datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
curobo
cuRobo Generated Subset
This private dataset contains generated successful RoboDojo incremental demonstrations staged by FWH.
Current subset after swap_blocks replacement:
align_blocks: 106 successful episodes, preserved from previous staging
swap_blocks: 200 successful episodes, regenerated with full vision observations
make_kong: removed from this repository version because the previous 300-episode batch used only 4 replay action templates
Replacement source:… See the full description on the dataset page: https://huggingface.co/datasets/Fwhze/curobo.harbour-fwh-dataset
Harbour / FiveWin (FWH) instruction dataset
Instruction-tuning dataset for the Harbour language and the FiveWin (FWH)
framework: PRG code, low-level Harbour Extend API C wrappers (HB_FUNC) and
xBase idioms, in Spanish and English.
Use harbour_fwh_dataset_v3.jsonl (default config).
Files
file
role
harbour_fwh_dataset_v3.jsonl
Recommended — cleaned, deduped, encoding-fixed
harbour_fwh_dataset_v2.jsonl
Compiler-verified snapshot (2026-07-16), pre-cleanup… See the full description on the dataset page: https://huggingface.co/datasets/fivetech/harbour-fwh-dataset.fw_hq_384_gpt