Pickapic

Welcome to Pick a Pic, an open-sourced application that plays a crucial role in academic research on AI-generated images. Our app focuses on collecting valuable human feedback for AI-generated images, contributing to the advancement of AI technology.

With Pick a Pic, users have the opportunity to generate images from text and provide rankings based on their preferences. As part of our commitment to transparency, all code, data, and models resulting from our research project are open-sourced and accessible to the public. This allows AI enthusiasts, researchers, and developers to explore and leverage our resources.

Prior to generating images and participating in this fascinating project, users need to authenticate via their Google accounts. By doing so, we ensure the integrity of the data and maintain the quality of the research outcomes.

Our mission is ambitious yet essential: to construct the most expansive and publicly available human-feedback dataset for text-to-image technology. Currently, our app utilizes Stable Diffusion 2.1 and ProtogenX3.4 models, empowering users to engage with cutting-edge AI advancements.

Rest assured, our data is consistently updated, and new information is periodically released into the public domain. We owe thanks to the Google TRC Program for their support in making this project possible.