FROM nvcr.io/nvidia/tritonserver:23.10-py3 CMD ["tritonserver", "--model-repository=/models", "--cache-config=local,size=1048576"]
