---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: unicode
    dtype: string
  - name: meaning
    sequence: string
  - name: primary_meaning
    dtype: string
  - name: prompt
    dtype: string
  splits:
  - name: train
    num_bytes: 31894933.0
    num_examples: 6410
  download_size: 32122546
  dataset_size: 31894933.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

Generated from Kanjivg.
128 x 128 images with meanings.
https://kanjivg.tagaini.net/index.html