Datasets:

Modalities:
Image
Text
Formats:
parquet
ArXiv:
Libraries:
Datasets
Dask
License:
MatteoOmenetti commited on
Commit
bf43f52
·
verified ·
1 Parent(s): f486472

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -12,7 +12,7 @@ size_categories:
12
  # SynthChartNet
13
 
14
  <div style="display: flex; justify-content: center; align-items: center;">
15
- <img src="https://cdn-uploads.huggingface.co/production/uploads/663e1254887b6f5645a0399f/Kgt6S5S_zPGGQ3IlmyRVB.png" alt="Chart Example" style="width: 500px; height: auto">
16
  </div>
17
 
18
  **SynthChartNet** is a multimodal dataset designed for training the **SmolDocling** model on chart-based document understanding tasks. It consists of **1,981,157** synthetically generated samples, where each image depicts a chart (e.g., line chart, bar chart, pie chart, stacked bar chart), and the associated ground truth is given in **OTSL** format.
 
12
  # SynthChartNet
13
 
14
  <div style="display: flex; justify-content: center; align-items: center;">
15
+ <img src="https://cdn-uploads.huggingface.co/production/uploads/663e1254887b6f5645a0399f/Kgt6S5S_zPGGQ3IlmyRVB.png" alt="Chart Example" style="width: 800px; height: auto">
16
  </div>
17
 
18
  **SynthChartNet** is a multimodal dataset designed for training the **SmolDocling** model on chart-based document understanding tasks. It consists of **1,981,157** synthetically generated samples, where each image depicts a chart (e.g., line chart, bar chart, pie chart, stacked bar chart), and the associated ground truth is given in **OTSL** format.