You can deploy this model and start earning money today!

Any2Poster/
any2poster-bench

text-to-image

document-question-answering

document-understanding

evaluation

Views

No views yet

Any2Poster Benchmark

A multi-modal benchmark for evaluating automated poster generation from diverse document types and content domains.

Dataset Description

Any2Poster Bench evaluates the task of generating a well-designed, content-faithful poster from a source document, where the source can be any of 8 modalities (PDF, DOCX, LaTeX, Markdown, Jupyter Notebook, PPTX, URL, Video) across 5 content domains (research, news, educational, business, fiction). The full benchmark… See the full description on the dataset page: https://huggingface.co/datasets/Any2Poster/any2poster-bench.