[CI] Refresh hybrid Model Runner V2 coverage (#51410)
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
This commit is contained in:
@@ -108,6 +108,7 @@ steps:
|
||||
- vllm/v1/worker/gpu_worker.py
|
||||
- tests/v1/spec_decode/test_max_len.py
|
||||
- tests/v1/spec_decode/test_rejection_sampler_utils.py
|
||||
- tests/v1/spec_decode/test_synthetic_rejection_sampler_utils.py
|
||||
commands:
|
||||
- >-
|
||||
bash .buildkite/scripts/hardware_ci/run-intel-test.sh
|
||||
|
||||
@@ -1851,7 +1851,7 @@ steps:
|
||||
commands:
|
||||
- set -x
|
||||
- export VLLM_USE_V2_MODEL_RUNNER=1
|
||||
- pytest -v -s distributed/test_pipeline_parallel.py -k "not ray and not Jamba"
|
||||
- pytest -v -s distributed/test_pipeline_parallel.py -k "not ray"
|
||||
- pytest -v -s distributed/test_pp_cudagraph.py -k "not ray"
|
||||
|
||||
- label: Model Runner V2 Spec Decode # TBD
|
||||
|
||||
@@ -96,7 +96,7 @@ steps:
|
||||
commands:
|
||||
- set -x
|
||||
- export VLLM_USE_V2_MODEL_RUNNER=1
|
||||
- pytest -v -s distributed/test_pipeline_parallel.py -k "not ray and not Jamba"
|
||||
- pytest -v -s distributed/test_pipeline_parallel.py -k "not ray"
|
||||
- pytest -v -s distributed/test_pp_cudagraph.py -k "not ray"
|
||||
- pytest -v -s v1/distributed/test_pp_dp_v2.py
|
||||
|
||||
@@ -110,6 +110,7 @@ steps:
|
||||
- vllm/v1/worker/gpu_worker.py
|
||||
- tests/v1/spec_decode/test_max_len.py
|
||||
- tests/v1/spec_decode/test_rejection_sampler_utils.py
|
||||
- tests/v1/spec_decode/test_synthetic_rejection_sampler_utils.py
|
||||
- tests/v1/e2e/spec_decode/
|
||||
commands:
|
||||
- set -x
|
||||
@@ -119,3 +120,6 @@ steps:
|
||||
- pytest -v -s v1/spec_decode/test_synthetic_rejection_sampler_utils.py
|
||||
- pytest -v -s v1/e2e/spec_decode/eagle/
|
||||
- pytest -v -s v1/e2e/spec_decode/speculators/
|
||||
- >-
|
||||
pytest -v -s
|
||||
'v1/e2e/spec_decode/mtp/test_mtp.py::test_mtp_correctness[qwen3_5-hybrid]'
|
||||
|
||||
@@ -13,13 +13,7 @@ steps:
|
||||
- vllm/model_executor/layers/quantization
|
||||
- tests/quantization
|
||||
commands:
|
||||
# temporary install here since we need nightly, will move to requirements/test/cuda.in
|
||||
# after torchao 0.12 release, and pin a working version of torchao nightly here
|
||||
|
||||
# since torchao nightly is only compatible with torch nightly currently
|
||||
# https://github.com/pytorch/ao/issues/2919, we'll have to skip new torchao tests for now
|
||||
# we can only upgrade after this is resolved
|
||||
# TODO(jerryzh168): resolve the above comment
|
||||
# Pin torchao to the version compatible with the CI PyTorch and CUDA stack.
|
||||
- uv pip install --system torchao==0.17.0 --index-url https://download.pytorch.org/whl/cu130
|
||||
- uv pip install --system conch-triton-kernels
|
||||
# The SM90-only checkpoint currently contains a removed weight_chan_scale
|
||||
|
||||
@@ -1,7 +1,6 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
import os
|
||||
from typing import Any
|
||||
|
||||
import pytest
|
||||
@@ -72,12 +71,6 @@ def test_mtp_correctness(
|
||||
draft_model = None
|
||||
_skip_if_insufficient_gpus_for_tp(tp_size)
|
||||
|
||||
if "Qwen3.5" in model_name and os.environ.get("VLLM_USE_V2_MODEL_RUNNER"):
|
||||
pytest.skip(
|
||||
"Model Runner V2 does not yet support hybrid models "
|
||||
"(Qwen3.5 mixes Mamba-style GDN with attention layers)."
|
||||
)
|
||||
|
||||
attn_backend = "TRITON_ATTN" if current_platform.is_rocm() else "auto"
|
||||
|
||||
# Skip multimodal profiling for models that don't need it in this test.
|
||||
|
||||
Reference in New Issue
Block a user