mms_app_gpu.conf

[MMS Arguments]
--models
squeezenet=https://s3.amazonaws.com/model-server/models/squeezenet_v1.1/squeezenet_v1.1.model

--service
optional

--gen-api
optional

--log-file
optional

--log-rotation-time
optional

--log-level
optional

--metrics-write-to
optional

--num-gpu
optional

[Gunicorn Arguments]

--bind
unix:/tmp/mms_app.sock

###Number of Gunicorn workers optimised for p3.8xlarge instance ##########
--workers
optional


--worker-class
gevent

--limit-request-line
0

[Nginx Configurations]
server {
    listen       8080;

    location / {
        proxy_pass http://unix:/tmp/mms_app.sock;
    }
}