Update app.py

app.py

@@ -39,12 +39,25 @@ TTS_MODEL_PATH = "path/to/your/espnet/kan-bayashi_ljspeech_vits"  # Replace with
 TTS_CONFIG_PATH = os.path.join(TTS_MODEL_PATH, "config.yaml") # Replace with your config.yaml
 TTS_VOCAB_PATH = os.path.join(TTS_MODEL_PATH, "train.json") # Replace with your train.json
 
+# Load the configuration using yaml
+with open(TTS_CONFIG_PATH, "r") as f:
+    config = yaml.safe_load(f)
+
+# Correctly handle the threshold parameter
+threshold = config.get("threshold", 0.5)  # Get threshold, default to 0.5 if not found
+if isinstance(threshold, str):  # Check if it's a string
+    try:
+        threshold = float(threshold)  # Try converting to float
+    except ValueError:
+        threshold = 0.5  # If conversion fails, use a default value
+
+
 tts = Text2Speech(
     TTS_MODEL_PATH,
     TTS_CONFIG_PATH,
     TTS_VOCAB_PATH,
     device="cuda" if torch.cuda.is_available() else "cpu",
-    # You can customize the speed and other parameters here if needed
+    threshold=threshold  # Pass the corrected threshold value
 )
 fastspeech_config = get_fastspeech_config(TTS_CONFIG_PATH)