From 772ce5af9745393d90408eaa1b8b090014ab551d Mon Sep 17 00:00:00 2001 From: Ming Yang Date: Wed, 23 Jul 2025 20:22:42 -0700 Subject: [PATCH] [Misc] Add dummy maverick test to CI (#21324) Signed-off-by: Ming Yang Co-authored-by: Cyrus Leung --- .buildkite/test-pipeline.yaml | 1 + tests/models/multimodal/generation/test_maverick.py | 3 +++ 2 files changed, 4 insertions(+) diff --git a/.buildkite/test-pipeline.yaml b/.buildkite/test-pipeline.yaml index c7378bf8ba..c2e56557ba 100644 --- a/.buildkite/test-pipeline.yaml +++ b/.buildkite/test-pipeline.yaml @@ -718,6 +718,7 @@ steps: - VLLM_USE_V1=0 CUDA_VISIBLE_DEVICES=0,1 pytest -v -s test_sharded_state_loader.py - VLLM_USE_V1=0 CUDA_VISIBLE_DEVICES=0,1 pytest -v -s kv_transfer/test_disagg.py - CUDA_VISIBLE_DEVICES=0,1 pytest -v -s v1/shutdown + - pytest -v -s models/multimodal/generation/test_maverick.py - label: Plugin Tests (2 GPUs) # 40min mirror_hardwares: [amdexperimental] diff --git a/tests/models/multimodal/generation/test_maverick.py b/tests/models/multimodal/generation/test_maverick.py index 083dc66148..306cf39002 100644 --- a/tests/models/multimodal/generation/test_maverick.py +++ b/tests/models/multimodal/generation/test_maverick.py @@ -23,6 +23,8 @@ from transformers import (AutoConfig, AutoProcessor, AutoTokenizer, from vllm import LLM, SamplingParams +from ....utils import multi_gpu_test + # Sample prompts for testing PROMPTS: list[str] = [ "Hello, my name is", @@ -541,6 +543,7 @@ def run_reduced_model(model_path: str, print("-" * 40) +@multi_gpu_test(num_gpus=2) @pytest.mark.parametrize( "original_model_name,text_layers,num_experts,vision_layers,", [("meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", 4, 4, 2)])