次の方法で共有


クラス DeployClient

azureml.deploy.DeployClient(host, auth=None, use=None)

デプロイ クライアントを作成するためのファクトリを定義します。

use プロパティを使用した基本的な使用モジュール実装プラグイン:

インポート参照から "モジュール" 検索して読み込む:

from azureml.deploy import DeployClient
from azureml.deploy.server import MLServer

host = 'http://localhost:12800'
ctx = ('username', 'password')
mls_client = DeployClient(host, use=MLServer, auth=ctx)

名前空間 str から use による定義に基づいて "モジュール" を検索して読み込む:

host = 'http://localhost:12800'
ctx = ('username', 'password')

mls_client = DeployClient(host, use=MLServer, auth=ctx)
mls_client = DeployClient(host, use='azureml.deploy.server.MLServer',
auth=ctx)

ファイルとパスのタプルから "モジュール" を検索して読み込む:

host = 'http://localhost:12800'
ctx = ('username', 'password')

use = ('azureml.deploy.server.MLServer', '/path/to/mlserver.py')
mls_client = DeployClient(host, use=use, auth=ctx)

新しいデプロイ クライアントを作成します。

引数

host

サーバーの HTTP/HTTPS エンドポイント (ポート番号を含む)。

auth

(省略可能) 認証コンテキスト。 すべてのデプロイ クライアントで認証が必要になるわけではありません。 MLServer では auth必要です

use

(必須) 使用するデプロイの実装 (例) ML Server を使用する場合は use='MLServer'