cjwbw / pix2struct

Pix2Struct: Screenshot Parsing as Pretraining for Visual Language Understanding

Demo API Examples Versions (e32d7748)

Run time and cost

Predictions run on Nvidia A100 (40GB) GPU hardware. Predictions typically complete within 2 seconds. The predict time for this model varies significantly based on the inputs.


Models included in this demo:

model_urls = {
    "textcaps": "google/pix2struct-textcaps-large", # Finetuned on TextCaps
    "screen2words": "google/pix2struct-screen2words-large", # Finetuned on Screen2Words
    "widgetcaption": "google/pix2struct-widget-captioning-large", # Finetuned on Widget Captioning (captioning a UI component on a screen)
    "infographics": "google/pix2struct-infographics-vqa-large", # Infographics
    "docvqa": "google/pix2struct-docvqa-large", # Visual question answering
    "ai2d": "google/pix2struct-ai2d-large", # Scienfic diagram


Pix2Struct is an image encoder - text decoder model that is trained on image-text pairs for various tasks, including image captionning and visual question answering. The full list of available models can be found on the Table 1 of the paper:

Table 1 - paper

The abstract of the model states that:

Visually-situated language is ubiquitous—sources range from textbooks with diagrams to web pages with images and tables, to mobile apps with buttons and forms. Perhaps due to this diversity, previous work has typically relied on domainspecific recipes with limited sharing of the underlying data, model architectures, and objectives. We present Pix2Struct, a pretrained image-to-text model for purely visual language understanding, which can be finetuned on tasks containing visually-situated language. Pix2Struct is pretrained by learning to parse masked screenshots of web pages into simplified HTML. The web, with its richness of visual elements cleanly reflected in the HTML structure, provides a large source of pretraining data well suited to the diversity of downstream tasks. Intuitively, this objective subsumes common pretraining signals such as OCR, language modeling, image captioning. In addition to the novel pretraining strategy, we introduce a variable-resolution input representation and a more flexible integration of language and vision inputs, where language prompts such as questions are rendered directly on top of the input image. For the first time, we show that a single pretrained model can achieve state-of-the-art results in six out of nine tasks across four domains: documents, illustrations, user interfaces, and natural images.


This model was originally contributed by Kenton Lee, Mandar Joshi et al. and added to the Hugging Face ecosystem by Younes Belkada.


If you want to cite this work, please consider citing the original paper:

  doi = {10.48550/ARXIV.2210.03347},

  url = {https://arxiv.org/abs/2210.03347},

  author = {Lee, Kenton and Joshi, Mandar and Turc, Iulia and Hu, Hexiang and Liu, Fangyu and Eisenschlos, Julian and Khandelwal, Urvashi and Shaw, Peter and Chang, Ming-Wei and Toutanova, Kristina},

  keywords = {Computation and Language (cs.CL), Computer Vision and Pattern Recognition (cs.CV), FOS: Computer and information sciences, FOS: Computer and information sciences},

  title = {Pix2Struct: Screenshot Parsing as Pretraining for Visual Language Understanding},

  publisher = {arXiv},

  year = {2022},

  copyright = {Creative Commons Attribution 4.0 International}