From 5d62fcd7171f1d0588045fcf77ea6e8d4c8ecedf Mon Sep 17 00:00:00 2001 From: zhumingjue138 Date: Tue, 10 Mar 2026 17:39:37 +0800 Subject: [PATCH] [CI]: Add core_model and cpu markers for L1 use case. (#1709) Signed-off-by: zhumingjue --- docs/contributing/ci/CI_5levels.md | 4 ++-- tests/diffusion/distributed/test_hsdp.py | 2 +- tests/diffusion/distributed/test_parallel_state_sp_groups.py | 5 +---- tests/diffusion/distributed/test_sp_plan_hooks.py | 2 ++ tests/diffusion/distributed/test_vae_patch_parallel.py | 2 ++ .../models/nextstep_1_1/test_nextstep_cfg_parallel_layout.py | 2 ++ tests/diffusion/models/z_image/test_zimage_tp_constraints.py | 2 ++ tests/diffusion/test_multiproc_executor_concurrency.py | 2 +- .../omni_connectors/test_mooncake_transfer_engine_buffer.py | 2 +- .../omni_connectors/test_omni_connector_configs.py | 2 ++ .../models/qwen2_5_omni/test_qwen2_5_omni_embed.py | 2 ++ 11 files changed, 18 insertions(+), 9 deletions(-) diff --git a/docs/contributing/ci/CI_5levels.md b/docs/contributing/ci/CI_5levels.md index dbf87d352..c8493af0a 100644 --- a/docs/contributing/ci/CI_5levels.md +++ b/docs/contributing/ci/CI_5levels.md @@ -67,8 +67,8 @@ Through five levels (L1-L5) and common (Common) specifications, the system clari Online & Offline (multiple deployment scenarios):
real model, normal inference function, normal accuracy <30min - /tests/e2e/online_serving/test_{model_name}_expansion.py
- /tests/e2e/offline_inference/test_{model_name}_expansion.py + /tests/e2e/online_serving/test_{model_name}.py
+ /tests/e2e/offline_inference/test_{model_name}.py Chapter 2
diff --git a/tests/diffusion/distributed/test_hsdp.py b/tests/diffusion/distributed/test_hsdp.py index 6fcaf1e81..e6d782c0a 100644 --- a/tests/diffusion/distributed/test_hsdp.py +++ b/tests/diffusion/distributed/test_hsdp.py @@ -11,7 +11,7 @@ import torch.nn as nn from vllm_omni.diffusion.data import DiffusionParallelConfig from vllm_omni.diffusion.distributed.hsdp import HSDPInferenceConfig -pytestmark = [pytest.mark.diffusion, pytest.mark.parallel, pytest.mark.cpu] +pytestmark = [pytest.mark.diffusion, pytest.mark.parallel, pytest.mark.cpu, pytest.mark.core_model] class TestHSDPInferenceConfig: diff --git a/tests/diffusion/distributed/test_parallel_state_sp_groups.py b/tests/diffusion/distributed/test_parallel_state_sp_groups.py index ac8b1700a..9dda5610b 100644 --- a/tests/diffusion/distributed/test_parallel_state_sp_groups.py +++ b/tests/diffusion/distributed/test_parallel_state_sp_groups.py @@ -11,10 +11,7 @@ import torch from vllm_omni.diffusion.distributed.parallel_state import RankGenerator, set_seq_parallel_pg -pytestmark = [ - pytest.mark.diffusion, - pytest.mark.parallel, -] +pytestmark = [pytest.mark.diffusion, pytest.mark.parallel, pytest.mark.core_model, pytest.mark.cpu] def _fake_new_group_factory(created_groups: list[SimpleNamespace]): diff --git a/tests/diffusion/distributed/test_sp_plan_hooks.py b/tests/diffusion/distributed/test_sp_plan_hooks.py index 6883ae09f..72ec9aaf2 100644 --- a/tests/diffusion/distributed/test_sp_plan_hooks.py +++ b/tests/diffusion/distributed/test_sp_plan_hooks.py @@ -47,6 +47,8 @@ requires_distributed = pytest.mark.skipif( pytestmark = [ pytest.mark.diffusion, pytest.mark.parallel, + pytest.mark.core_model, + pytest.mark.cpu, ] # ============================================================================= diff --git a/tests/diffusion/distributed/test_vae_patch_parallel.py b/tests/diffusion/distributed/test_vae_patch_parallel.py index 43bbc71d5..db74fff1b 100644 --- a/tests/diffusion/distributed/test_vae_patch_parallel.py +++ b/tests/diffusion/distributed/test_vae_patch_parallel.py @@ -8,6 +8,8 @@ import torch from vllm_omni.diffusion.distributed import vae_patch_parallel as vae_patch_parallel +pytestmark = [pytest.mark.core_model, pytest.mark.cpu] + class _DummyConfig: def __init__(self, **attrs): diff --git a/tests/diffusion/models/nextstep_1_1/test_nextstep_cfg_parallel_layout.py b/tests/diffusion/models/nextstep_1_1/test_nextstep_cfg_parallel_layout.py index 9caf667be..dd87d2249 100644 --- a/tests/diffusion/models/nextstep_1_1/test_nextstep_cfg_parallel_layout.py +++ b/tests/diffusion/models/nextstep_1_1/test_nextstep_cfg_parallel_layout.py @@ -8,6 +8,8 @@ import vllm_omni.diffusion.models.nextstep_1_1.pipeline_nextstep_1_1 as nextstep from vllm_omni.diffusion.models.nextstep_1_1.modeling_nextstep_heads import FlowMatchingHead from vllm_omni.diffusion.models.nextstep_1_1.pipeline_nextstep_1_1 import NextStep11Pipeline +pytestmark = [pytest.mark.core_model, pytest.mark.cpu] + class _DummyImageHead: def __init__(self, token_dim: int): diff --git a/tests/diffusion/models/z_image/test_zimage_tp_constraints.py b/tests/diffusion/models/z_image/test_zimage_tp_constraints.py index a276274a0..b87d51c36 100644 --- a/tests/diffusion/models/z_image/test_zimage_tp_constraints.py +++ b/tests/diffusion/models/z_image/test_zimage_tp_constraints.py @@ -2,6 +2,8 @@ import pytest from vllm_omni.diffusion.models.z_image.z_image_transformer import validate_zimage_tp_constraints +pytestmark = [pytest.mark.core_model, pytest.mark.cpu] + def test_validate_zimage_tp_constraints_tp2_ok(): ffn_hidden_dim, final_out_dims, supported_tp = validate_zimage_tp_constraints( diff --git a/tests/diffusion/test_multiproc_executor_concurrency.py b/tests/diffusion/test_multiproc_executor_concurrency.py index ee3b8f4b8..76caa9844 100644 --- a/tests/diffusion/test_multiproc_executor_concurrency.py +++ b/tests/diffusion/test_multiproc_executor_concurrency.py @@ -12,7 +12,7 @@ from vllm_omni.diffusion.data import DiffusionOutput from vllm_omni.diffusion.executor.multiproc_executor import MultiprocDiffusionExecutor from vllm_omni.diffusion.scheduler import Scheduler -pytestmark = [pytest.mark.diffusion] +pytestmark = [pytest.mark.diffusion, pytest.mark.core_model, pytest.mark.cpu] # ───────────────────────────────────────────── helpers ───────────────────── diff --git a/tests/distributed/omni_connectors/test_mooncake_transfer_engine_buffer.py b/tests/distributed/omni_connectors/test_mooncake_transfer_engine_buffer.py index 968109976..55351efcb 100755 --- a/tests/distributed/omni_connectors/test_mooncake_transfer_engine_buffer.py +++ b/tests/distributed/omni_connectors/test_mooncake_transfer_engine_buffer.py @@ -17,7 +17,7 @@ from vllm_omni.distributed.omni_connectors.connectors.mooncake_transfer_engine_c ) # All tests in this file are pure-CPU unit tests for the memory allocator. -pytestmark = [pytest.mark.cpu, pytest.mark.parallel] +pytestmark = [pytest.mark.cpu, pytest.mark.parallel, pytest.mark.core_model] @pytest.mark.core_model diff --git a/tests/distributed/omni_connectors/test_omni_connector_configs.py b/tests/distributed/omni_connectors/test_omni_connector_configs.py index ca73e1ec4..3dc80b57b 100644 --- a/tests/distributed/omni_connectors/test_omni_connector_configs.py +++ b/tests/distributed/omni_connectors/test_omni_connector_configs.py @@ -8,6 +8,8 @@ import pytest # Use the new import path for initialization utilities from vllm_omni.distributed.omni_connectors.utils.initialization import load_omni_transfer_config +pytestmark = [pytest.mark.core_model, pytest.mark.cpu] + def get_config_files(): """Helper to find config files.""" diff --git a/tests/model_executor/models/qwen2_5_omni/test_qwen2_5_omni_embed.py b/tests/model_executor/models/qwen2_5_omni/test_qwen2_5_omni_embed.py index 9bf230d55..84112336e 100644 --- a/tests/model_executor/models/qwen2_5_omni/test_qwen2_5_omni_embed.py +++ b/tests/model_executor/models/qwen2_5_omni/test_qwen2_5_omni_embed.py @@ -19,6 +19,8 @@ from vllm.model_executor.models.qwen2_5_omni_thinker import ( merge_interleaved_embeddings, ) +pytestmark = [pytest.mark.core_model, pytest.mark.cpu] + # Fake token IDs AUDIO_TOKEN_ID = 1001 IMAGE_TOKEN_ID = 1002