Update rtx6000-vllm.yaml

This commit is contained in:
sirius0xdev 2026-05-24 15:55:27 -04:00 committed by GitHub
parent 2b9630d8dc
commit f6040befbd
No known key found for this signature in database
GPG key ID: B5690EEEBB952194

View file

@ -40,7 +40,7 @@ spec:
- --gpu-memory-utilization=0.90
- --max-model-len=262144
- --enable-auto-tool-choice
- --kv-cache-dtype=bfloat16
- --kv-cache-dtype=auto
- --max-num-batched-tokens=16384
- --max-num-seqs=4
- --block-size=32