toshi456's picture
Update README.md
998104d verified
metadata
license: apache-2.0
task_categories:
  - image-to-text
language:
  - ja
size_categories:
  - 100K<n<1M

Dataset details

Dataset type: LLaVA JP Instruct 108K created by changing the data sets in Japanese Visual Genome VQA dataset and train data of docci_ja to the LLaVA-Instruct format.

Acknowledgement

Lisence

Apach lisense 2.0