feat(tts): migrate Kokoro TTS to Hugging Face facebook/mms-tts-eng with in-memory optimization

- Replaced previous Kokoro-based TTS pipeline with Hugging Face AutoModelForTextToWaveform.
- Updated to generate WAV in-memory without saving to disk.
- Uses tokenizer + model from app.state for FastAPI conventions.
- Returns audio as bytes or URL (depending on implementation), improving performance and reducing I/O.

app/routes/tts.py

@@ -4,7 +4,6 @@ from fastapi.responses import StreamingResponse
 import numpy as np
 import scipy.io.wavfile as sf
 import torch
-import uuid
 import io
 
 
@@ -22,6 +21,7 @@ def load_model_tts():
 
 
 def save_audio(request: Request, text: str,) -> bytes:
+    """ function to generate audio from text using TTS model """
     model, tokenizer = request.app.state.model_tts, request.app.state.tokenizer_tts
     inputs = tokenizer(text, return_tensors="pt")