From 188433c7a75b3d689fdefc48bb48d77abdef07cc Mon Sep 17 00:00:00 2001 From: Helen Ngo Date: Fri, 31 Oct 2025 13:02:20 -0700 Subject: [PATCH 1/2] Do not test MoE throughput yet --- .../python_test_utils/test_inference_regular_pipeline.py | 3 --- .../golden_values_dev_dgx_h100.json | 3 +-- .../model_config.yaml | 2 +- 3 files changed, 2 insertions(+), 6 deletions(-) diff --git a/tests/functional_tests/python_test_utils/test_inference_regular_pipeline.py b/tests/functional_tests/python_test_utils/test_inference_regular_pipeline.py index b9c70046a4d..bbbb924dbfd 100644 --- a/tests/functional_tests/python_test_utils/test_inference_regular_pipeline.py +++ b/tests/functional_tests/python_test_utils/test_inference_regular_pipeline.py @@ -24,9 +24,6 @@ def test_inference_pipeline(golden_values_path: str, test_values_path: str) -> N # Handle JSONL output, assume only one line in this case. output_current = json.loads(output_current) - assert set(output_groundtruth.keys()).issuperset( - set(output_current.keys()) - ), f"Some IDs from groundtruth are missing in current: {output_groundtruth.keys()} vs {output_current.keys()}" if set(output_groundtruth.keys()) != set(output_current.keys()): logger.warning( f"Some IDs from groundtruth are missing in output, only the subset of ids in groundtruth will be tested: {output_groundtruth.keys()} vs {output_current.keys()}" diff --git a/tests/functional_tests/test_cases/moe/gpt_dynamic_inference_tp4_pp1_ep4_16B_logitsmatch/golden_values_dev_dgx_h100.json b/tests/functional_tests/test_cases/moe/gpt_dynamic_inference_tp4_pp1_ep4_16B_logitsmatch/golden_values_dev_dgx_h100.json index 711eeddfb25..aaa789b3696 100644 --- a/tests/functional_tests/test_cases/moe/gpt_dynamic_inference_tp4_pp1_ep4_16B_logitsmatch/golden_values_dev_dgx_h100.json +++ b/tests/functional_tests/test_cases/moe/gpt_dynamic_inference_tp4_pp1_ep4_16B_logitsmatch/golden_values_dev_dgx_h100.json @@ -156,6 +156,5 @@ -1.2342015504837036, -0.5394397377967834 ] - }, - "throughput": 6.543502517233578 + } } \ No newline at end of file diff --git a/tests/functional_tests/test_cases/moe/gpt_dynamic_inference_tp4_pp1_ep4_16B_logitsmatch/model_config.yaml b/tests/functional_tests/test_cases/moe/gpt_dynamic_inference_tp4_pp1_ep4_16B_logitsmatch/model_config.yaml index 1b9eaaf1f65..669ff9532ad 100644 --- a/tests/functional_tests/test_cases/moe/gpt_dynamic_inference_tp4_pp1_ep4_16B_logitsmatch/model_config.yaml +++ b/tests/functional_tests/test_cases/moe/gpt_dynamic_inference_tp4_pp1_ep4_16B_logitsmatch/model_config.yaml @@ -75,7 +75,7 @@ MODEL_ARGS: --output-path: ${TENSORBOARD_PATH} --prompts: "Time travel to 2008, and go to a bar or a club or one of the myriad disco-basements on the Lower East Side that does not quite know which of those it is. Dance awkwardly in a room full of other glittered-up nerds, and wait for something to happen, buoyed on the feeling that this is the big swollen heart of life, that this is New York like the movies." --incoming-requests-per-sec: -1 # all requests arrive up front. - --inference-repeat-n: 8 + --inference-repeat-n: 1 METRICS: - "generated_tokens" - "logprobs" From 3852a72e37a5d816dc9b40fc8a6174b88699e24a Mon Sep 17 00:00:00 2001 From: Helen Ngo Date: Fri, 31 Oct 2025 13:26:39 -0700 Subject: [PATCH 2/2] Copyright header --- .../python_test_utils/test_inference_regular_pipeline.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/tests/functional_tests/python_test_utils/test_inference_regular_pipeline.py b/tests/functional_tests/python_test_utils/test_inference_regular_pipeline.py index bbbb924dbfd..1df21273ee5 100644 --- a/tests/functional_tests/python_test_utils/test_inference_regular_pipeline.py +++ b/tests/functional_tests/python_test_utils/test_inference_regular_pipeline.py @@ -1,3 +1,5 @@ +# Copyright (c) 2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved. + import json import logging import math