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, 14:45 UTC
[BugFix] logger is not callable (#16312)
Signed-off-by: yihong0618 <zouzou0208@gmail.com>
yihong committed
Apr 9, 2025 at 15:59 UTC
819d548e8a4e56f0e68ea8b4f9bf41f759548191
1 file changed
+2
-2
vllm/attention/backends/hpu_attn.py
+2
-2
@@ -149,8 +149,8 @@ class HPUAttentionImpl(AttentionImpl, torch.nn.Module):
149
self.fused_scaled_dot_product_attention = ModuleFusedSDPA(
150
FusedSDPA)
151
except ImportError:
152
- logger().warning("Could not import HPU FusedSDPA kernel. "
153
- "vLLM will use native implementation.")
152
+ logger.warning("Could not import HPU FusedSDPA kernel. "
153
+ "vLLM will use native implementation.")
154
155
suppored_head_sizes = HPUPagedAttention.get_supported_head_sizes()
156
if head_size not in suppored_head_sizes: