[Bugfix][CI/CD][CPU] Fix CPU CI tests (#20383)

Signed-off-by: jiang1.li <jiang1.li@intel.com>

[Bugfix][CI/CD][CPU] Fix CPU CI tests (#20383)
Signed-off-by: jiang1.li <jiang1.li@intel.com>
0ec3779d · Li, Jiang · GitHub · b616f6a5 · 0ec3779d
Unverified Commit 0ec3779d authored Jul 03, 2025 by Li, Jiang Committed by GitHub Jul 02, 2025
Hide whitespace changes
Inline Side-by-side

Showing with 7 additions and 6 deletions

vllm/model_executor/layers/quantization/utils/fp8_utils.py vllm/model_executor/layers/quantization/utils/fp8_utils.py +7 -6

No files found.
--- a/vllm/model_executor/layers/quantization/utils/fp8_utils.py
+++ b/vllm/model_executor/layers/quantization/utils/fp8_utils.py
@@ -201,12 +201,13 @@ def apply_w8a8_block_fp8_linear_fake(
    return torch.empty(output_shape, dtype=input.dtype, device=input.device)


-direct_register_custom_op(
-    op_name="apply_w8a8_block_fp8_linear",
-    op_func=apply_w8a8_block_fp8_linear,
-    mutates_args=[],
-    fake_impl=apply_w8a8_block_fp8_linear_fake,
-)
+if not current_platform.is_cpu():
+    direct_register_custom_op(
+        op_name="apply_w8a8_block_fp8_linear",
+        op_func=apply_w8a8_block_fp8_linear,
+        mutates_args=[],
+        fake_impl=apply_w8a8_block_fp8_linear_fake,
+    )


 def input_to_float8(