anodyine/vllm-omni
1
0 Fork 0
Code Issues 4 Pull Requests 2 Actions Packages Projects Releases Wiki Activity

[CI/Build] Restrict ERNIE fused RoPE tests to NVIDIA CUDA (#7771)

Browse Source 
Signed-off-by: andyluo7 <andy.luo@amd.com>
This commit is contained in:
andyluo7 authored and GitHub committed 2026-09-19 13:23:50 +08:00
1 parent 4c7a98c26f
commit 232dbc82f8
1 file changed
+14 -6
Download Patch File Download Diff File Expand all files Collapse all files
tests/diffusion/models/ernie_image/test_ernie_image_fused_rope.py
+14 -6
View File
@@ -5,9 +5,11 @@ import pytest
import torch
from vllm.triton_utils import HAS_TRITON
from tests.helpers.mark import hardware_test
from vllm_omni.diffusion.models.ernie_image.ernie_image_transformer import _apply_rotary_emb
from vllm_omni.platforms import current_omni_platform
pytestmark = [pytest.mark.core_model, pytest.mark.cuda, pytest.mark.diffusion]
pytestmark = [pytest.mark.core_model, pytest.mark.diffusion]
@pytest.fixture(autouse=True)
@@ -35,7 +37,8 @@ def _inputs(shape: tuple[int, int, int, int]):
return query, key, freqs_cos, freqs_sin
@pytest.mark.skipif(not torch.cuda.is_available(), reason="CUDA required")
@hardware_test(res={"cuda": "L4"}, num_cards=1)
@pytest.mark.skipif(not current_omni_platform.is_cuda(), reason="NVIDIA CUDA required")
@pytest.mark.skipif(not HAS_TRITON, reason="Triton required")
@pytest.mark.parametrize(
"shape",
@@ -60,7 +63,8 @@ def test_fused_qk_rope_is_bit_exact(shape):
assert torch.equal(actual_key, expected_key)
@pytest.mark.skipif(not torch.cuda.is_available(), reason="CUDA required")
@hardware_test(res={"cuda": "L4"}, num_cards=1)
@pytest.mark.skipif(not current_omni_platform.is_cuda(), reason="NVIDIA CUDA required")
@pytest.mark.skipif(not HAS_TRITON, reason="Triton required")
def test_launch_failure_disables_runtime_key(monkeypatch):
from vllm_omni.diffusion.models.ernie_image import fused_rope
@@ -85,6 +89,7 @@ def test_launch_failure_disables_runtime_key(monkeypatch):
assert len(fused_rope._FAILED_KEYS) == 1
@hardware_test(res={"cuda": "L4", "rocm": "MI325"}, num_cards=1)
def test_failed_runtime_key_cache_is_bounded():
from vllm_omni.diffusion.models.ernie_image import fused_rope
@@ -96,7 +101,8 @@ def test_failed_runtime_key_cache_is_bounded():
assert (fused_rope._FAILED_KEYS_MAX_SIZE,) in fused_rope._FAILED_KEYS
@pytest.mark.skipif(not torch.cuda.is_available(), reason="CUDA required")
@hardware_test(res={"cuda": "L4"}, num_cards=1)
@pytest.mark.skipif(not current_omni_platform.is_cuda(), reason="NVIDIA CUDA required")
@pytest.mark.skipif(not HAS_TRITON, reason="Triton required")
def test_compile_path_does_not_launch_fused_kernel(monkeypatch):
from vllm_omni.diffusion.models.ernie_image import fused_rope
@@ -116,7 +122,8 @@ def test_compile_path_does_not_launch_fused_kernel(monkeypatch):
assert fused_rope.try_fused_qk_rotary_emb(query, key, freqs_cos, freqs_sin) is None
@pytest.mark.skipif(not torch.cuda.is_available(), reason="CUDA required")
@hardware_test(res={"cuda": "L4"}, num_cards=1)
@pytest.mark.skipif(not current_omni_platform.is_cuda(), reason="NVIDIA CUDA required")
@pytest.mark.skipif(not HAS_TRITON, reason="Triton required")
@pytest.mark.parametrize("requires_grad_input", ["query", "key", "cos", "sin"])
def test_gradient_inputs_do_not_launch_fused_kernel(monkeypatch, requires_grad_input):
@@ -142,7 +149,8 @@ def test_gradient_inputs_do_not_launch_fused_kernel(monkeypatch, requires_grad_i
)
@pytest.mark.skipif(not torch.cuda.is_available(), reason="CUDA required")
@hardware_test(res={"cuda": "L4"}, num_cards=1)
@pytest.mark.skipif(not current_omni_platform.is_cuda(), reason="NVIDIA CUDA required")
@pytest.mark.skipif(not HAS_TRITON, reason="Triton required")
def test_mismatched_key_dtype_or_device_does_not_launch(monkeypatch):
from vllm_omni.diffusion.models.ernie_image import fused_rope