xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-22 04:45:01 +08:00
Code Issues Packages Projects Releases Wiki Activity

[core] Avoid metrics log noise when idle - include speculative decodi… (#10809)

Browse Source
This commit is contained in:
cduk 2024-12-02 02:49:48 +01:00 committed by GitHub
parent b18c9bbaba
commit b7954776fd
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
vllm/engine/metrics.py
View File

@ -473,13 +473,13 @@ class LoggingStatLogger(StatLoggerBase):
) )
if (stats.cpu_prefix_cache_hit_rate >= 0 if (stats.cpu_prefix_cache_hit_rate >= 0
or stats.gpu_prefix_cache_hit_rate >= 0): or stats.gpu_prefix_cache_hit_rate >= 0):
logger.info( log_fn(
"Prefix cache hit rate: GPU: %.2f%%, CPU: %.2f%%", "Prefix cache hit rate: GPU: %.2f%%, CPU: %.2f%%",
stats.gpu_prefix_cache_hit_rate * 100, stats.gpu_prefix_cache_hit_rate * 100,
stats.cpu_prefix_cache_hit_rate * 100, stats.cpu_prefix_cache_hit_rate * 100,
) )
if self.spec_decode_metrics is not None: if self.spec_decode_metrics is not None:
logger.info( log_fn(
self._format_spec_decode_metrics_str( self._format_spec_decode_metrics_str(
self.spec_decode_metrics)) self.spec_decode_metrics))