41554 blogs · 4,504,310 posts Claim your blog

Record · corpus.blog/posts/01a08794-6bcf-70a7-95f4-7d3070c121fd

Battle-Tested LLM Training: From Dataset to Data Iterator

kehang.github.io · published 11 August 2024

A record is what was published and who pointed at it. The text of the post is not held here: read it at the source, or at its Wayback capture.

Read the post at kehang.github.io Wayback capture

0blogs citing

First cited: —
Most recent: —
Rank this month: not ranked
Links from this post: 6
Words captured: 1,589

Cited by

No blog we hold has cited this post yet.

Links from this post

Link

Host

https://github.com/google/maxtext/blob/da50760ac0baf3920305a365215f6f0c5f110ad2/MaxText/input_pipeline/_hf_data_processing.py

github.com

https://huggingface.co/datasets/stas/openwebtext-10k

huggingface.co

https://github.com/google/maxtext/blob/ed21f6ad1bc60285958d585753dda01e1ddfa664/MaxText/input_pipeline/_input_pipeline_utils.py

github.com

https://github.com/google/grain/blob/9b984e8a2ccdd3d377cb43b17591080b44c07009/grain/_src/python/experimental/example_packing/packing.py

github.com

https://colab.research.google.com/drive/1MrIvDAiWcTSma3mDKwmd5F_cxVznfxmQ

colab.research.google.com

https://github.com/google/maxtext/blob/ead18fbe6f2d8a6cbae6bbd38568146919e20e18/MaxText/multihost_dataloading.py

github.com