C2-W1 assignment processing job failing

When running this cell:

%%time

running_processor = sagemaker.processing.ProcessingJob.from_processing_name(
    processing_job_name=scikit_processing_job_name,
    sagemaker_session=sess
)

running_processor.wait(logs=False)

I get a failure -
UnexpectedStatusException: Error for Processing job sagemaker-scikit-learn-2022-09-29-08-59-09-114: Failed. Reason: AlgorithmError: See job logs for more information

The CloudWatch logs end with the following:

	2022-09-29T11:56:24.993+03:00
Traceback (most recent call last): File "/opt/ml/processing/input/code/prepare_data.py", line 26, in <module> subprocess.check_call([sys.executable, "-m", "conda", "install", "-c", "pytorch", "pytorch==1.6.0", "-y"]) File "/miniconda3/lib/python3.7/subprocess.py", line 363, in check_call raise CalledProcessError(retcode, cmd)

2022-09-29T11:56:24.993+03:00
subprocess.CalledProcessError: Command '['/miniconda3/bin/python3', '-m', 'conda', 'install', '-c', 'pytorch', 'pytorch==1.6.0', '-y']' returned non-zero exit status 1.

This is not an issue with completing the task - I did edit the “convert_to_bert_input_ids” method as required.

Hi @Edden_Gerber, if you can please share how did you solve this previous exercise you have mentioned?

Thanks!

C2_W1_Assignment.ipynb (49.1 KB)
Hi, I attached my solution that evoked the above error.
Edden

I tried to start the assignment again just in case the problem went away, but now it fails even sooner: when first importing the prepare_data.py script, it fails with the error:

CalledProcessError: Command '['/opt/conda/bin/python', '-m', 'conda', 'install', '-c', 'conda-forge', 'transformers==3.5.1', '-y']' died with <Signals.SIGKILL: 9>.

Attaching my script again, the error is under the cell with the “This cell will take approximately 1-2 minutes to run” title.

I proceeded with the course without completing the task but it’s a real pain.
C2_W1_Assignment.ipynb (41.3 KB)

thanks for sharing. I’ll report this to the teaching staff.

@Edden_Gerber (cc @Raul) thank you for reporting the issue and for the notebook. We are investigating the problem and hopefully it will be resolved as soon as possible. We will keep you updated.

@Edden_Gerber (cc @Raul) the problem is fixed, that line in the prepare_data.py script should be changed with the other one:
subprocess.check_call([sys.executable, "-m", "pip", "install", "transformers==3.5.1"])
The lab files are updated in the S3 Bucket.

Thank you for reporting the issue!

Thank you, it’s now working.