GET
/
v2
/
functions
/
{owner_name}
/
{function_name}
/
versions
/
{version_slug}

Authorizations

X-API-Key
string
headerrequired

The Sieve API key to authenticate with.

Path Parameters

owner_name
string
required

The owner of the function.

function_name
string
required

The name of the function.

version_slug
string
required

The ID or ID prefix of the function version, or 'latest'.

Response

200 - application/json
id
string
required

Id of the model

build_status
string
required

Build status of the function version. Statuses are 'queued', 'building', 'pushing', 'deploying', 'ready', 'error'

queued_at
string | null
required

The UTC time that the function build is queued.

built_at
string | null
required

The UTC time that the function build finished building.

ready_at
string | null
required

The UTC time that the model is ready to process jobs.

compute_type
string
required

Machine type to use. See possible values for GPU machine types

python_packages
string[]
required

An array of the model's python pip packages

system_packages
string[]
required

An array of the model's system packages

gpu
boolean
required

Whether this model uses gpu

split
integer
required

How many functions to allocate per node

inputs
object[]
required

The inputs to a function

outputs
object[]
required

The outputs to a function

environment_variables
required

The outputs to a function

function_dependencies
string[]
required

Function dependencies

cuda_version
string | null

Cuda Version

minimum_replicas
integer
default: 0

The minimum number of replicas of the model. Scale your model's minimum replicas to avoid cold starts

maximum_replicas
integer | null
default: 0

The maximum number of replicas of the model. Prevents model replicas from scaling beyond a certain point

current_workers
object | null

The current number and status of the model's workers

stream_output
boolean | null
default: false

Stream output