app.py

import gradio as gr
import torch
from transformers import AutoProcessor, AutoModelForAudioClassification
import torchaudio

# Load model and processor
model = AutoModelForAudioClassification.from_pretrained("jacksonwambali/my-bird-sound-classifier/bird_model")
processor = AutoProcessor.from_pretrained("jacksonwambali/my-bird-sound-classifier/bird_model")

def classify_sound(audio):
    waveform, sample_rate = audio
    inputs = processor(waveform, sampling_rate=sample_rate, return_tensors="pt")
    with torch.no_grad():
        logits = model(**inputs).logits
    predicted_class_id = logits.argmax().item()
    predicted_label = model.config.id2label[predicted_class_id]
    return predicted_label

iface = gr.Interface(fn=classify_sound, 
                     inputs=gr.Audio(type="torchaudio"), 
                     outputs="label",
                     title="Bird Sound Classifier",
                     description="Upload a bird audio clip to identify the species.")
iface.launch()