Create Sklearn model.tar.gz in one sagemaker, use the model in another sagemaker without using endpoint #1854

Activity

1. chuyang-deng commented on Sep 1, 2020

   

   chuyang-deng

   on Sep 1, 2020

   Contributor

   More actions

   Hi @radhakrishnang, if you have a pre-trained model, assuming that you have the S3 uri of the model artifact and your SageMaker role has access to the object, you can do the following to deploy the model to an endpoint:

   https://github.com/aws/sagemaker-python-sdk/blob/master/src/sagemaker/sklearn/model.py#L53
   https://github.com/aws/sagemaker-python-sdk/blob/master/src/sagemaker/model.py#L406

   model = SKLearnModel(model_data=<your pre-trained model location in S3>,
                                            role=<your SageMaker role>,
                                            entrypoint=<your inference script>,
                                            ... # other optional params)
   
   model.deploy(initial_instance_count=<instance count>,
                          instance_type=<chosen instance type>,
                          ... # other optional params)
   

   As for the batch transform, after your endpoint is in service, you can proceed batch transforming as normal.
2. chuyang-deng

   added

   type: question

   on Sep 1, 2020
3. radhakrishnang commented on Sep 2, 2020

   

   radhakrishnang

   on Sep 2, 2020

   Author

   More actions

   Hi @radhakrishnang, if you have a pre-trained model, assuming that you have the S3 uri of the model artifact and your SageMaker role has access to the object, you can do the following to deploy the model to an endpoint:

   https://github.com/aws/sagemaker-python-sdk/blob/master/src/sagemaker/sklearn/model.py#L53
   https://github.com/aws/sagemaker-python-sdk/blob/master/src/sagemaker/model.py#L406

   model = SKLearnModel(model_data=<your pre-trained model location in S3>,
                                            role=<your SageMaker role>,
                                            entrypoint=<your inference script>,
                                            ... # other optional params)
   
   model.deploy(initial_instance_count=<instance count>,
                          instance_type=<chosen instance type>,
                          ... # other optional params)
   

   As for the batch transform, after your endpoint is in service, you can proceed batch transforming as normal.

   Ni hao @ChuyangDeng,

   I want to use model.tar.gz created from already trained model in sagemaker to be directly deployed in another sagemaker environment for batch transform without training again or having endpoint creation ... is it feasible?

   Xiexie
4. bveeramani

   closed this as completedon May 20, 2021
5. aws

   locked and limited conversation to collaborators on May 20, 2021