# Efficient Neural Architecture for Text-to-Image Synthesis

arXiv: https://arxiv.org/abs/2004.11437 · HTML: https://mlsanity.com/paper/2004.11437/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Apr 2020.

## Built on by (most cited first, 11 of 11)

- [Vector Quantized Diffusion Model for Text-to-Image Synthesis](https://mlsanity.com/paper/2111.14822/) (2021)
- [DE-FAKE: Detection and Attribution of Fake Images Generated by Text-to-Image Generation Models](https://mlsanity.com/paper/2210.06998/) (2022)
- [Towards Implicit Text-Guided 3D Shape Generation](https://mlsanity.com/paper/2203.14622/) (2022)
- [Improved Vector Quantized Diffusion Models](https://mlsanity.com/paper/2205.16007/) (2022)
- [Adversarial Text-to-Image Synthesis: A Review](https://mlsanity.com/paper/2101.09983/) (2021)
- [Multimodal Image Synthesis and Editing: The Generative AI Era](https://mlsanity.com/paper/2112.13592/) (2021)
- [ISS: Image as Stepping Stone for Text-Guided 3D Shape Generation](https://mlsanity.com/paper/2209.04145/) (2022)
- [Bi-LORA: A Vision-Language Approach for Synthetic Image Detection](https://mlsanity.com/paper/2404.01959/) (2024)
- [CRD-CGAN: Category-Consistent and Relativistic Constraints for Diverse Text-to-Image Generation](https://mlsanity.com/paper/2107.13516/) (2021)
- [DreamStone: Image as Stepping Stone for Text-Guided 3D Shape Generation](https://mlsanity.com/paper/2303.15181/) (2023)
- [Looks Like Magic: Transfer Learning in GANs to Generate New Card Illustrations](https://mlsanity.com/paper/2205.14442/) (2022)

## Abstract (arXiv)

Text-to-image synthesis is the task of generating images from text descriptions. Image generation, by itself, is a challenging task. When we combine image generation and text, we bring complexity to a new level: we need to combine data from two different modalities. Most of recent works in text-to-image synthesis follow a similar approach when it comes to neural architectures. Due to aforementioned difficulties, plus the inherent difficulty of training GANs at high resolutions, most methods have adopted a multi-stage training strategy. In this paper we shift the architectural paradigm currently used in text-to-image methods and show that an effective neural architecture can achieve state-of-the-art performance using a single stage training with a single generator and a single discriminator. We do so by applying deep residual networks along with a novel sentence interpolation strategy that enables learning a smooth conditional space. Finally, our work points a new direction for text-to-image research, which has not experimented with novel neural architectures recently.