data for training

[liuBY001]

I saw that the training data path is an internal path(/weka/oe-training-default/ai2-llm). Do you have plan to open source the training data?

[liuBY001]

Or could you guys provide a transforming script or format specification for the internal files?

FileNotFoundError: http://flexolmo-data.org/preprocessed/dclm/baseline_type_classified/news_article/allenai/dolma2-tokenizer/0000/part-000-00000.csv.gz