More cleanup from hwoo CR

triton-inference-server · Oct 18, 2023 · 51aedfa · 51aedfa
1 parent d6e7c74
commit 51aedfa
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 13 deletions.
diff --git a/model_analyzer/config/generate/perf_analyzer_config_generator.py b/model_analyzer/config/generate/perf_analyzer_config_generator.py
@@ -379,21 +379,19 @@ def _update_perf_config_based_on_parameter_combination(
     ) -> None:
         if "request-parameter" in parameter_combination:
             request_parameter = parameter_combination["request-parameter"]
-            max_token = self._extract_max_token_from_request_parameter(
+            max_tokens = self._extract_max_tokens_from_request_parameter(
                 request_parameter
             )
             parameter_combination["request-period"] = (
-                max_token if max_token < 10 else 10
+                max_tokens if max_tokens < 10 else 10
             )
 
         perf_config.update_config(parameter_combination)
 
-    def _extract_max_token_from_request_parameter(self, request_parameter: str) -> int:
-        max_token_start = request_parameter.find(":")
-        max_token_stop = request_parameter.find(":", max_token_start + 1)
-        max_token = int(request_parameter[max_token_start + 1 : max_token_stop])
+    def _extract_max_tokens_from_request_parameter(self, request_parameter: str) -> int:
+        _, max_tokens, _ = request_parameter.split(":")
 
-        return max_token
+        return int(max_tokens)
 
     def _update_perf_config_based_on_perf_analyzer_flags(
         self, perf_config: PerfAnalyzerConfig

diff --git a/tests/common/test_utils.py b/tests/common/test_utils.py
@@ -305,15 +305,15 @@ def construct_perf_analyzer_config(
         pa_config._args["concurrency-range"] = concurrency
 
     if llm_search_mode:
-        pa_config._args["request-parameter"] = (
-            "max_tokens:" + str(max_token_count) + ":int"
-        )
+        pa_config._args["request-parameter"] = f"max_tokens:{str(max_token_count)}:int"
+
         pa_config._args["request-period"] = (
             max_token_count if max_token_count < 10 else 10
         )
-        pa_config._args["input-data"] = (
-            DEFAULT_INPUT_JSON_PATH + "/input-data-" + str(text_input_length) + ".json"
-        )
+        pa_config._args[
+            "input-data"
+        ] = f"{DEFAULT_INPUT_JSON_PATH}/input-data-{str(text_input_length)}.json"
+
         pa_config._args["streaming"] = "True"
 
     pa_config._args["measurement-mode"] = DEFAULT_MEASUREMENT_MODE