metadata

license: cc-by-nc-sa-4.0
language:
  - en
  - de
  - zh
  - fr
  - nl
  - el
  - it
  - es
  - my
  - he
  - sv
  - fa
  - tr
  - ur
library_name: transformers
pipeline_tag: audio-classification
tags:
  - Speech Emotion Recognition
  - SER
  - Transformer
  - HuBERT
  - PyTorch

ExHuBERT: Enhancing HuBERT Through Block Extension and Fine-Tuning on 37 Emotion Datasets

Authors: Shahin Amiriparian, Filip Packań, Maurice Gerczuk, Björn W. Schuller

Fine-tuned HuBERT Large on EmoSet++, comprising 37 datasets, totaling 150,907 samples and spanning a cumulative duration of 119.5 hours. The model is expecting a 3 second long raw waveform resampled to 16 kHz. The original 6 Ouput classes are combinations of low/high arousal and negative/neutral/positive valence. Further details are available in the corresponding paper

Note: This model is for research purpose only.

EmoSet++ subsets used for fine-tuning the model:


ABC	AD	BES	CASIA	CVE
Crema-D	DES	DEMoS	EA-ACT	EA-BMW
EA-WSJ	EMO-DB	EmoFilm	EmotiW-2014	EMOVO
eNTERFACE	ESD	EU-EmoSS	EU-EV	FAU Aibo
GEMEP	GVESS	IEMOCAP	MES	MESD
MELD	PPMMK	RAVDESS	SAVEE	ShEMO
SmartKom	SIMIS	SUSAS	SUBSECO	TESS
TurkishEmo	Urdu

Usage

import torch
import torch.nn as nn
from transformers import AutoModelForAudioClassification, Wav2Vec2FeatureExtractor



# CONFIG and MODEL SETUP
model_name = 'amiriparian/HuBERT-EmoSet'
feature_extractor = Wav2Vec2FeatureExtractor.from_pretrained("facebook/hubert-base-ls960")
AutoModelForAudioClassification.from_pretrained(model_name, trust_remote_code=True)

sampling_rate=16000 
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model = model.to(device)

Citation Info

@inproceedings{Amiriparian24-EEH,
  author = {Shahin Amiriparian and Filip Packan and Maurice Gerczuk and Bj\"orn W.\ Schuller},
  title = {{ExHuBERT: Enhancing HuBERT Through Block Extension and Fine-Tuning on 37 Emotion Datasets}},
  booktitle = {{Proc. INTERSPEECH}}, 
  year = {2024},
  editor = {},
  volume = {},
  series = {},
  address = {Kos Island, Greece},
  month = {September},
  publisher = {ISCA},
}