Flame-Eval-React is the first benchmarking dataset specifically designed to evaluate the accuracy, functionality, and visual fidelity of vision-language models (VLMs) for React code generation.
This dataset includes manually curated React components that serve as the gold standard for evaluating the performance of image-to-code translation models.
Here we provide the code guidance… See the full description on the dataset page:
https://huggingface.co/datasets/Flame-Code-VLM/Flame-Eval-React.
