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, 21:00 UTC
[V0][Metrics] Remove unimplemented `vllm:tokens_total` (#14134)
Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Mark McLoughlin committed
Mar 3, 2025 at 17:50 UTC
c41d27156b7c9123bd38387afca639631dfc2ed0
1 file changed
-4
vllm/engine/metrics.py
-4
@@ -115,10 +115,6 @@ class Metrics:
115
name="vllm:generation_tokens_total",
116
documentation="Number of generation tokens processed.",
117
labelnames=labelnames)
118
- self.counter_tokens = self._counter_cls(
119
- name="vllm:tokens_total",
120
- documentation="Number of prefill plus generation tokens processed.",
121
- labelnames=labelnames)
118
buckets = [1, 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8096]
119
if not vllm_config.model_config.enforce_eager:
120
buckets = vllm_config.compilation_config.\