0

I'm learning about SageMaker Experiments and I gotcha a client error that I did not figure out what it is. If I understood correctly, I think that is something when the SM tries to save the generated training model. Besides, the training job completes successfully, but it shows this error in the end. I'm using the sklearn.estimator to train the random forest algorithm. Tkx in advance for this help.

The code and the error are the following:

code:

from sagemaker.sklern.estimator import SKLearn

with Run() as run:
   sklearn_estimator = SKLearn(entry_point="rfc_script.py",
                               source_dir="../experiments/scripts",
                               role=get_execution_role(),
                               instance_count=1,
                               intance_type="ml.c5.xlarge",
                               framework_version="1.2-1",
                               output_path=f"s3://{bucket}/{prefix}/outputs",
                               base_job_name="rf-scikit",
                               metric_definitions=metric_definitions,
                               hyperparamenters=hyperparamenters,)

   sklearn_estimator.fit({"train":s3_inputs_train, "test":s3_inputs_test, "validation":s3_inputs_validation}, job_name=f"rf-job{id}", wait= True)

SageMaker Training Job status: Completed

ClientError: enter image description here

Yuri Santos
  • 55
  • 7
  • Did the Training Job complete successfully? Are you using the latest version of the SageMaker SDK? – Marc Karp Aug 07 '23 at 06:39
  • Hi @MarcKarp. Yes, the Training Job completes successfully as we can see in the image before the error arise. Yes. I'm using the latest version. – Yuri Santos Aug 07 '23 at 12:44
  • I think this error is interfering in the `run.log_parameters` because they are not showing up on experiments vis. – Yuri Santos Aug 07 '23 at 14:46

0 Answers0