File size: 2,989 Bytes
b9bfdc9 a71bf01 b9bfdc9 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 |
---
language: fa
datasets:
- ShEMO
tags:
- audio
- automatic-speech-recognition
- speech
- speech-emotion-recognition
license: apache-2.0
---
# Emotion Recognition in Persian (Farsi - fa) Speech using Wav2Vec 2.0
## How to use
### Requirements
```bash
# 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
```
### Prediction
```python
import torch
import torch.nn as nn
import torch.nn.functional as F
import torchaudio
from transformers import AutoConfig, Wav2Vec2FeatureExtractor
import librosa
import IPython.display as ipd
import numpy as np
import pandas as pd
```
```python
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model_name_or_path = "m3hrdadfi/wav2vec2-xlsr-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 = Wav2Vec2ForSpeechClassification.from_pretrained(model_name_or_path).to(device)
```
```python
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
```
```python
path = "/path/to/sadness.wav"
outputs = predict(path, sampling_rate)
```
```bash
[
{'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 | 0.95 | 0.95 | 0.95 | |
| Fear | 0.33 | 0.17 | 0.22 | |
| Happiness | 0.69 | 0.69 | 0.69 | |
| Neutral | 0.91 | 0.94 | 0.93 | |
| Sadness | 0.92 | 0.85 | 0.88 | |
| Surprise | 0.81 | 0.88 | 0.84 | |
| | | | Overal | 0.90 |
## Questions?
Post a Github issue from [HERE](https://github.com/m3hrdadfi/soxan/issues). |