Last active
October 1, 2025 01:04
-
-
Save curtishall/9549f34240ee7446dee7fa4cd4cf861b to your computer and use it in GitHub Desktop.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| version: "3.8" | |
| services: | |
| big-coder: | |
| image: vllm/vllm-openai:latest | |
| container_name: big-coder | |
| restart: unless-stopped | |
| runtime: nvidia | |
| environment: | |
| - CUDA_VISIBLE_DEVICES=${CUDA_VISIBLE_DEVICES:-0,1} | |
| ports: | |
| - "8000:8000" | |
| ipc: host | |
| shm_size: "16gb" | |
| volumes: | |
| - ${HF_CACHE_DIR:-~/.cache/huggingface}:/root/.cache/huggingface | |
| command: > | |
| --model ${MODEL_REPO} | |
| --tensor-parallel-size ${TP_SIZE:-2} | |
| --max-model-len ${MAX_LEN:-16384} | |
| --gpu-memory-utilization ${GPU_UTIL:-0.92} | |
| --trust-remote-code | |
| ---------------------------- | |
| .env | |
| MODEL_REPO=Qwen/Qwen2.5-Coder-32B-Instruct | |
| # Use 2 GPU, or change to 0 | |
| CUDA_VISIBLE_DEVICES=0,1 | |
| TP_SIZE=2 | |
| MAX_LEN=16000 | |
| GPU_UTIL=0.92 | |
| HF_TOKEN= | |
| HF_CACHE_DIR=/opt/hf-cache | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment