fix: specify device for Whisper model loading

- Updated the Whisper model loading to explicitly set the device to "cpu" for improved compatibility and performance.

apps/audio_cloning/vallex/main.py

@@ -170,7 +170,9 @@ if not os.path.exists(OUTPUT_DIR_WHISPER):
 try:
     logger.info("Loading Whisper model...")
     model_name = "tiny"
-    whisper_model = whisper.load_model(model_name, download_root=OUTPUT_DIR_WHISPER)
+    whisper_model = whisper.load_model(
+        model_name, device="cpu", download_root=OUTPUT_DIR_WHISPER
+    )
     logger.info("Whisper model loaded successfully")
 except NotImplementedError as e:
     logger.error("Error on loading Whisper model: %s", e)