Refine code clarification
From your RefineCode dataset statement: "Currently, we have uploaded the meta data covered by The Stack V2 (About 50% file volume). Due to complex legal considerations, we are unable to provide the complete source code currently. We are working hard to make the remaining part available."
And the training data from your paper as below, I would like to confirm what you have released is the 60 billion tokens Stack v2 data or 480B tokens (GitHub Code + Jupyter Notebooks + Stack v2)?
Raw Code Data:
GitHub Code: 755 billion tokens (78.4%)
Jupyter Notebooks: 11 billion tokens (1.1%)
The Stack v2: 120 billion tokens (12.5%)
Code-related Web Data:
Processed CC: 13 billion tokens (1.4%)
Processed SkyPile: 3 billion tokens (0.3%)
Processed FineWeb: 55 billion tokens (5.7%)
OpenSource Data:
Processed AutoMathText: 3 billion tokens (0.3%)
Hi! This meta data contains the raw code from Github repositories which appear in The Stack V2, so it belongs to 755B Github Code.
Thanks for the clarification! It sounds like you've shared about 50% of the 755B GitHub code. Downloading 330 million GitHub code files is indeed a significant challenge. Would it be possible for you to provide the data in a more accessible format, such as a pre-packaged dataset or through an alternative sharing method?
Thanks for the clarification! It sounds like you've shared about 50% of the 755B GitHub code. Downloading 330 million GitHub code files is indeed a significant challenge. Would it be possible for you to provide the data in a more accessible format, such as a pre-packaged dataset or through an alternative sharing method?
Same question! The StackV2 dataset was extremely difficult to download, making it nearly impossible to use this meta data to join the original contents...