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

refactor: upgrade vLLM brain to uncensored Qwen 2.5 32B AWQ

Browse source
This commit is contained in:
SiriusClaw 2026-04-18 15:26:19 +00:00
parent e68774da48
commit 55a3e9eb28
1 changed files with 6 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

10
apps/base/customer1/openclaw/vllm-gemma.yaml
View file

@ -23,16 +23,18 @@ spec:
value: "present" value: "present"
effect: "NO_SCHEDULE" effect: "NO_SCHEDULE"
containers: containers:
- name: vllm-gemma - name: vllm-brain
image: vllm/vllm-openai:latest image: vllm/vllm-openai:latest
command: ["python3", "-m", "vllm.entrypoints.openai.api_server"] command: ["python3", "-m", "vllm.entrypoints.openai.api_server"]
args: args:
- "--model" - "--model"
- "cognitivecomputations/dolphin-2.8-gemma-2b-v0" # Uncensored/Distilled Dolphin-Gemma model - "cognitivecomputations/dolphin-2.9.4-qwen2.5-32b-AWQ" # Uncensored 32B Reasoning Heavyweight (Quantized)
- "--quantization"
- "awq"
- "--dtype" - "--dtype"
- "bfloat16" - "half"
- "--max-model-len" - "--max-model-len"
- "8192" - "16384"
ports: ports:
- containerPort: 8000 - containerPort: 8000
resources: resources: