lilyzhng/uigen-ui-code-gen-full

UIGEN UI/UX Code Generation Dataset

This dataset contains UI/UX code generation examples formatted for training code generation models. Each example consists of a task description and the corresponding HTML/CSS code implementation using Tailwind CSS.

Dataset Structure

The dataset has a single text column containing formatted prompts and completions:

# Task: Generate HTML/CSS code using Tailwind CSS
# Requirements: [specific requirements]

[HTML/CSS code implementation]

Splits

• —train: 715 examples
• —validation: 10 examples
• —test: 80 examples

Usage

from datasets import load_dataset

dataset = load_dataset("lilyzhng/uigen-ui-code-gen-full")

# Access training data
train_data = dataset["train"]
print(train_data[0]["text"])

Use with Modal Training

# In your Modal training script
from datasets import load_dataset

dataset = load_dataset("lilyzhng/uigen-ui-code-gen-full", split="train")

Source

This dataset is derived from the UIGEN dataset, formatted specifically for code generation training with base (non-instruct) models.

Citation

If you use this dataset, please cite the original UIGEN paper:

@article{uigen2023,
  title={UIGEN: A Dataset for UI Code Generation},
  author={...},
  year={2023}
}