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 17, 03:00 UTC
[Bugfix] fix_log_time_in_metrics (#4050)
zspo committed
Apr 13, 2024 at 22:52 UTC
ec8e3c695f2dce080bde569746180300e91084a3
1 file changed
+1
-1
vllm/engine/metrics.py
+1
-1
@@ -130,7 +130,7 @@ class StatLogger:
130
131
def __init__(self, local_interval: float, labels: Dict[str, str]) -> None:
132
# Metadata for logging locally.
133
- self.last_local_log = time.monotonic()
133
+ self.last_local_log = time.time()
134
self.local_interval = local_interval
135
136
# Tracked stats over current local logging interval.