docker run --gpus all -it \
  --ipc=host \
  -e VLLM_DISABLE_PYNCCL=1 \
  -p 127.0.0.1:8001:8000 \
  -v /mnt/models:/models \
  -v /mnt/lectures:/lectures \
  absence209/vllm-sm70 \
    --model /models/Qwen2.5-VL-7B-Instruct-AWQ \
    --served-model-name qwen-vl \
    --dtype float16 \
    --quantization awq \
    --tensor-parallel-size 2 \
    --gpu-memory-utilization 0.80 \
    --max-model-len 8192 \
    --max-num-seqs 1 \
    --max-num-batched-tokens 1024 \
    --attention-backend TRITON_ATTN \
    --disable-custom-all-reduce \
    --trust-remote-code \
    --limit-mm-per-prompt '{"image":1,"video":0}' \
    --allowed-local-media-path /lectures \
    --mm-processor-kwargs '{"max_pixels": 802816}' \
    --host 0.0.0.0 \
    --port 8000