| set -ex | |
| echo "--- Starting setup and server ---" | |
| echo "Starting server..." | |
| exec vllm serve ${HF_MODEL_PATH} \ | |
| --host 0.0.0.0 \ | |
| --port ${PORT} \ | |
| --tensor-parallel-size 1 \ | |
| --gpu-memory-utilization 0.8 \ | |
| --chat-template-content-format string \ | |
| --served-model-name dotsocr-model \ | |
| --trust-remote-code | |