README.md

metadata

language: fa
datasets:
  - ShEMO
tags:
  - audio
  - speech
  - speech-emotion-recognition
license: apache-2.0

Emotion Recognition in Persian (fa) Speech using HuBERT

How to use

Requirements

# requirement packages
!pip install git+https://github.com/huggingface/datasets.git
!pip install git+https://github.com/huggingface/transformers.git
!pip install torchaudio
!pip install librosa

!git clone https://github.com/m3hrdadfi/soxan.git .

Prediction

import torch
import torch.nn as nn
import torch.nn.functional as F
import torchaudio
from transformers import AutoConfig, Wav2Vec2FeatureExtractor
from src.models import Wav2Vec2ForSpeechClassification, HubertForSpeechClassification

import librosa
import IPython.display as ipd
import numpy as np
import pandas as pd

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model_name_or_path = "m3hrdadfi/hubert-base-persian-speech-emotion-recognition"
config = AutoConfig.from_pretrained(model_name_or_path)
feature_extractor = Wav2Vec2FeatureExtractor.from_pretrained(model_name_or_path)
sampling_rate = feature_extractor.sampling_rate
model = HubertForSpeechClassification.from_pretrained(model_name_or_path).to(device)

def speech_file_to_array_fn(path, sampling_rate):
    speech_array, _sampling_rate = torchaudio.load(path)
    resampler = torchaudio.transforms.Resample(_sampling_rate)
    speech = resampler(speech_array).squeeze().numpy()
    return speech


def predict(path, sampling_rate):
    speech = speech_file_to_array_fn(path, sampling_rate)
    inputs = feature_extractor(speech, sampling_rate=sampling_rate, return_tensors="pt", padding=True)
    inputs = {key: inputs[key].to(device) for key in inputs}

    with torch.no_grad():
        logits = model(**inputs).logits

    scores = F.softmax(logits, dim=1).detach().cpu().numpy()[0]
    outputs = [{"Label": config.id2label[i], "Score": f"{round(score * 100, 3):.1f}%"} for i, score in enumerate(scores)]
    return outputs

path = "/path/to/sadness.wav"
outputs = predict(path, sampling_rate)

[
{'Label': 'Anger', 'Score': '0.0%'},
{'Label': 'Fear', 'Score': '0.0%'},
{'Label': 'Happiness', 'Score': '0.0%'},
{'Label': 'Neutral', 'Score': '0.0%'},
{'Label': 'Sadness', 'Score': '99.9%'},
{'Label': 'Surprise', 'Score': '0.0%'}
]

Evaluation

The following tables summarize the scores obtained by model overall and per each class.

Emotions	precision	recall	f1-score	accuracy
Anger	0.96	0.96	0.96
Fear	1.00	0.50	0.67
Happiness	0.79	0.87	0.83
Neutral	0.93	0.94	0.93
Sadness	0.87	0.94	0.91
Surprise	0.97	0.75	0.85
			Overal	0.92

Questions?

Post a Github issue from HERE.