Datasets:
image image | svg string |
|---|---|
<svg width="300" height="200" xmlns="http://www.w3.org/2000/svg" shape-rendering="crispEdges"><rect x="238" y="59" width="5" height="3" fill="#5e6f70"/><rect x="238" y="56" width="5" height="3" fill="#425252"/><rect x="238" y="53" width="5" height="3" fill="#232f30"/><rect x="220" y="46" width="5" height="4" fill="#676... | |
"<svg width=\"300\" height=\"200\" xmlns=\"http://www.w3.org/2000/svg\" shape-rendering=\"crispEdges(...TRUNCATED) | |
"<svg width=\"300\" height=\"200\" xmlns=\"http://www.w3.org/2000/svg\" shape-rendering=\"crispEdges(...TRUNCATED) | |
"<svg width=\"300\" height=\"200\" xmlns=\"http://www.w3.org/2000/svg\" shape-rendering=\"crispEdges(...TRUNCATED) | |
"<svg width=\"300\" height=\"200\" xmlns=\"http://www.w3.org/2000/svg\" shape-rendering=\"crispEdges(...TRUNCATED) | |
"<svg width=\"300\" height=\"200\" xmlns=\"http://www.w3.org/2000/svg\" shape-rendering=\"crispEdges(...TRUNCATED) | |
"<svg width=\"300\" height=\"200\" xmlns=\"http://www.w3.org/2000/svg\" shape-rendering=\"crispEdges(...TRUNCATED) | |
"<svg width=\"300\" height=\"200\" xmlns=\"http://www.w3.org/2000/svg\" shape-rendering=\"crispEdges(...TRUNCATED) | |
"<svg width=\"300\" height=\"200\" xmlns=\"http://www.w3.org/2000/svg\" shape-rendering=\"crispEdges(...TRUNCATED) | |
"<svg width=\"300\" height=\"200\" xmlns=\"http://www.w3.org/2000/svg\" shape-rendering=\"crispEdges(...TRUNCATED) |
End of preview. Expand in Data Studio
This dataset is designed to let Vision Language models transform standarrd PNG images into SVG code. The algorithm used to generate this code doesn't capture every single quality in the image, and downscales the resolution to make it more token efficient.
Example
Original Image
Image Converted to SVG
This output uses 61,400 tokens according to the OpenAI Tokenizer.
- Downloads last month
- 37