fix tests A10-PyTorch-1.test_e2e.test_openai_chat_multimodal_example and A10-PyTorch-1.test_e2e.test_trtllm_serve_multimodal_example

johncalesp · johncalesp · commit 85ec2fcfad95 · 2025-10-12T18:43:24.000-04:00
Signed-off-by: John Calderon &lt;jcalderon@nvidia.com&gt;
diff --git a/tests/unittest/llmapi/apps/_test_openai_chat_multimodal.py b/tests/unittest/llmapi/apps/_test_openai_chat_multimodal.py
@@ -48,7 +48,7 @@ def server(model_name: str, temp_extra_llm_api_options_file: str):
     model_path = get_model_path(model_name)
     args = [
         "--extra_llm_api_options", temp_extra_llm_api_options_file,
-        "--max_batch_size", "64"
+        "--max_batch_size", "64", "--enable_chunked_prefill"
     ]
     with RemoteOpenAIServer(model_path, args) as remote_server:
         yield remote_server
diff --git a/tests/unittest/llmapi/apps/_test_trtllm_serve_multimodal_example.py b/tests/unittest/llmapi/apps/_test_trtllm_serve_multimodal_example.py
@@ -46,7 +46,10 @@ def temp_extra_llm_api_options_file(request):
 @pytest.fixture(scope="module")
 def server(model_name: str, temp_extra_llm_api_options_file: str):
     model_path = get_model_path(model_name)
-    args = ["--extra_llm_api_options", temp_extra_llm_api_options_file]
+    args = [
+        "--enable_chunked_prefill", "--extra_llm_api_options",
+        temp_extra_llm_api_options_file
+    ]
     with RemoteOpenAIServer(model_path, port=8000,
                             cli_args=args) as remote_server:
         yield remote_server

Original file line number	Diff line number	Diff line change
`@@ -48,7 +48,7 @@ def server(model_name: str, temp_extra_llm_api_options_file: str):`
`48`	`48`	`model_path = get_model_path(model_name)`
`49`	`49`	`args = [`
`50`	`50`	`"--extra_llm_api_options", temp_extra_llm_api_options_file,`
`51`		`- "--max_batch_size", "64"`
	`51`	`+ "--max_batch_size", "64", "--enable_chunked_prefill"`
`52`	`52`	`]`
`53`	`53`	`with RemoteOpenAIServer(model_path, args) as remote_server:`
`54`	`54`	`yield remote_server`