问题描述
我正在尝试向Azure ML
模型注册数据砖块模型tp mlflow.azure.base_image
工作区。但是通过这种方法,我们可以将Azure ML
图像保存到连接到ACR
工作区的默认Azure ML
。
但是我想将Azure ML
图像保存到另一个现有的ACR
中。在确定设计时需要帮助。
我使用的方法如下
workspace = Workspace.create(name = workspace_name,location = workspace_location,resource_group = resource_group,subscription_id = subscription_id,auth=svc_pr,exist_ok=True)
import mlflow.azureml
model_image,azure_model = mlflow.azureml.build_image(model_uri=model_uri,workspace=workspace,model_name="winequality",image_name="winequality",description="Sklearn ElasticNet image for predicting wine quality",synchronous=True)
#model_image.wait_for_creation(show_output=True)
print("Access the following URI for build logs: {}".format(model_image.image_build_log_uri))
解决方法
在创建Azure ML工作区并为使用的服务主体提供必要的权限(参与者)角色时附加现有的ACR。
workspace = Workspace.create(name = workspace_name,location = workspace_location,resource_group = resource_group,subscription_id = subscription_id,auth=svc_pr,container_registry=<resource_id>,exist_ok=True)
import mlflow.azureml
model_image,azure_model = mlflow.azureml.build_image(model_uri=model_uri,workspace=workspace,model_name="winequality",image_name="winequality",description="Sklearn ElasticNet image for predicting wine quality",synchronous=True)
#model_image.wait_for_creation(show_output=True)
print("Access the following URI for build logs: {}".format(model_image.image_build_log_uri))