Clarification: dataset row count differs from EMBER2024 docs (3,232,000 vs 6,464,000)

[tvquynh]

Hi authors,
The EMBER2024 notebook/paper lists the dataset size as 3,232,000 rows, but the file I just downloaded shows 6,464,000 rows (rows doubled, columns unchanged).

Is this the intended updated release, or did I download the wrong artifact?

If intended, could you confirm the expected size and update the docs/changelog (and ideally provide a version tag/checksum)?

Thanks!

[tvquynh]

I found the duplicated sha256 in the same .jsonl file. Please help to check at your side.

Thanks!