Dockerfile

FROM nvidia/cuda:12.1.1-cudnn8-runtime-ubuntu22.04

RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y \
    git ffmpeg python3-pip python3-dev gcc && \
    rm -rf /var/lib/apt/lists/*

RUN pip3 install --upgrade pip
RUN pip3 install --no-cache-dir \
    ms-swift bitsandbytes decord qwen-vl-utils gradio hf_transfer

ENV GRADIO_SERVER_NAME=0.0.0.0 \
    GRADIO_SERVER_PORT=7860 \
    HF_HUB_ENABLE_HF_TRANSFER=1 
    
ENV FPS_MIN_FRAMES=24 \
    FPS_MAX_FRAMES=24 \
    VIDEO_MAX_PIXELS=100352 \
    HF_HOME=/tmp/huggingface

    
RUN mkdir -p $HF_HOME && chmod -R 777 $HF_HOME

RUN python3 - <<'PY'
from huggingface_hub import snapshot_download
snapshot_download("Qwen/Qwen2-VL-7B-Instruct")
snapshot_download("SimulaMet/SoccerChat-qwen2-vl-7b")
PY

EXPOSE 7860

RUN mkdir -p /.cache && \
    chmod -R 777 /.cache && chmod -R 777 /root/.cache

RUN 
    
CMD swift app \
  --adapters "SimulaMet/SoccerChat-qwen2-vl-7b" \
  --model "Qwen/Qwen2-VL-7B-Instruct" \
  --use_hf true \
  --attn_impl sdpa \
  --quant_method bnb --quant_bits 4 \
  --bnb_4bit_quant_type nf4 \
  --bnb_4bit_use_double_quant true \
  --bnb_4bit_compute_dtype float16 \
  --max_batch_size 1 \
  --is_multimodal true \
  --studio_title "SoccerChat" \
  --stream true \
  --server_name 0.0.0.0 --server_port 7860 --share false --lang en