CogView3：通過中繼擴散實現更精細更快速的文本到圖像生成

CogView3: Finer and Faster Text-to-Image Generation via Relay Diffusion

March 8, 2024

作者: Wendi Zheng, Jiayan Teng, Zhuoyi Yang, Weihan Wang, Jidong Chen, Xiaotao Gu, Yuxiao Dong, Ming Ding, Jie Tang

cs.AI

摘要

最近在文本到圖像生成系統方面的進展主要是由擴散模型推動的。然而，單階段文本到圖像擴散模型仍然面臨著計算效率和圖像細節的改進方面的挑戰。為了應對這個問題，我們提出了CogView3，這是一個創新的級聯框架，可以增強文本到圖像擴散的性能。CogView3是第一個在文本到圖像生成領域實現中繼擴散的模型，通過首先創建低分辨率圖像，然後應用基於中繼的超分辨率來執行任務。這種方法不僅產生具有競爭力的文本到圖像輸出，還大大降低了訓練和推理成本。我們的實驗結果表明，CogView3在人類評估方面比當前最先進的開源文本到圖像擴散模型SDXL表現優異，性能提高了77.0％，同時推理時間僅需SDXL的約1/2。CogView3的精煉變體實現了可比的性能，同時只利用SDXL推理時間的1/10。

English

Recent advancements in text-to-image generative systems have been largely driven by diffusion models. However, single-stage text-to-image diffusion models still face challenges, in terms of computational efficiency and the refinement of image details. To tackle the issue, we propose CogView3, an innovative cascaded framework that enhances the performance of text-to-image diffusion. CogView3 is the first model implementing relay diffusion in the realm of text-to-image generation, executing the task by first creating low-resolution images and subsequently applying relay-based super-resolution. This methodology not only results in competitive text-to-image outputs but also greatly reduces both training and inference costs. Our experimental results demonstrate that CogView3 outperforms SDXL, the current state-of-the-art open-source text-to-image diffusion model, by 77.0\% in human evaluations, all while requiring only about 1/2 of the inference time. The distilled variant of CogView3 achieves comparable performance while only utilizing 1/10 of the inference time by SDXL.