sirius/gcloud-lab
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Update rtx6000-vllm.yaml

Browse source
This commit is contained in:
sirius0xdev 2026-05-16 22:35:02 -04:00 • committed by GitHub
parent c12b31a47a
commit e4ce253e80
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 2 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

3
infrastructure/gpus/base/vllm-servers/rtx6000-vllm.yaml
View file

@ -39,7 +39,7 @@ spec:
- --tokenizer-mode=hf - --tokenizer-mode=hf
- --gpu-memory-utilization=0.90 - --gpu-memory-utilization=0.90
#- --enforce-eager #- --enforce-eager
- --max-model-len=98304 - --max-model-len=136876
- --enable-auto-tool-choice - --enable-auto-tool-choice
- --kv-cache-dtype=fp8 - --kv-cache-dtype=fp8
- --max-num-batched-tokens=32768 - --max-num-batched-tokens=32768
@ -51,6 +51,7 @@ spec:
- --enable-prefix-caching - --enable-prefix-caching
- --tool-call-parser=qwen3_xml - --tool-call-parser=qwen3_xml
- --reasoning-parser=qwen3 - --reasoning-parser=qwen3
- --speculative-config '{"method": "mtp", "num_speculative_tokens": 2}'
ports: ports:
- containerPort: 8000 - containerPort: 8000