months ago · JSON representation

Repository

Another version of https://github.com/bagustris/A-VB2022 using CCC loss function

Basic Info
  • Host: GitHub
  • Owner: bagustris
  • Language: Python
  • Default Branch: master
  • Size: 46.9 KB
Statistics
  • Stars: 1
  • Watchers: 1
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created over 3 years ago · Last pushed almost 3 years ago
Metadata Files
Readme

README.md

The paper for this repo has been published: Evaluating Variants of wav2vec 2.0 on Affective Vocal Burst Tasks.

A-VB Feature-based

Another version of https://github.com/bagustris/A-VB2022 using CCC loss function (please take a look that repository before trying code in this repository). Feature extraction for 'w2v2-r-er' and 'w2v2-r-vad' are given there. Also install requirements from there (inside feature_based directory).

We provide a 'feature-based' approach for all four tasks to reproduce results in our ICASSP paper (see citation).

Installation

First, make sure you download the data and features, placing them within your working directory. For more info and instructions on how to access the competition data visit competitions.hume.ai.

We suggest creating a virtual environment, and installing the requirements.txt

conda create -n avb-2022-ccc python=3.8 conda activate avb-2022-ccc pip install -r requirements.txt

Example

After the data, labels, and features are downloaded to your working directory, when running main.py set to -d ./ i.e., where features/ and labels/ are, and run:

A-VB High Baseline

python main.py -d /data/A-VB/ -f w2v2-R-emo-vad -t type -e 100 -lr 0.0005 -bs 8 -p 10 --n_seeds 20

Follow the same procedure for each task altering -t.

| Option | Description | | -------------- | -------------------------------------------- | | -d | Set the path to working dir | | -f | Feature set e.g., eGeMAPS | | -t | Task ['high','two','culture','type'] | | -e | Number of Epochs (default: 20) | | -lr | Learning Rate (default: 0.001) | | -bs | Batch Size (default: 8) | | -p | Early Stopping Patience (default: 5) | | --n_seeds | Number of Seeds to run for | | --verbose | Maximum verbosity, (default: quiet) |

Citation

B. T. Atmaja and A. Sasou, “Evaluating Variants of wav2vec 2.0 on Affective Vocal Burst Tasks,” in ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Jun. 2023, pp. 1–5, doi: 10.1109/ICASSP49357.2023.10096552.

Owner

  • Name: Bagus Tris Atmaja
  • Login: bagustris
  • Kind: user
  • Location: Tsukuba
  • Company: AIST

Researcher @aistairc @VibrasticLab

GitHub Events

Total
Last Year

Committers

Last synced: 10 months ago

All Time
  • Total Commits: 17
  • Total Committers: 1
  • Avg Commits per committer: 17.0
  • Development Distribution Score (DDS): 0.0
Past Year
  • Commits: 0
  • Committers: 0
  • Avg Commits per committer: 0.0
  • Development Distribution Score (DDS): 0.0
Top Committers
Name Email Commits
Bagus Tris Atmaja b****s@y****m 17

Issues and Pull Requests

Last synced: 10 months ago