Datasets:

Modalities:
Text
Formats:
json
ArXiv:
Libraries:
Datasets
pandas
License:
GlyphByT5's picture
feat: upd dataset card
fab5785 verified
metadata
license: apache-2.0

GlyphByT5 Pretraining

This is the pretraining data for Glyph-ByT5.

Dataset Details

Glyph-ByT5: A Customized Text Encoder for Accurate Visual Text Rendering
Zeyu Liu, Weicong Liang, Zhanhao Liang, Chong Luo, Ji Li, Gao Huang, Yuhui Yuan Microsoft Research Asia; Tsinghua University; Peking University; The Australian National University
Preprint

Dataset Description

Dataset Structure

This dataset contains a json file containing the annotation file needed for Glyph-ByT5 pretraining. The json file is organized as a list of dicts. Each dict contains the following fields:

Note: As font family and color are randomly selected during pretraining, 'color' and 'font-family' fields are None and later filled randomly in the corresponding dataset code.

Currently, we release a subset of fonts that we use, containing 100 with free commercial use license, see repo for more details.

{
  'texts': [
    text_1,
    ...
    text_k,
  ],
  'styles': [
    {
      'color': color_1,
      'font-family': 'font_1',
    },
    ...
    {
      'color': color_k,
      'font-family': 'font_k',
    },
  ],
  'bbox': [
    [x1, y1, w1, h1, a1],
    ...
    [xk, yk, wk, hk, ak],
  ]
}