root
85 lines
2.1 KiB
YAML
85 lines
2.1 KiB
YAML
services:
|
|
qwen-dense:
|
|
image: nvcr.io/nvidia/vllm:26.04-py3
|
|
container_name: vllm-qwen-dense
|
|
profiles: ["qwen-dense"]
|
|
restart: unless-stopped
|
|
ipc: host
|
|
ports:
|
|
- "8000:8000"
|
|
volumes:
|
|
- /root/.cache/huggingface:/root/.cache/huggingface
|
|
deploy:
|
|
resources:
|
|
reservations:
|
|
devices:
|
|
- driver: nvidia
|
|
count: all
|
|
capabilities: [gpu]
|
|
command: >
|
|
vllm serve Qwen/Qwen3.6-27B-FP8
|
|
--quantization fp8
|
|
--gpu-memory-utilization 0.40
|
|
--kv-cache-dtype fp8
|
|
--max-model-len 32768
|
|
--trust-remote-code
|
|
--enable-auto-tool-choice
|
|
--tool-call-parser hermes
|
|
--served-model-name qwen-3.6-27b-dense
|
|
|
|
qwen-moe:
|
|
image: nvcr.io/nvidia/vllm:26.04-py3
|
|
container_name: vllm-qwen-moe
|
|
profiles: ["qwen-moe"]
|
|
restart: unless-stopped
|
|
ipc: host
|
|
ports:
|
|
- "8001:8000"
|
|
volumes:
|
|
- /root/.cache/huggingface:/root/.cache/huggingface
|
|
deploy:
|
|
resources:
|
|
reservations:
|
|
devices:
|
|
- driver: nvidia
|
|
count: all
|
|
capabilities: [gpu]
|
|
command: >
|
|
vllm serve Qwen/Qwen3.6-35B-A3B-FP8
|
|
--quantization fp8
|
|
--kv-cache-dtype fp8
|
|
--max-model-len 32768
|
|
--gpu-memory-utilization 0.40
|
|
--trust-remote-code
|
|
--enable-auto-tool-choice
|
|
--tool-call-parser hermes
|
|
--served-model-name qwen-3.6-35b-moe
|
|
|
|
gemma:
|
|
image: nvcr.io/nvidia/vllm:26.04-py3
|
|
container_name: vllm-gemma-moe
|
|
profiles: ["gemma-moe"]
|
|
ipc: host
|
|
restart: unless-stopped
|
|
ports:
|
|
- "8002:8000"
|
|
volumes:
|
|
- /root/.cache/huggingface:/root/.cache/huggingface
|
|
environment:
|
|
- HF_TOKEN=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
|
|
deploy:
|
|
resources:
|
|
reservations:
|
|
devices:
|
|
- driver: nvidia
|
|
count: all
|
|
capabilities: [gpu]
|
|
command: >
|
|
vllm serve google/gemma-4-26B-A4B-it
|
|
--quantization fp8
|
|
--kv-cache-dtype fp8
|
|
--max-model-len 8192
|
|
--gpu-memory-utilization 0.40
|
|
--trust-remote-code
|
|
--served-model-name gemma-26b-moe
|