[https://nvbugs/5596343] [test] Waive flaky GPT-OSS cases (NVIDIA#8904)

Signed-off-by: Xiwen Yu <[email protected]>
Signed-off-by: FredricZ-2007 <[email protected]>

Author: VALLIS-NERIA

tests/integration/defs/accuracy/test_llm_api_pytorch.py

@@ -3591,10 +3591,6 @@ def test_dummy_load_format(self):
     def test_w4_4gpus(self, kv_cache_dtype, moe_backend, tp_size, pp_size,
                       ep_size, attention_dp, cuda_graph, overlap_scheduler,
                       mocker):
-        if get_sm_version() < 100:
-            pytest.skip(
-                "https://nvbugs/5596343: Skip Hopper due to accuracy issue.")
-
         if moe_backend == "TRITON":
             if not IS_TRITON_KERNELS_AVAILABLE:
                 pytest.skip("Triton kernels are not available")

tests/integration/test_lists/waives.txt

@@ -360,6 +360,8 @@ full:RTX/accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_w4_2gpus[ep2-trtllm-
 full:RTX/accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_w4_2gpus[ep2-trtllm-fp8] SKIP (https://nvbugs/5569719)
 full:RTX/accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_w4_2gpus[dp2-trtllm-auto] SKIP (https://nvbugs/5569719)
 full:RTX/accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_w4_2gpus[dp2-trtllm-fp8] SKIP (https://nvbugs/5569719)
+accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_w4_4gpus[ep4-cutlass-auto] SKIP (https://nvbugs/5596343)
+accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_w4_4gpus[tp4-cutlass-auto] SKIP (https://nvbugs/5596343)
 examples/test_phi.py::test_llm_phi_lora_1gpu[Phi-3-mini-4k-instruct-ru-lora-Phi-3-mini-4k-instruct-lora_fp16-base_fp16] SKIP (https://nvbugs/5612313)
 accuracy/test_disaggregated_serving.py::TestGemma3_1BInstruct::test_auto_dtype[False] SKIP (https://nvbugs/5569696)
 accuracy/test_disaggregated_serving.py::TestGemma3_1BInstruct::test_auto_dtype[True] SKIP (https://nvbugs/5569696)