Update orpheus-tts/engine_class.py

orpheus-tts/engine_class.py

@@ -112,23 +112,31 @@ class OrpheusModel:
  
 
 
-    def generate_tokens_sync(self, prompt, voice=None, request_id="req-001", temperature=0.6, top_p=0.8, max_tokens=1200, stop_token_ids = [49158], repetition_penalty=1.3):
+    def generate_tokens_sync(self, prompt, voice=None, request_id="req-001", temperature=0.6, top_p=0.8, max_tokens=1200, stop_token_ids = [128258], repetition_penalty=1.3):
         prompt_string = self._format_prompt(prompt, voice)
-        print(prompt)
+        print(f"DEBUG: Original prompt: {prompt}")
+        print(f"DEBUG: Formatted prompt: {prompt_string}")
+        
         sampling_params = SamplingParams(
         temperature=temperature,
         top_p=top_p,
         max_tokens=max_tokens,  # Adjust max_tokens as needed.
-        stop_token_ids = stop_token_ids, 
-        repetition_penalty=repetition_penalty, 
+        stop_token_ids = stop_token_ids,
+        repetition_penalty=repetition_penalty,
         )
 
         token_queue = queue.Queue()
+        token_count = 0
 
         async def async_producer():
             async for result in self.engine.generate(prompt=prompt_string, sampling_params=sampling_params, request_id=request_id):
                 # Place each token text into the queue.
-                token_queue.put(result.outputs[0].text)
+                token_text = result.outputs[0].text
+                print(f"DEBUG: Generated token {token_count}: {repr(token_text)}")
+                token_queue.put(token_text)
+                nonlocal token_count
+                token_count += 1
+            print(f"DEBUG: Generation completed. Total tokens: {token_count}")
             token_queue.put(None)  # Sentinel to indicate completion.
 
         def run_async():