AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

FlowTok: Flowing Seamlessly Across Text and Image Tokens

Ju He, Qihang Yu, Qihao Liu, Liang-Chieh Chen

ICCV 2025 · Image generation › Autoregressive and latent generation · approach: Generative latent space tooling · 30 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
learn the text alignment target with an MLP instead of average pooling (ablation)FlowTok-B · COCO FID-30KFID-30K: 36.02 -> 29.14Improved
use contrastive text alignment loss instead of cosine similarity loss (ablation)FlowTok-B · COCO FID-30KFID-30K: 31.80 -> 29.14Improved
raise text alignment loss weight from 1.0 to 2.0 (ablation)FlowTok-B · COCO FID-30KFID-30K: 29.14 -> 30.59Hurt
scale FlowTok-XL (698M) to FlowTok-H (1.1B)FlowTok · MJHQ-30KFID: 7.68 -> 7.15Improved
scale FlowTok-XL (698M) to FlowTok-H (1.1B)FlowTok · COCO FID-30KFID-30K: 10.06 -> 9.67Improved
use compact 1D text-to-image flow instead of 2D CrossFlowFlowTok-H · text-to-image throughputthroughput (samples/s): 1.1 -> 18.2Improved
use compact 1D tokens instead of 2D CrossFlowFlowTok-H · training cost8 A100 days: 78.8 -> 26.1Improved
use 1D flow instead of CrossFlow on COCOFlowTok-H · COCO FID-30KFID-30K: 9.63 -> 9.67No clear effect
use direct image-to-text 1D flow instead of CrossFlowFlowTok-XL · COCO KarpathyCIDEr: 116.2 -> 117.0Improved
use direct image-to-text flow instead of SCD-NetFlowTok-XL · COCO KarpathyCIDEr: 118.0 -> 117.0Hurt
use 1D flow over text-as-condition Stable-Diffusion-2.1FlowTok-H · MJHQ-30KFID: 26.96 -> 7.15Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy