[Feature] Control the stage init timeout threshold by --stage-init-timeout (#393)

Signed-off-by: Taichang Zhou <tzhouam@connect.ust.hk>
Signed-off-by: Zhou Taichang <tzhouam@connect.ust.hk>
Signed-off-by: tzhouam <tzhouam@connect.ust.hk>
Signed-off-by: gcanlin <canlinguosdu@gmail.com>
Signed-off-by: yinpeiqi <yinpeiqi809@gmail.com>
Signed-off-by: Li Shunyang <gmsv2544@nus.edu.sg>
Signed-off-by: Li Shunyang <lishunyang12@163.com>
Signed-off-by: Peiqi Yin <60515999+yinpeiqi@users.noreply.github.com>
Signed-off-by: dengyunyang <584797741@qq.com>
Signed-off-by: Samit <285365963@qq.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Canlin Guo <canlinguosdu@gmail.com>
Co-authored-by: Peiqi Yin <60515999+yinpeiqi@users.noreply.github.com>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>
Co-authored-by: SYLAR <125541396+lishunyang12@users.noreply.github.com>
Co-authored-by: Jiangyun Zhu <riverclouds.zhu@qq.com>
Co-authored-by: dengyunyang <584797741@qq.com>
Co-authored-by: Samit <285365963@qq.com>
This commit is contained in:
authored and GitHub committed 2025-12-30 23:37:10 +08:00
1 parent 51c0303945
commit f0b8d3db82
11 files changed
+64 -58

No files matched your search

+18 -23
View File
@@ -5,9 +5,6 @@ from typing import Any
from unittest.mock import MagicMock
import pytest
from vllm.sampling_params import SamplingParams
from vllm_omni.entrypoints.stage_utils import _to_dict
# Suppress noisy DeprecationWarnings from optional Swig bindings imported by vLLM dependencies.
warnings.filterwarnings(
@@ -71,7 +68,7 @@ class _FakeQueue:
class _FakeStage:
"""Lightweight Stage stub for multi-process pipeline version with queue support."""
def __init__(self, config):
def __init__(self, config, stage_init_timeout: int = 300):
# Handle both dict and object configs
if isinstance(config, dict):
config = _FakeStageConfig(config)
@@ -95,9 +92,7 @@ class _FakeStage:
self._in_q = None
self._out_q = None
self._proc = None # Mock process reference
default_sampling_params = getattr(config, "default_sampling_params", {})
self.default_sampling_params = SamplingParams(**_to_dict(default_sampling_params))
self._stage_init_timeout = max(0, int(stage_init_timeout))
def attach_queues(self, in_q, out_q):
"""Attach input and output queues."""
@@ -481,7 +476,7 @@ def test_initialize_stage_configs_called_when_none(monkeypatch, fake_stage_confi
# Replace OmniStage
monkeypatch.setattr(
"vllm_omni.entrypoints.omni_stage.OmniStage",
lambda cfg: _FakeStage(cfg),
lambda cfg, **kwargs: _FakeStage(cfg, **kwargs),
raising=False,
)
@@ -490,7 +485,7 @@ def test_initialize_stage_configs_called_when_none(monkeypatch, fake_stage_confi
# Patch the imported function and class in the module
monkeypatch.setattr(omni_module, "load_stage_configs_from_model", _fake_loader)
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg: _FakeStage(cfg))
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg, **kwargs: _FakeStage(cfg, **kwargs))
from vllm_omni.entrypoints.omni import Omni
@@ -543,14 +538,14 @@ def test_generate_raises_on_length_mismatch(monkeypatch, fake_stage_config):
)
monkeypatch.setattr(
"vllm_omni.entrypoints.omni_stage.OmniStage",
lambda cfg: _FakeStage(cfg),
lambda cfg, **kwargs: _FakeStage(cfg, **kwargs),
raising=False,
)
import vllm_omni.entrypoints.omni as omni_module
monkeypatch.setattr(omni_module, "load_stage_configs_from_model", _fake_loader)
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg: _FakeStage(cfg))
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg, **kwargs: _FakeStage(cfg, **kwargs))
from vllm_omni.entrypoints.omni import Omni
@@ -595,14 +590,14 @@ def test_generate_pipeline_and_final_outputs(monkeypatch, fake_stage_config):
)
monkeypatch.setattr(
"vllm_omni.entrypoints.omni_stage.OmniStage",
lambda cfg: _FakeStage(cfg),
lambda cfg, **kwargs: _FakeStage(cfg, **kwargs),
raising=False,
)
import vllm_omni.entrypoints.omni as omni_module
monkeypatch.setattr(omni_module, "load_stage_configs_from_model", _fake_loader)
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg: _FakeStage(cfg))
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg, **kwargs: _FakeStage(cfg, **kwargs))
# Mock uuid.uuid4() to return a predictable value for request ID generation
test_uuid = uuid.UUID("00000000-0000-0000-0000-000000000000")
@@ -688,14 +683,14 @@ def test_generate_no_final_output_returns_empty(monkeypatch, fake_stage_config):
)
monkeypatch.setattr(
"vllm_omni.entrypoints.omni_stage.OmniStage",
lambda cfg: _FakeStage(cfg),
lambda cfg, **kwargs: _FakeStage(cfg, **kwargs),
raising=False,
)
import vllm_omni.entrypoints.omni as omni_module
monkeypatch.setattr(omni_module, "load_stage_configs_from_model", _fake_loader)
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg: _FakeStage(cfg))
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg, **kwargs: _FakeStage(cfg, **kwargs))
# Mock uuid.uuid4() to return a predictable value for request ID generation
test_uuid = uuid.UUID("00000000-0000-0000-0000-000000000000")
@@ -762,14 +757,14 @@ def test_generate_sampling_params_none_use_default(monkeypatch, fake_stage_confi
)
monkeypatch.setattr(
"vllm_omni.entrypoints.omni_stage.OmniStage",
lambda cfg: _FakeStage(cfg),
lambda cfg, **kwargs: _FakeStage(cfg, **kwargs),
raising=False,
)
import vllm_omni.entrypoints.omni as omni_module
monkeypatch.setattr(omni_module, "load_stage_configs_from_model", _fake_loader)
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg: _FakeStage(cfg))
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg, **kwargs: _FakeStage(cfg, **kwargs))
# Mock uuid.uuid4() to return a predictable value for request ID generation
test_uuid = uuid.UUID("00000000-0000-0000-0000-000000000000")
@@ -841,14 +836,14 @@ def test_wait_for_stages_ready_timeout(monkeypatch, fake_stage_config):
monkeypatch.setattr(
"vllm_omni.entrypoints.omni_stage.OmniStage",
lambda cfg: _FakeStageNoReady(cfg),
lambda cfg, **kwargs: _FakeStageNoReady(cfg, **kwargs),
raising=False,
)
import vllm_omni.entrypoints.omni as omni_module
monkeypatch.setattr(omni_module, "load_stage_configs_from_model", _fake_loader)
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg: _FakeStageNoReady(cfg))
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg, **kwargs: _FakeStageNoReady(cfg, **kwargs))
from vllm_omni.entrypoints.omni import Omni
@@ -886,14 +881,14 @@ def test_generate_handles_error_messages(monkeypatch, fake_stage_config):
)
monkeypatch.setattr(
"vllm_omni.entrypoints.omni_stage.OmniStage",
lambda cfg: _FakeStage(cfg),
lambda cfg, **kwargs: _FakeStage(cfg, **kwargs),
raising=False,
)
import vllm_omni.entrypoints.omni as omni_module
monkeypatch.setattr(omni_module, "load_stage_configs_from_model", _fake_loader)
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg: _FakeStage(cfg))
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg, **kwargs: _FakeStage(cfg, **kwargs))
# Mock uuid.uuid4() to return a predictable value for request ID generation
test_uuid = uuid.UUID("00000000-0000-0000-0000-000000000000")
@@ -962,14 +957,14 @@ def test_close_sends_shutdown_signal(monkeypatch, fake_stage_config):
)
monkeypatch.setattr(
"vllm_omni.entrypoints.omni_stage.OmniStage",
lambda cfg: _FakeStage(cfg),
lambda cfg, **kwargs: _FakeStage(cfg, **kwargs),
raising=False,
)
import vllm_omni.entrypoints.omni as omni_module
monkeypatch.setattr(omni_module, "load_stage_configs_from_model", _fake_loader)
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg: _FakeStage(cfg))
monkeypatch.setattr(omni_module, "OmniStage", lambda cfg, **kwargs: _FakeStage(cfg, **kwargs))
from vllm_omni.entrypoints.omni import Omni