CLI Arguments

inference

Make an inference to ModelZ deployment

inference [OPTIONS] [UNKNOWN]...

Options

-k, --key <key>

Required API key for Modelz, will read from env $MODELZ_API_KEY if not provided

-e, --endpoint <endpoint>

Required Inference endpoint for Modelz deployment

--serde <serde>

Serialization/deserialization method

Options:

json | msgpack | raw

--read-stdin

Read bytes from stdin

--write-file <write_file>

Write received data to file

Arguments

UNKNOWN

Optional argument(s)

Environment variables

MODELZ_API_KEY

Provide a default for --key

metrics

Get metric from ModelZ deployment

metrics [OPTIONS]

Options

-k, --key <key>

Required API key for Modelz, will read from env $MODELZ_API_KEY if not provided

-d, --deployment <deployment>

Required Deployment id

Environment variables

MODELZ_API_KEY

Provide a default for --key

build

Build image by ModelZ builder

build [OPTIONS]

deployment

Operate to ModelZ deployments

deployment [OPTIONS] COMMAND [ARGS]...

create

deployment create [OPTIONS]

Options

--host <host>

Control Apiserver host for Modelz, will read from env $MODELZ_CTRL_HOST if not provided

-u, --user-id <user_id>

Required login name for Modelz, will read from env $MODELZ_USER if not provided

-k, --key <key>

Required API key for Modelz, will read from env $MODELZ_API_KEY if not provided

--image-source <image_source>

Image pull source

Options:

docker | huggingface

--image <image>

Required URL of Docker image path or HuggingFace project Path

--server-resource <server_resource>

Required Server Resource used for deployment

Options:

cpu-4c-16g | nvidia-ada-l4-2-24c-96g | nvidia-ada-l4-4-48c-192g | nvidia-ada-l4-8c-32g | nvidia-ampere-a100-40g-12c-85g | nvidia-tesla-t4-4c-16g

--framework <framework>

Required Framework of deployment

Options:

gradio | mosec | other | streamlit | unknown

--name <name>

Required Name of deployment

--min-replicas <min_replicas>

MinReplicas is the minimum number of replicas of the deployment

--max-replicas <max_replicas>

MaxReplicas is the maximum number of replicas of the deployment

--target-load <target_load>

TargetLoad is the target load of the deployment. (inflight requests per replica)

--startup-duration <startup_duration>

StartupDuration is the startup timeout

--zero-duration <zero_duration>

ZeroDuration is the idle timeout before scaling to zero

--http-probe-path <http_probe_path>

HTTPProbePath is the user defined path of the http probe

--port <port>

Port is the port of the deployment

--command <command>

Command is the command to run

--env-vars <env_vars>

EnvVars is the environment variables of the deployment, input it by minified json

Environment variables

MODELZ_CTRL_HOST

Provide a default for --host

MODELZ_USER

Provide a default for --user-id

MODELZ_API_KEY

Provide a default for --key

delete

deployment delete [OPTIONS]

Options

--host <host>

Control Apiserver host for Modelz, will read from env $MODELZ_CTRL_HOST if not provided

-u, --user-id <user_id>

Required login name for Modelz, will read from env $MODELZ_USER if not provided

-k, --key <key>

Required API key for Modelz, will read from env $MODELZ_API_KEY if not provided

-d, --deployment <deployment>

Required Deployment id

Environment variables

MODELZ_CTRL_HOST

Provide a default for --host

MODELZ_USER

Provide a default for --user-id

MODELZ_API_KEY

Provide a default for --key

get

deployment get [OPTIONS]

Options

--host <host>

Control Apiserver host for Modelz, will read from env $MODELZ_CTRL_HOST if not provided

-u, --user-id <user_id>

Required login name for Modelz, will read from env $MODELZ_USER if not provided

-k, --key <key>

Required API key for Modelz, will read from env $MODELZ_API_KEY if not provided

-d, --deployment <deployment>

Required Deployment id

Environment variables

MODELZ_CTRL_HOST

Provide a default for --host

MODELZ_USER

Provide a default for --user-id

MODELZ_API_KEY

Provide a default for --key

list

deployment list [OPTIONS]

Options

--host <host>

Control Apiserver host for Modelz, will read from env $MODELZ_CTRL_HOST if not provided

-u, --user-id <user_id>

Required login name for Modelz, will read from env $MODELZ_USER if not provided

-k, --key <key>

Required API key for Modelz, will read from env $MODELZ_API_KEY if not provided

Environment variables

MODELZ_CTRL_HOST

Provide a default for --host

MODELZ_USER

Provide a default for --user-id

MODELZ_API_KEY

Provide a default for --key

update

deployment update [OPTIONS]

Options

--host <host>

Control Apiserver host for Modelz, will read from env $MODELZ_CTRL_HOST if not provided

-u, --user-id <user_id>

Required login name for Modelz, will read from env $MODELZ_USER if not provided

-k, --key <key>

Required API key for Modelz, will read from env $MODELZ_API_KEY if not provided

-d, --deployment <deployment>

Required Deployment id

--min-replicas <min_replicas>

MinReplicas is the minimum number of replicas of the deployment

--max-replicas <max_replicas>

MaxReplicas is the maximum number of replicas of the deployment

--target-load <target_load>

TargetLoad is the target load of the deployment. (inflight requests per replica)

--startup-duration <startup_duration>

StartupDuration is the startup timeout

--zero-duration <zero_duration>

ZeroDuration is the idle timeout before scaling to zero

--env-vars <env_vars>

EnvVars is the environment variables of the deployment, input it by minified json

Environment variables

MODELZ_CTRL_HOST

Provide a default for --host

MODELZ_USER

Provide a default for --user-id

MODELZ_API_KEY

Provide a default for --key