zhiyuanhucs/general-game-reasoning-0919-history15

0919 history15 reasoning training set Three games only. No LLaVA mixture. Images are packed as one uncompressed tar per chunk so the upload is a few hundred large files instead of millions of jpgs. After download, extract next to the jsonl. Each tar contains chunk_XXX/..., and the jsonl paths are relative to this dataset root. find . -name 'chunk_*.tar' -print0 | xargs -0 -P 16 -I{} tar -xf {} -C "$(dirname {})" find . -name 'chunk_*.tar' -delete game train val… See the full description on the dataset page: https://huggingface.co/datasets/zhiyuanhucs/general-game-reasoning-0919-history15.

<iframe src="https://silytics.si/embed/item/059a5309-bc8d-45e3-a2e8-4c49c8584bb4" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.