diff --git a/model_analyzer/perf_analyzer/perf_analyzer.py b/model_analyzer/perf_analyzer/perf_analyzer.py index 0075aa6b9..83e95f763 100755 --- a/model_analyzer/perf_analyzer/perf_analyzer.py +++ b/model_analyzer/perf_analyzer/perf_analyzer.py @@ -550,7 +550,7 @@ def _calculate_avg_token_to_token_latency(self, llm_output: Dict) -> float: total_response_to_response_latency = 0 prev_response = request["response_timestamps"][0] for response in request["response_timestamps"][1:]: - total_response_to_reponse_latency = response - prev_response + total_response_to_response_latency = response - prev_response prev_response = response total_token_to_token_latency += total_response_to_response_latency / len(