Keep sagemaker notebook up during model training?

Question

I start a training job from a sagemaker notebook using boto3.client():

client = boto3.client(service_name='sagemaker')
client.create_training_job(**training_params)

I see the training job in progress on the console. Is it then an error to stop and delete the notebook even if the model has not finished training?

Accepted Answer

Hi @fascani, no, once the training job has been created, you don't need to keep the notebook running. You can continue other explorations/model building on the notebook, check the progress of the training job using APIs, but that's not required.

Keep sagemaker notebook up during model training?

Contenuto pertinente