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 9, 02:30 UTC
[UX] Format attention backend log line (#33570)
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Matthew Bonanni committed
Feb 2, 2026 at 13:57 UTC
5d1aef3004f0e666c613d7edb37b1998d712c18a
1 file changed
+2
-2
vllm/platforms/cuda.py
+2
-2
@@ -362,9 +362,9 @@ class CudaPlatformBase(Platform):
362
selected_index = sorted_indices[0]
363
selected_backend = valid_backends_priorities[selected_index][0]
364
logger.info_once(
365
- "Using %s attention backend out of potential backends: %s",
365
+ "Using %s attention backend out of potential backends: %s.",
366
selected_backend.name,
367
- tuple(b[0].name for b in valid_backends_priorities),
367
+ "[" + ", ".join(f"'{b[0].name}'" for b in valid_backends_priorities) + "]",
368
scope="local",
369
)
370