No project description provided
Project description
NAACL Utils
A command line tool to help authors submit Docker images to the NAACL Reproducibility Track.
Requires Docker and Python 3.7 or newer. If you don't have a suitable Python installation, see Installing Python below.
Prerequisites
Please follow these carefully before doing anything else.
- Create an account on Beaker.org. You can sign up either with GitHub or your email, both are fine.
- Find and copy your Beaker username at beaker.org/user under the "User Details" section.
- Open a new Beaker add request issue and fill in all of the information that it asks you for, including your Beaker username. This will notify a NAACL reviewer that they need to add you to the NAACL organization on Beaker.
- Wait for a NAACL reviewer to close the issue with confirmation that they've added you to the Beaker organization.
- Then proceed to install and configure naacl-utils.
After completing these steps, you are ready to submit a Docker image for review.
Installing and configuring naacl-utils
Before attempting to install naacl-utils, make sure you have Python 3.7 or newer installed on your machine.
You can check this by running python --version
or python3 --version
. If both of those commands return an error or display a version that's older than 3.7, see Installing Python below.
Once you have a suitable Python installation on your machine, you can install naacl-utils directly with pip:
pip install naacl-utils
After the installation completes, you'll need to run the setup
command once to configure naacl-utils:
naacl-utils setup
Then follow the prompts to complete the setup.
Submitting a Docker image
You can submit a new Docker image with the naacl-utils submit
command. For example:
docker pull nvidia/cuda:11.0-base
naacl-utils submit nvidia/cuda:11.0-base cuda-test-run-1 --cmd nvidia-smi
The first argument to naacl-utils submit
is the name of your Docker image (e.g. "nvidia/cuda:11.0-base"), and the second is an arbitrary unique name you assign to your submission ("cuda-test-run-1").
If you make another submission you'll have to use a different name.
The --cmd
parameter is optional. You can use this to override the CMD
of your Docker image. Similarly there is also the --entrypoint
parameter for overriding the ENTRYPOINT
of your image.
If the submission was successful it will print out a link on Beaker.org that you can follow to track the progress of your submission and view the logs. If the run fails for some reason you can use the logs to debug it and then resubmit when you're ready.
Installing Python
naacl-utils requires Python 3.7 or newer. If you don't already have a suitable Python installation on your machine, the easiest way to get one is with Miniconda.
Miniconda is straight-forward to install on MacOS, Linux, and Windows. On MacOS, for example, you can install Miniconda via Homebrew:
brew install miniconda
Otherwise just use the official installer links.
Once you have Miniconda installed, you can create and activate a new Python 3.7 virtual environment by running:
conda create -n naacl python=3.7
conda activate naacl
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for naacl_utils-0.1.4-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | ed7dd1ad6ef8f3bf030ee829071fcb510a5c0268a8f6f44f787742f67d0ce346 |
|
MD5 | 75ca2a76d55e79b6d4c3eea19d37362b |
|
BLAKE2b-256 | 75589b8e197cd3fa6c0647a984249f88f1d35743317517cdd44f62133a180206 |