From 8433091630408538ef0b8b4fe8c8b43ce3017986 Mon Sep 17 00:00:00 2001 From: Mike Iovine Date: Wed, 4 Jun 2025 20:49:37 -0400 Subject: [PATCH] [infra] Unwaive unittests/_torch (#4919) Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> --- tests/integration/test_lists/waives.txt | 7 ++++++- tests/unittest/_torch/speculative/test_eagle3.py | 1 + 2 files changed, 7 insertions(+), 1 deletion(-) diff --git a/tests/integration/test_lists/waives.txt b/tests/integration/test_lists/waives.txt index 36f0273bd0..2ba9648489 100644 --- a/tests/integration/test_lists/waives.txt +++ b/tests/integration/test_lists/waives.txt @@ -381,9 +381,14 @@ accuracy/test_cli_flow.py::TestMixtral8x22B::test_int8_plugin_tp8[renormalize-te test_e2e.py::test_ptp_quickstart_advanced_8gpus[Nemotron-Ultra-253B-nemotron-nas/Llama-3_1-Nemotron-Ultra-253B-v1] SKIP (https://nvbugs/5273697) examples/test_gpt.py::test_starcoder_fp8_quantization_2gpu[starcoder] SKIP (https://nvbugs/5144931) examples/test_gpt.py::test_starcoder_fp8_quantization_2gpu[starcoderplus] SKIP (https://nvbugs/5144931) -unittest/_torch -k "not (modeling or multi_gpu or auto_deploy)" SKIP (https://nvbugs/5280806) examples/test_whisper.py::test_llm_whisper_general[large-v3-disable_gemm_plugin-disable_attention_plugin-disable_weight_only-float16-nb:1-use_python_runtime] SKIP (https://nvbugs/5244570) unittest/_torch/speculative/test_eagle3.py SKIP (https://nvbugs/5280806) +unittest/_torch/modules/test_fused_moe.py SKIP (https://nvbugspro.nvidia.com/bug/5324229) +unittest/_torch/modules/test_moe_load_balancer.py SKIP (https://nvbugspro.nvidia.com/bug/5324229) +unittest/_torch/speculative/test_ngram.py SKIP (https://nvbugspro.nvidia.com/bug/5324239) +unittest/_torch/test_pytorch_model_engine.py SKIP (https://nvbugspro.nvidia.com/bug/5324248) +unittest/_torch/test_resource_manager.py SKIP (https://nvbugspro.nvidia.com/bug/5324252) +unittest/_torch/thop/test_selective_scan_op.py SKIP (https://nvbugspro.nvidia.com/bug/5324258) triton_server/test_triton_rcca.py::test_mistral_beam_search[rcca_4714407-True-10---False-True-False-0-128-disableDecoupleMode-inflight_fused_batching-disableTrtOverlap--guaranteed_no_evict---1-1-1-False-ensemble] SKIP (https://nvbugs/5240060) triton_server/test_triton.py::test_triton_extensive[triton-extensive] SKIP triton_server/test_triton.py::test_gpt_speculative_decoding[gpt-speculative-decoding] SKIP diff --git a/tests/unittest/_torch/speculative/test_eagle3.py b/tests/unittest/_torch/speculative/test_eagle3.py index b698c8a389..1906f47163 100644 --- a/tests/unittest/_torch/speculative/test_eagle3.py +++ b/tests/unittest/_torch/speculative/test_eagle3.py @@ -17,6 +17,7 @@ from utils.llm_data import llm_models_root [[True, "TRTLLM"], [False, "TRTLLM"], [True, "FLASHINFER"], [False, "FLASHINFER"]]) def test_llama_eagle3(use_cuda_graph: bool, attn_backend: str): + pytest.skip("Test is hanging") total_mem_gb = torch.cuda.get_device_properties(0).total_memory / 1e9 if total_mem_gb < 35: pytest.skip("Not enough memory to load target + draft model")