app.py

# Import necessary modules
import gradio as gr
from transformers import pipeline

# Load the fine-tuned model from Hugging Face
pipe = pipeline("automatic-speech-recognition", model="Futuresony/whisper-small-sw")

# Function to transcribe audio
def transcribe(audio):
    if audio is None:
        return "Please upload or record an audio file."
    print("Transcribing audio...")
    result = pipe(audio)["text"]
    return result

# Gradio App
with gr.Blocks() as demo:
    gr.Markdown("# 🎙️ Swahili Speech-to-Text Transcription App")
    
    with gr.Row():
        audio_input = gr.Audio(source="microphone", type="filepath", label="🎤 Record Audio")
        file_input = gr.Audio(source="upload", type="filepath", label="📂 Upload Audio File")

    transcribe_button = gr.Button("Transcribe")
    output_text = gr.Textbox(label="📝 Transcription Output")

    transcribe_button.click(transcribe, inputs=[audio_input], outputs=output_text)
    transcribe_button.click(transcribe, inputs=[file_input], outputs=output_text)

# Launch the app
demo.launch()