ops_data/http_grpc/run.sh

docker run --rm --gpus all \
  -p8000:8000 \
  -p8001:8001 \
  -p8002:8002 \
  -v $PWD/model_repo:/models \
  nvcr.io/nvidia/tritonserver:25.04-py3 \
  tritonserver --model-repository=/models