File size: 1,367 Bytes
35545b0 c3d029a 8dda170 35545b0 c3d029a f9b0e0d 8dda170 35545b0 d3aaef0 c3d029a d3aaef0 35545b0 8dda170 35545b0 8dda170 d3aaef0 |
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 |
from transformers import pipeline
from transformers import AutoModelForAudioClassification
import gradio as gr
import librosa
import torch
import numpy as np
mean, std = -8.278621631819787e-05, 0.08485510250851999
id2label = {0: 'arousal', 1: 'dominance', 2: 'valence'}
def classify_audio(audio_file):
model = AutoModelForAudioClassification.from_pretrained("3loi/SER-Odyssey-Baseline-WavLM-Multi-Attributes", trust_remote_code=True)
sr, raw_wav = audio_file
y = raw_wav.astype(np.float32)
y /= np.max(np.abs(y))
norm_wav = (y - mean) / (std+0.000001)
mask = torch.ones(1, len(norm_wav))
wavs = torch.tensor(norm_wav).unsqueeze(0)
pred = model(wavs, mask).detach().numpy()
output = ''
# for i, audio_pred in enumerate(pred):
# output[i] = {}
for att_i, att_val in enumerate(pred[0]):
output += "{}: \t{:0.4f}\n".format(id2label[att_i], att_val)
return output
def main():
iface = gr.Interface(fn=classify_audio, inputs=gr.Audio(sources=["upload", "microphone"], label="Audio file"),
outputs=gr.Text(), title="Speech Emotion Recognition App",
description="Upload an audio file and hit the 'Submit'\
button")
iface.launch()
if __name__ == '__main__':
main()
|