A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 03:30 UTC
[Bugfix] Fix deepep tests (#20288)
Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com> Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Varun Sundar Rabindranath committed
Jul 1, 2025 at 03:29 UTC
08d81f1014d174d4dd96518914c7ed9767c67a3f
2 files changed
+2
-2
tests/kernels/moe/test_deepep_deepgemm_moe.py
+1
-1
@@ -30,7 +30,7 @@ if has_deep_ep():
30
from vllm.model_executor.layers.fused_moe.deepep_ll_prepare_finalize import ( # noqa: E501
31
DeepEPLLPrepareAndFinalize)
32
33
- from .deepep_utils import DeepEPHTArgs, DeepEPLLArgs, make_deepep_a2a
33
+ from .utils import DeepEPHTArgs, DeepEPLLArgs, make_deepep_a2a
34
35
if has_deep_gemm():
36
import deep_gemm
tests/kernels/moe/test_deepep_moe.py
+1
-1
@@ -31,7 +31,7 @@ if has_deep_ep():
31
from vllm.model_executor.layers.fused_moe.deepep_ll_prepare_finalize import ( # noqa: E501
32
DeepEPLLPrepareAndFinalize)
33
34
- from .deepep_utils import DeepEPHTArgs, DeepEPLLArgs, make_deepep_a2a
34
+ from .utils import DeepEPHTArgs, DeepEPLLArgs, make_deepep_a2a
35
36
requires_deep_ep = pytest.mark.skipif(
37
not has_deep_ep(),