llm_server/routes/openai_request_handler.py

@@ -25,7 +25,6 @@ class OpenAIRequestHandler(RequestHandler):
         self.prompt = None
 
     def handle_request(self) -> Tuple[flask.Response, int]:
-        print('recieved request')
         assert not self.used
 
         if opts.openai_silent_trim:
@@ -39,7 +38,6 @@ class OpenAIRequestHandler(RequestHandler):
             return invalid_response
 
         if opts.openai_moderation_enabled and opts.openai_api_key and is_api_key_moderated(self.token):
-            print('moderating', self.token)
             try:
                 # Gather the last message from the user and all preceding system messages
                 msg_l = self.request.json['messages'].copy()
@@ -60,17 +58,14 @@ class OpenAIRequestHandler(RequestHandler):
                 traceback.print_exc()
 
         # TODO: support Ooba
-        print('converting to vllm')
         self.parameters = oai_to_vllm(self.parameters, hashes=True, mode=self.cluster_backend_info['mode'])
 
-        print('generating')
         llm_request = {**self.parameters, 'prompt': self.prompt}
         (success, _, _, _), (backend_response, backend_response_status_code) = self.generate_response(llm_request)
 
         model = self.request_json_body.get('model')
 
         if success:
-            print('sent success response')
             return self.build_openai_response(self.prompt, backend_response.json['results'][0]['text'], model=model), backend_response_status_code
         else:
             return backend_response, backend_response_status_code