From e9f78c687a5f31011b5de6c6a3280e9515635ba6 Mon Sep 17 00:00:00 2001 From: Fanrong Li <23290157+lfr-0531@users.noreply.github.com> Date: Mon, 3 Nov 2025 16:34:52 +0800 Subject: [PATCH] [https://nvbugs/5625962][chore] unwaive DS-v32-fp4 tests (#8853) Signed-off-by: Fanrong Li <23290157+lfr-0531@users.noreply.github.com> --- tests/integration/defs/accuracy/references/mmlu.yaml | 5 +++++ tests/integration/test_lists/waives.txt | 2 -- 2 files changed, 5 insertions(+), 2 deletions(-) diff --git a/tests/integration/defs/accuracy/references/mmlu.yaml b/tests/integration/defs/accuracy/references/mmlu.yaml index efe0850e0a..9fe9a7404c 100644 --- a/tests/integration/defs/accuracy/references/mmlu.yaml +++ b/tests/integration/defs/accuracy/references/mmlu.yaml @@ -198,6 +198,11 @@ deepseek-ai/DeepSeek-V3.2-Exp: - quant_algo: FP8_BLOCK_SCALES spec_dec_algo: MTP accuracy: 88.2 + - quant_algo: NVFP4 + accuracy: 87.81 + - quant_algo: NVFP4 + spec_dec_algo: MTP + accuracy: 87.81 Qwen3/Qwen3-8B: - quant_algo: W4A8_MXFP4_FP8 accuracy: 72.70 diff --git a/tests/integration/test_lists/waives.txt b/tests/integration/test_lists/waives.txt index 1825cf07fa..2fc3a0b191 100644 --- a/tests/integration/test_lists/waives.txt +++ b/tests/integration/test_lists/waives.txt @@ -394,5 +394,3 @@ accuracy/test_disaggregated_serving.py::TestQwen3_30B_A3B::test_mixed_ctx_gen_mo accuracy/test_disaggregated_serving.py::TestGPTOSS::test_auto_dtype[True] SKIP (https://nvbugs/5624367) accuracy/test_disaggregated_serving.py::TestGPTOSS::test_auto_dtype[False] SKIP (https://nvbugs/5624367) disaggregated/test_disaggregated.py::test_disaggregated_benchmark_on_diff_backends[llama-v3-8b-hf] SKIP (https://nvbugs/5587574) -accuracy/test_llm_api_pytorch.py::TestDeepSeekV32::test_nvfp4_multi_gpus[baseline] SKIP (https://nvbugs/5625962) -accuracy/test_llm_api_pytorch.py::TestDeepSeekV32::test_nvfp4_multi_gpus[baseline_mtp1] SKIP (https://nvbugs/5625962)