SER-Odyssey-Baseline-WavLM-Valence

The model was trained on MSP-Podcast for the Odyssey 2024 Emotion Recognition competition baseline<br> This particular model is the single-task specialized valence model, which predict valence in a range of approximately 0...1.

Benchmarks

CCC based on Test3 and Development sets of the Odyssey Competition

<table style="width:500px"> <tr><th colspan=2 align="center"> Sinle-Task Setup </th></tr> <tr><th colspan=1 align="center">Test 3</th><th colspan=1 align="center">Development</th></tr> <tr> <td align="center">Val</td> <td align="center">Val</td> </tr> <tr> <td align="center"> 0.607</td> <td align="center" >0.709 </td> </tr> </table>

For more details: demo, paper, and GitHub.

@InProceedings{Goncalves_2024,
            author={L. Goncalves and A. N. Salman and A. {Reddy Naini} and L. Moro-Velazquez and T. Thebaud and L. {Paola Garcia} and N. Dehak and B. Sisman and C. Busso},
            title={Odyssey2024 - Speech Emotion Recognition Challenge: Dataset, Baseline Framework, and Results},
            booktitle={Odyssey 2024: The Speaker and Language Recognition Workshop)},
            volume={To appear},
            year={2024},
            month={June},
            address =  {Quebec, Canada},
}

Usage

from transformers import AutoModelForAudioClassification
import librosa, torch

#load model
model = AutoModelForAudioClassification.from_pretrained("3loi/SER-Odyssey-Baseline-WavLM-Valence", trust_remote_code=True)

#get mean/std
mean = model.config.mean
std = model.config.std


#load an audio file
audio_path = "/path/to/audio.wav"
raw_wav, _ = librosa.load(audio_path, sr=model.config.sampling_rate)

#normalize the audio by mean/std
norm_wav = (raw_wav - mean) / (std+0.000001)

#generate the mask
mask = torch.ones(1, len(norm_wav))

#batch it (add dim)
wavs = torch.tensor(norm_wav).unsqueeze(0)


#predict
with torch.no_grad():
    pred = model(wavs, mask)

print(model.config.id2label) 
print(pred)
#{0: 'valence'}
#tensor([[0.3670]])