50 lines
1.1 KiB
YAML
50 lines
1.1 KiB
YAML
services:
|
|
vllm:
|
|
image: vllm/vllm-openai:latest
|
|
container_name: vllm
|
|
restart: unless-stopped
|
|
|
|
ipc: host
|
|
gpus: all
|
|
|
|
ports:
|
|
- "8001:8000"
|
|
|
|
volumes:
|
|
- /opt/vLLM/models:/root/.cache/huggingface
|
|
- /opt/vLLM/config:/vllm-workspace/config
|
|
|
|
environment:
|
|
HUGGING_FACE_HUB_TOKEN: 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
|
|
NVIDIA_VISIBLE_DEVICES: "0,1"
|
|
NVIDIA_DRIVER_CAPABILITIES: compute,utility
|
|
|
|
shm_size: "16g"
|
|
|
|
command:
|
|
- "--model"
|
|
- "Qwen/Qwen3-14B-AWQ"
|
|
- "--served-model-name"
|
|
- "qwen3-14b"
|
|
- "--host"
|
|
- "0.0.0.0"
|
|
- "--port"
|
|
- "8000"
|
|
- "--tensor-parallel-size"
|
|
- "2"
|
|
- "--dtype"
|
|
- "half"
|
|
- "--gpu-memory-utilization"
|
|
- "0.90"
|
|
- "--max-model-len"
|
|
- "65536"
|
|
- "--rope-scaling"
|
|
- '{"rope_type":"yarn","factor":2.0,"original_max_position_embeddings":40960}'
|
|
- "--max-num-seqs"
|
|
- "16"
|
|
- "--kv-cache-dtype"
|
|
- "fp8"
|
|
- "--enable-prefix-caching"
|
|
- "--enable-auto-tool-choice"
|
|
- "--tool-call-parser"
|
|
- "hermes" |