The Plug and Play of Language Models for Text-to-image Generation

Can Qin, Ning Yu, Chen Xing, Shu Zhang, Stefano Ermon, Yun Fu, Caiming Xiong, Ran Xu

Keywords: Text-to-Image Generation, Language Models, Efficiency

TL;DR: This paper introduces a new method to efficiently plug new language models to exiting text-to-image generation models as enhancement in scalability.

Abstract: Text-to-image (T2I) models enable controllable image generation through user-provided captions. A text encoder is typically used to map captions to a latent space, and it has been shown to be critical for model's performance. However, replacing or upgrading the text encoder in a T2I model is challenging due to the tight bond between the current encoder and the image decoder. It requires training the model from scratch, which can be prohibitively expensive. To address this problem, we introduce a more efficient approach to align a pre-trained language model with the latent space of an existing T2I model. We propose a Model Translation Network (MTN) and a new training objective to align the representation spaces of the two text encoders using only a corpus of unlabeled text. We empirically find that MTN can be trained efficiently and can boost the performance of existing T2I models by upgrading their text encoder. Moreover, we find that MTN can align multilingual language models such as XLM-Roberta, thus allowing existing T2I models to generate high-quality images from captions beyond English.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning