Papers › Fluid: Scaling Autoregressive Text-to-image Generative Models with Continuous Tokens

Fluid: Scaling Autoregressive Text-to-image Generative Models with Continuous Tokens

17 Oct 2024arXiv:2410.13863archive 2025-07-28

Lijie Fan, Tianhong Li, Siyang Qin, Yuanzhen Li, Chen Sun, Michael Rubinstein, Deqing Sun, Kaiming He, Yonglong Tian

Scaling up autoregressive models in vision has not proven as beneficial as in large language models. In this work, we investigate this scaling problem in the context of text-to-image generation, focusing on two critical factors: whether models use discrete or continuous tokens, and whether tokens are generated in a random or fixed raster order using BERT- or GPT-like transformer architectures. Our empirical results show that, while all models scale effectively in terms of validation loss, their evaluation performance -- measured by FID, GenEval score, and visual quality -- follows different trends. Models based on continuous tokens achieve significantly better visual quality than those using discrete tokens. Furthermore, the generation order and attention mechanisms significantly affect the GenEval score: random-order models achieve notably better GenEval scores compared to raster-order models. Inspired by these findings, we train Fluid, a random-order autoregressive model on continuous tokens. Fluid 10.5B model achieves a new state-of-the-art zero-shot FID of 6.16 on MS-COCO 30K, and 0.69 overall score on the GenEval benchmark. We hope our findings and results will encourage future efforts to further bridge the scaling gap between vision and language models.

PaperPDFCode

In Syntology View this paper on Syntology: its repositories, every harvested function with whether it ran, its licence and the call to fetch it.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

xianfengwu01/lightgen mentioned on GitHubpytorchMIT report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Image GenerationText to Image GenerationText-to-Image Generation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Text-to-Image Generation GenEval Fluid (10.5B) Color Attri. 0.51 #13 of 20 Archive leaderboard report
Text-to-Image Generation GenEval Fluid (10.5B) Colors 0.80 #13 of 20 Archive leaderboard report
Text-to-Image Generation GenEval Fluid (10.5B) Counting 0.63 #13 of 20 Archive leaderboard report
Text-to-Image Generation GenEval Fluid (10.5B) Overall 0.69 #13 of 20 Archive leaderboard report
Text-to-Image Generation GenEval Fluid (10.5B) Position 0.39 #13 of 20 Archive leaderboard report
Text-to-Image Generation GenEval Fluid (10.5B) Single Obj. 0.96 #13 of 20 Archive leaderboard report
Text-to-Image Generation GenEval Fluid (10.5B) Two Obj. 0.83 #13 of 20 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

AttentionSoftmax

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections