is this the correct HF inference code for this model?

[sleepingcat4]

this is the code I used for inference for SONICS model:

from sonics import HFAudioClassifier
model = HFAudioClassifier.from_pretrained("awsaf49/sonics-spectttra-gamma-5s")

audio_file = "audio2.mp3"; import torchaudio 
waveform, sr = torchaudio.load(audio_file)

predictions = model(waveform)

import torch

probs = torch.sigmoid(predictions)
print("Probabilities:", probs)

pred_class = torch.argmax(predictions, dim=0) # or torch.argmax(probs, dim=0)
print("Predicted class index:", pred_class.item())

@awsaf49 is this the correct way of performing inference?

[josx]

Hey here is a complete example on HF, check