Some cleanup

triton-inference-server · May 14, 2024 · 2093d1b · 2093d1b
1 parent ee960ee
commit 2093d1b
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 9 deletions.
diff --git a/src/c++/perf_analyzer/genai-perf/genai_perf/llm_metrics.py b/src/c++/perf_analyzer/genai-perf/genai_perf/llm_metrics.py
@@ -429,8 +429,6 @@ def _get_profile_metadata(self, data: dict) -> None:
                     self._response_format = ResponseFormat.OPENAI_CHAT_COMPLETIONS
                 elif "text_completion" in response:
                     self._response_format = ResponseFormat.OPENAI_COMPLETIONS
-                elif "input_ids" in response:
-                    self._response_format = ResponseFormat.OPENAI_COMPLETIONS
                 else:
                     raise RuntimeError("Unknown OpenAI response format.")
 
@@ -538,7 +536,6 @@ def _parse_requests(self, requests: dict) -> LLMMetrics:
             time_to_first_tokens.append(res_timestamps[0] - req_timestamp)
 
             # number of input tokens
-            # input_tokens = self._tokenize_request_inputs(req_inputs)
             len_input_tokens = self._tokenize_request_inputs(req_inputs)
             num_input_tokens.append(len_input_tokens)
 

diff --git a/src/c++/perf_analyzer/genai-perf/genai_perf/main.py b/src/c++/perf_analyzer/genai-perf/genai_perf/main.py
@@ -145,12 +145,13 @@ def run():
 def main():
     # Interactive use will catch exceptions and log formatted errors rather than
     # tracebacks.
-    run()
-    # except Exception as e:
-    #     traceback.print_exc()
-    #     logger = logging.getLogger(__name__)
-    #     logger.error(e)
-    #     return 1
+    try:
+        run()
+    except Exception as e:
+        traceback.print_exc()
+        logger = logging.getLogger(__name__)
+        logger.error(e)
+        return 1
 
     return 0