Restore order of split names in dataset_info for canonical datasets

[albertvillanova]

After a bulk edit of canonical datasets to create the YAML dataset_info metadata, the split names were accidentally sorted alphabetically. See for example:

• https://huggingface.co/datasets/bc2gm_corpus/commit/2384629484401ecf4bb77cd808816719c424e57c

Note that this order is the one appearing in the preview of the datasets.

I'm making a bulk edit to align the order of the splits appearing in the metadata info with the order appearing in the loading script.

Related to:

• CI fails after bulk edit of canonical datasets #5202

[albertvillanova]

The bulk edit is running...

See for example:

• A single config: https://huggingface.co/datasets/acronym_identification/discussions/2
• Multiple configs: https://huggingface.co/datasets/babi_qa/discussions/1

[albertvillanova]

TODO: Add "dataset_info" YAML metadata to:

• "chr_en" has no metadata JSON file, nor "dataset_info" YAML tag in its card
  • Fixing PR: https://huggingface.co/datasets/chr_en/discussions/1
• "conll2000" has no metadata JSON file, but it has "dataset_info" YAML tag in its card
• "crime_and_punish" has no metadata JSON file, but it has "dataset_info" YAML tag in its card
• "dart" has no metadata JSON file, but it has "dataset_info" YAML tag in its card
• "iwslt2017" has no metadata JSON file, but it has "dataset_info" YAML tag in its card
• "mc4" has no metadata JSON file, nor "dataset_info" YAML tag in its card
• "the_pile" has no metadata JSON file, nor "dataset_info" YAML tag in its card
• "timit_asr" has no metadata JSON file, nor "dataset_info" YAML tag in its card

[albertvillanova]

The bulk edit is finished.