[Log] Reduce duplicate log (#37313)

Signed-off-by: yewentao256 <zhyanwentao@126.com>
This commit is contained in:
Wentao Ye
2026-03-18 10:57:44 -04:00
committed by GitHub
parent de1a86b7de
commit c373b5c00d
8 changed files with 20 additions and 10 deletions
@@ -227,7 +227,9 @@ class MMEncoderAttention(CustomOp):
if self.attn_backend == AttentionBackendEnum.FLASHINFER:
_get_flashinfer_workspace_buffer()
logger.info_once(f"Using {self.attn_backend} for MMEncoderAttention.")
logger.info_once(
f"Using {self.attn_backend} for MMEncoderAttention.", scope="local"
)
@classmethod
def enabled(cls) -> bool:
+4 -3
View File
@@ -192,14 +192,15 @@ class ChunkGatedDeltaRule(CustomOp):
use_flashinfer = supports_flashinfer
if use_flashinfer:
logger.info_once("Using FlashInfer GDN prefill kernel")
logger.info_once("Using FlashInfer GDN prefill kernel", scope="local")
logger.info_once(
"FlashInfer GDN prefill kernel is JIT-compiled; first run may "
"take a while to compile. Set `--gdn-prefill-backend triton` to "
"avoid JIT compile time."
"avoid JIT compile time.",
scope="local",
)
else:
logger.info_once("Using Triton/FLA GDN prefill kernel")
logger.info_once("Using Triton/FLA GDN prefill kernel", scope="local")
self._forward_method = (
self.forward_cuda if use_flashinfer else self.forward_native