Update app.py

app.py

@@ -198,7 +198,7 @@ def tts_streaming(text: str, voice: str = "af_heart", speed: float = 1.0, format
             if i == 0:
                 tokens_to_send = [0] + chunk_tokens + [0]
             else:
-                tokens_to_send = [prev_last_token] + chunk_tokens + [0]
+                tokens_to_send = [0] + [prev_last_token] + chunk_tokens + [0]
             #     token_to_send = [0] + chunk_tokens + [0]
 
             # Save the last token of this chunk for the next iteration.
@@ -206,9 +206,10 @@ def tts_streaming(text: str, voice: str = "af_heart", speed: float = 1.0, format
 
             # Prepare the model input (a batch of one sequence).
             final_token = [tokens_to_send]
+            print(final_token)
 
             # Use the number of tokens to select the appropriate style vector.
-            style_index = len(chunk_tokens) + 2
+            style_index = len(chunk_tokens) + 1
             if style_index >= len(voices):
                 style_index = len(voices) - 1  # Fallback if index is out-of-bounds.
             ref_s = voices[style_index]