|
|
--- |
|
|
license: mit |
|
|
task_categories: |
|
|
- text-classification |
|
|
language: |
|
|
- en |
|
|
tags: |
|
|
- language |
|
|
- machine-learning |
|
|
- ascii |
|
|
configs: |
|
|
- config_name: default_uppercase |
|
|
data_files: |
|
|
- split: main_split |
|
|
path: uppercase/**/*.txt |
|
|
--- |
|
|
|
|
|
## Dataset Description |
|
|
|
|
|
This dataset provides programmatically generated ASCII representations of the English alphabet rendered using 571 fonts from the PyFiglet library. Each letter (A–Z) is available in multiple typographic styles, resulting in a structured and high-variability dataset suitable for research, experimentation, and creative applications. |
|
|
|
|
|
The dataset was created to support tasks involving text-based pattern recognition, synthetic data generation, typography analysis, and machine learning workflows. Because all samples are generated from a deterministic process, the dataset ensures consistency while preserving strong visual diversity across fonts. |
|
|
|
|
|
### Dataset Features |
|
|
|
|
|
- **571 fonts** sourced from PyFiglet |
|
|
- **26 letters per font** (English alphabet, A–Z) |
|
|
- **Programmatically generated** for reproducibility |
|
|
- **Structured format** for easy parsing and downstream processing |
|
|
- **High visual variability** across character styles |
|
|
- In addition to the individual files (e.g., `file_01.txt`, `file_02.txt`, etc.), the dataset also includes an additional file named `labels.txt` in uppercase, which contains the labels corresponding to the file names. |
|
|
|
|
|
### Potential Use Cases |
|
|
|
|
|
- Machine learning and deep learning experiments |
|
|
- Optical character recognition (OCR) research on text-based representations |
|
|
- Synthetic dataset benchmarking |
|
|
- Generative art and creative coding |
|
|
- Typography and character-style analysis |
|
|
- NLP experiments involving symbolic representations |
|
|
|
|
|
### Generation Methodology |
|
|
|
|
|
All samples were generated using Python and the PyFiglet library, which converts text into ASCII art using FIGlet fonts. The generation process was automated to ensure that every font consistently renders the full alphabet under the same conditions. |
|
|
|
|
|
### Notes |
|
|
|
|
|
- This dataset includes only the standard English alphabet and does not contain extended characters such as "ñ". |
|
|
- Outputs may vary in width and height depending on the font design. |
|
|
- The dataset is synthetic and intended primarily for research and experimental purposes. |
