Create app.py

app.py

@@ -0,0 +1,65 @@
+# Import libraries 
+import io
+import os 
+import json
+import subprocess
+import mmap
+import numpy
+import soundfile
+import torchaudio
+import torch
+import gradio as gr 
+from collections import defaultdict
+from pathlib import Path
+from pydub import AudioSegment
+from seamless_communication.inference import Translator
+from seamless_communication.streaming.dataloaders.s2tt import SileroVADSilenceRemover
+
+
+# Mapping of language names to their respective codes
+language_codes = {
+    "English": "eng",
+    "Spanish": "spa",
+    "French": "fra",
+    "German": "deu",
+    "Italian": "ita",
+    "Chinese": "cmn"
+}
+
+# Function to handle the translation to target language 
+def translate_audio(audio_file, target_language):
+    language_code = language_codes[target_language] 
+    output_file = "translated_audio.wav"
+    
+    command = f"expressivity_predict {audio_file} --tgt_lang {language_code} \
+               --model_name seamless_expressivity --vocoder_name vocoder_pretssel \
+               --gated-model-dir seamlessmodel --output_path {output_file}"
+    subprocess.run(command, shell=True)
+
+    if os.path.exists(output_file):
+        print(f"File created successfully: {output_file}")
+    else:
+        print(f"File not found: {output_file}")
+
+    return output_file
+
+
+# Define inputs
+inputs = [
+    gr.Audio(type = "filepath", label = "Audio_file"), 
+    gr.Dropdown(["English", "Spanish", "French", "German", "Italian", "Chinese"], label = "Target_Language")
+]
+
+
+# Create Gradio interface
+iface = gr.Interface(
+    fn=translate_audio,
+    inputs= inputs, 
+    outputs=gr.Audio(label="Translated Audio"),
+    title="Seamless Expressive Audio Translator",
+    description="Translate your audio into different languages with expressive styles."
+)
+
+# Run the application 
+if __name__ == "__main__":
+    iface.launch()