Datasets:

Modalities:
Tabular
Text
Formats:
parquet
Languages:
code
ArXiv:
Libraries:
Datasets
Dask
License:

Where is the-stack-v2-train-extras?

#17
by Chinglin - opened

Where is the extra?
But I did not find them in the dataset https://huggingface.co/datasets/bigcode/the-stack-v2, Could you please give me a hint ✨where these data are🧐?
Thanks a lot. 🙏🙏

Chinglin changed discussion title from Where are other data apart from SWH data? to Where is the-stack-v2-train-extras?

+1, where can we access github issues, pull requests, notebooks and parsed documentation that was used in the training data?
It is written here https://huggingface.co/blog/starcoder2 that "all models, datasets, and the processing as well as the training code" are getting released but I could not find them anywhere.
Thank you 🙏

BigCode org

The code documentation is released at https://huggingface.co/datasets/SivilTaram/starcoder2-documentation (non-official release)

@SivilTaram Thanks for the great resource! I notice there are different categories of extra data such as pull request, issues, Jupyter structured...
I wonder if there is a way to extract data based on categories? Thank you!

Sign up or log in to comment