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, 01:45 UTC
[Misc] Fix logger format typo (#4396)
Roy committed
Apr 27, 2024 at 13:45 UTC
87f545ba6fdbbbe9813736fc398874563e2604a7
1 file changed
+2
-2
vllm/engine/metrics.py
+2
-2
@@ -230,8 +230,8 @@ class StatLogger:
230
"Avg prompt throughput: %.1f tokens/s, "
231
"Avg generation throughput: %.1f tokens/s, "
232
"Running: %d reqs, Swapped: %d reqs, "
233
- "Pending: %d reqs, GPU KV cache usage: %.1f%, "
234
- "CPU KV cache usage: %.1f%",
233
+ "Pending: %d reqs, GPU KV cache usage: %.1f%%, "
234
+ "CPU KV cache usage: %.1f%%",
235
prompt_throughput,
236
generation_throughput,
237
stats.num_running,