Review changes

Signed-off-by: Xin Yang <[email protected]>

Author: xyang16

tests/lora/test_gptoss_tp.py

@@ -77,7 +77,6 @@ def test_gpt_oss_lora(gptoss20b_lora_files):
         max_loras=4,
         max_lora_rank=8,
         compilation_config=vllm.config.CompilationConfig(  # Avoid OOM
-            max_cudagraph_capture_size=48,
             cudagraph_specialize_lora=False,
         ),
     )
@@ -99,7 +98,6 @@ def test_gpt_oss_lora_tp2(gptoss20b_lora_files, fully_sharded_loras):
         tensor_parallel_size=2,
         fully_sharded_loras=fully_sharded_loras,
         compilation_config=vllm.config.CompilationConfig(  # Avoid OOM
-            max_cudagraph_capture_size=48,
             cudagraph_specialize_lora=False,
         ),
     )