Pith. sign in

REVIEW 5 cited by

1.58-bit FLUX

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2412.18653 v1 pith:XHRWC26I submitted 2024-12-24 cs.CV cs.AIcs.LG

Chenglin Yang , Celong Liu , Xueqing Deng , Dongwon Kim , Xing Mei , Xiaohui Shen , Liang-Chieh Chen This is my paper · ORCID
classification cs.CVcs.AIcs.LG
keywords fluxmodelgenerationinferencemaintainingreductionwhileaccess
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

We present 1.58-bit FLUX, the first successful approach to quantizing the state-of-the-art text-to-image generation model, FLUX.1-dev, using 1.58-bit weights (i.e., values in {-1, 0, +1}) while maintaining comparable performance for generating 1024 x 1024 images. Notably, our quantization method operates without access to image data, relying solely on self-supervision from the FLUX.1-dev model. Additionally, we develop a custom kernel optimized for 1.58-bit operations, achieving a 7.7x reduction in model storage, a 5.1x reduction in inference memory, and improved inference latency. Extensive evaluations on the GenEval and T2I Compbench benchmarks demonstrate the effectiveness of 1.58-bit FLUX in maintaining generation quality while significantly enhancing computational efficiency.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 5 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Investigation of Accuracy and Bias in Face Recognition Trained with Synthetic Data

    cs.CV 2025-07 conditional novelty 6.0 of 10

    Balanced synthetic face data from Stable Diffusion v3.5 reduces racial bias in face recognition models but does not yet match real-data accuracy on hard benchmarks.

  2. Retrieval Visual Contrastive Decoding to Mitigate Object Hallucinations in Large Vision-Language Models

    cs.CV 2025-05 conditional novelty 6.0 of 10

    RVCD uses YOLO detections and retrieved single-concept AI images to adjust LVLM logits at decode time, cutting CHAIR hallucination rates by roughly half versus prior contrastive decoding baselines.

  3. Grouping First, Attending Smartly: Training-Free Acceleration for Diffusion Transformers

    cs.CV 2025-05 conditional novelty 6.0 of 10

    GRAT accelerates pretrained diffusion transformers by grouping tokens and restricting each group's attention to neighboring blocks or criss-cross rows and columns, achieving large speedups with near-full-attention quality.

  4. ID-Card Synthetic Generation: Toward a Simulated Bona fide Dataset

    cs.CV 2025-08 conditional novelty 5.0 of 10

    Synthetic ID cards made with Stable Diffusion and hybrid face/text compositing are recognized as bona fide by a commercial PAD system and could augment scarce real training data.

  5. DFVEdit: Conditional Delta Flow Vector for Zero-shot Video Editing

    cs.CV 2025-06 conditional novelty 4.0 of 10

    DFVEdit edits videos by iteratively subtracting a conditional delta flow vector, the difference between the model's predictions under the target and source prompts, from the latent representation of the source video.

Pith tools