[bug] Can't build containers locally

[jbohnslav]

Checklist

• I've prepended issue tag with type of change: [bug]
• (If applicable) I've attached the script to reproduce the bug
• (If applicable) I've documented below the DLC image/dockerfile this relates to
• (If applicable) I've documented below the tests I've run on the DLC image
• I'm using an existing DLC image listed here: https://docs.aws.amazon.com/deep-learning-containers/latest/devguide/deep-learning-containers-images.html
• I've built my own container based off DLC (and I've attached the code used to build my own image)

Concise Description:
I want to be able to build PyTorch containers locally so that I can modify them and be sure that everything is running correctly (patching, etc).

Steps to reproduce, straight from the README

aws ecr get-login-password --region us-west-2 | docker login --username AWS --password-stdin $ACCOUNT_ID.dkr.ecr.us-west-2.amazonaws.com

python3 -m venv dlc
source dlc/bin/activate
pip install -r src/requirements.txt
pip install -e .
bash src/setup.sh pytorch

python src/main.py --buildspec pytorch/training/buildspec-2-5-sm.yml \
                   --framework pytorch \
                   --image_types training \
                   --device_types gpu \
                   --py_versions py3

Error message:

Traceback (most recent call last):
  File "/path/to/deep-learning-containers/src/main.py", line 140, in <module>
    main()
  File "/path/to/deep-learning-containers/src/main.py", line 136, in main
    image_builder(buildspec_file, image_types, device_types)
  File "/path/to/deep-learning-containers/src/image_builder.py", line 378, in image_builder
    patch_helper.initiate_multithreaded_autopatch_prep(
  File "/path/to/deep-learning-containers/src/patch_helper.py", line 352, in initiate_multithreaded_autopatch_prep
    run(f"aws s3 cp s3://patch-dlc {download_path} --recursive", hide=True)
  File "/path/to/deep-learning-containers/dlc/lib/python3.11/site-packages/invoke/__init__.py", line 50, in run
    return Context().run(command, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/path/to/deep-learning-containers/dlc/lib/python3.11/site-packages/invoke/context.py", line 104, in run
    return self._run(runner, command, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/path/to/deep-learning-containers/dlc/lib/python3.11/site-packages/invoke/context.py", line 113, in _run
    return runner.run(command, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/path/to/deep-learning-containers/dlc/lib/python3.11/site-packages/invoke/runners.py", line 395, in run
    return self._run_body(command, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/path/to/deep-learning-containers/dlc/lib/python3.11/site-packages/invoke/runners.py", line 451, in _run_body
    return self.make_promise() if self._asynchronous else self._finish()
                                                          ^^^^^^^^^^^^^^
  File "/path/to/deep-learning-containers/dlc/lib/python3.11/site-packages/invoke/runners.py", line 518, in _finish
    raise UnexpectedExit(result)
invoke.exceptions.UnexpectedExit: Encountered a bad command exit code!

Command: 'aws s3 cp s3://patch-dlc /path/to/patch-dlc --recursive'

Exit code: 1

Do I need to be on EC2 to run this? Is it not possible for anyone other than AWS employees to access the s3://patch-dlc bucket?

[BaiqingL]

Also running into the same issue, python version 3.12.9, following readme gets me the error

Traceback (most recent call last):
  File "/home/andy/sage/deep-learning-containers/src/main.py", line 10, in <module>
    from image_builder import image_builder
  File "/home/andy/sage/deep-learning-containers/src/image_builder.py", line 27, in <module>
    import patch_helper
  File "/home/andy/sage/deep-learning-containers/src/patch_helper.py", line 9, in <module>
    from output import OutputFormatter
  File "/home/andy/sage/deep-learning-containers/src/output.py", line 21, in <module>
    import pyfiglet
  File "/home/andy/sage/deep-learning-containers/dlc/lib/python3.12/site-packages/pyfiglet/__init__.py", line 11, in <module>
    import pkg_resources
ModuleNotFoundError: No module named 'pkg_resources'

[Yadan-Wei]

The package missing here is used for patch CVEs, you can build image without enabling patching by disabling this tag in buildspec https://github.com/aws/deep-learning-containers/blob/master/pytorch/training/buildspec-2-5-sm.yml#L8.