# Local

# Description

Tools used for local development including building, training and deploying a dev endpoint.

# Build

usage: mldock local build [--dir <path>] [--no-cache] [--tag <text>]
                          [--stage <text>]

# Description

Build MLDOCK container image locally

# Options

--dir Path to MLDOCK container project directory.

--no-cache (Optional) Flag which forces docker to build container without cache.

--tag (Default=latest) Selects which container image varient to run container as.

--stage (Optional) Selects which stage to run docker container in.

# Train

usage: mldock local train [--dir <path>] [--params --p <param=value>]
                          [--env_vars --e <env_var=value>] [--tag <text>]
                          [--stage <text>]

# Description

Run training locally on machine

# Options

--dir Path to MLDOCK container project directory.

--params -p (Optional) Temporarily set or override a hyperparameter when running container.

--env_vars -e (Optional) Temporarily set or override a hyperparameter when running container.

--tag (Default=latest) Selects which container image varient to run container as.

--stage (Optional) Selects which stage to run docker container in.

# Deploy

usage: mldock local deploy [--dir <path>] [--port <text>]
                           [--params -p <param=value>]
                           [--env_vars -e <env_var=value>] 
                           [--tag <text>] [--stage <text>]

# Description

Run deploy locally on machine at localhost

# Options

--dir Path to MLDOCK container project directory.

--port path to payload file

--params -p (Optional) Temporarily set or override a hyperparameter when running container.

--env_vars -e (Optional) Temporarily set or override a hyperparameter when running container.

--tag (Default=latest) Selects which container image varient to run container as.

--stage (Optional) Selects which stage to run docker container in.

# Predict

usage: mldock local predict [--payload <path>] [--host <text>]
                            [--content-type <text>]

# Description

Execute prediction request against ml endpoint

# Options

--host server url to send request to.

--payload path to payload file

--content-type (default=application/json) how to serialize the payload in the request to the server.

← Container Registry →