metadata
annotations_creators:
- crowdsourced
language:
- en
license: apache-2.0
multilinguality:
- monolingual
pretty_name: FlowGen
task_categories:
- visual-question-answering
- image-to-text
task_ids:
- visual-question-answering
- image-captioning
dataset_creators:
- Sorrystopper
size_categories:
- 10K<n<100K
source_datasets: []
tags:
- flowchart
- diagram-understanding
- visual-reasoning
- multimodal
🌟 FlowGen
FlowGen is a controllable flowchart synthesizer that generates diagrams with tunable structural features and supports multiple rendering styles.
📑 Dataset description
This dataset contains different types of renderer flowchart images with different difficulty levels.
| Types | Train (Easy) | Train (Medium) | Train (Hard) | Test (Graph Easy) | Test (Graph Medium) | Test (Graph Hard) | Test (Scanned Easy) | Test (Scanned Medium) | Test (Scanned Hard) |
|---|---|---|---|---|---|---|---|---|---|
| Mermaid | 960 | 960 | 960 | 360 | 360 | 360 | 360 | 360 | 360 |
| Graphviz | 960 | 960 | 960 | 360 | 360 | 360 | 360 | 360 | 360 |
| PlantUML | 960 | 960 | 960 | 360 | 360 | 360 | 360 | 360 | 360 |
| Diagrams | 960 | 960 | 960 | 360 | 360 | 360 | 360 | 360 | 360 |
| Total | 3840 | 3840 | 3840 | 1440 | 1440 | 1440 | 1440 | 1440 | 1440 |
Dataset Structure
Flowchart_Bench/
├── train/
│ ├── mermaid
| ├──no_disturbance
| ├──mermaid_1.mmd
| ├──mermaid_1.png
| ├──mermaid_2.mmd
| ├──mermaid_2.png
| └── ...
| ├──medium
| └── ...
│ ├── graphviz
| ├──easy
| ├──graphviz_1.dot
| ├──graphviz_1.png
| ├──graphviz_2.dot
| ├──graphviz_2.png
| └── ...
| ├──medium
| └── ...
│ ├── plantuml
| ├──easy
| ├──plantuml_1.puml
| ├──plantuml_1.png
| ├──plantumlz_2.puml
| ├──plantuml_2.png
| └── ...
| ├──medium
| └── ...
│ └── diagrams
| ├──easy
| ├──diagrams_1.dot
| ├──diagrams_1.png
| ├──diagrams_2.dot
| ├──diagrams_2.png
| └── ...
| ├──medium
| └── ...
├──...
└── README.md
For more information check here.