Skip to content
/ dolma Public
forked from allenai/dolma

OLMo's favorite snack! Data and tools for generating and inspecting OLMo pre-training data.

License

Notifications You must be signed in to change notification settings

ianand/dolma

 
 

Repository files navigation

dolma

Data to feed OLMo's Appetite

DOLMa logo. It's a watercolor of grape leaves with the word DOLMa in the top left.

Data and tools for generating and inspecting OLMo pre-training data.

To get started, install dolma using pip.

pip install dolma

Usage

The dolma CLI can be access using the dolma command. To see the available commands, use the --help flag.

dolma --help

At the moment, the CLI supports three commands: tag, dedupe, and mix.

For all commands, configurations can be specified from command line, or by passing a YAML or JSON file using the -c flag. For example:

dolma -c config.yaml dedupe --dedupe.name "test"

dolma tag

The tag command is used to run any of the built-in taggers on a set of documents. For example:

dolma tag \
    --experiment sample \
    --documents \
        's3:https://ai2-llm/pretraining-data/sources/common-crawl/test/v0/documents/**/*.json.gz' \
        's3:https://ai2-llm/pretraining-data/sources/common-crawl/test/v1/documents/*.json.gz' \
    --taggers random_number_v1 \
    --processes 2

This command will run the random_number_v1 tagger on all documents in the specified S3 paths. The results will be written to the s3:https://ai2-llm/pretraining-data/sources/common-crawl/test/v0/attributes/sample and s3:https://ai2-llm/pretraining-data/sources/common-crawl/test/v1/attributes/sample paths.

dolma dedupe

The dedupe command is used to deduplicate a set of documents at the attribute level using a bloom filter. For example configurations, see directory test/config. For example:

dolma dedupe -c test/config/dedupe-paragraphs.json

dolma mix

The mix command is used to mix documents from multiple sources, optionally filtering by attributes and/or performing string replacement. For example configurations, see directory test/config. For example:

dolma mix -c test/config/mixer.json

Development

Create a conda environment with Python >= 3.8. In this case, we use Python 3.10 and use Anaconda to create the environment.

conda create -n dolma python=3.10

After creating the environment, activate it and install necessary tools using the included makefile.

conda activate dolma
make setup

Finally, to begin development, install the repository in editable mode using maturin.

make develop

To run tests, use the following command.

make test

You can choose to run just the Python or Rust tests by calling make test-python or make test-rust respectively.

Citation

If you use this repository, please cite it as:

@software{dolma,
    author = {{Soldaini, Luca and Lo, Kyle and Kinney, Rodney and Naik, Aakanksha and Ravichander, Abhilasha and Bhagia, Akshita and Groeneveld, Dirk and Schwenk, Dustin and Magnusson, Ian and Chandu, Khyathi}},
    license = {{Apache-2.0}},
    title = {{DOLMa}},
    url = {https://github.com/allenai/dolma}
}

About

OLMo's favorite snack! Data and tools for generating and inspecting OLMo pre-training data.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 69.8%
  • Rust 29.2%
  • Makefile 1.0%