Pith. sign in

REVIEW 5 major objections 7 minor 88 references

ADIEE: Automatic Dataset Creation and Scorer for Instruction-Guided Image Editing Evaluation

T0 review · 5 major / 7 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read This paper claims that a VLM judge for instruction-guided image editing, trained entirely on a 100K-sample dataset labeled by automatic heuristics rather than human annotation, outperforms all open-source VLMs and Gemini-Pro 1.5 on…

desk verdict Useful training recipe for an editing judge, but the test set overlaps the training data and the headline gains need a clean-room check. read the letter →

arxiv 2507.07317 v2 pith:NJ4WUWYH submitted 2025-07-09 cs.CV

Sherry X. Chen , Yi Wei , Luowei Zhou , Suren Kumar This is my paper
classification cs.CV
keywords instruction-guidedimageeditingautomaticevaluationvision-languagemodeljudgerewardsynthetictrainingdatahumanalignmentmulti-turnCLIPdirectionalsimilarity
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to show that a reliable judge for instruction-guided image editing can be built without any human preference labels. It creates over 100,000 training examples automatically, by treating existing editing datasets and models as sources of successful, partially successful, and failed edits, and labels them with simple heuristics. A vision-language model fine-tuned on those labels scores edits more consistently with human raters than any open-source VLM and than Gemini-Pro 1.5 on three benchmarks. The same scorer also serves as a reward model that raises the average evaluation score of the MagicBrush editing model from 5.90 to 6.43. A sympathetic reader would care because cheap, transparent automatic evaluation would replace costly human annotation and opaque proprietary judges.

What carries the argument

The central mechanism is a fully automatic labeling pipeline. For single-turn edits, it scores each generated output by comparing it to the input and ground-truth images: CLIP directional similarity below 0.2 marks partial success, CLIP-I and DINO-I below the dataset's fifth percentile mark failure, and all outputs from DiffEdit, Pix2Pix-Zero, SDEdit, and Text2LIVE are treated as failures because those models produce inconsistent edits. For multi-turn sequences, the score function f in Eq. 4 assigns 0 to earlier images, the fraction $(k-j_1)/(j_2-j_1)$ to intermediate turns, 1 to the ground truth, and 0.5 to over-edited later images. These examples fine-tune a LLaVA-NeXT-8B whose vocabulary is extended with a special [SCORE] token, and a small MLP decoder turns that token's embedding into the numeric rating; LoRA keeps the fine-tune efficient.

What would settle it

Take a random sample of the constructed 100K training examples, have several human raters assign scores, and compute the agreement between the heuristic label and the human average. Alternatively, feed the scorer pairs where the heuristic labels are demonstrably wrong—such as a high-quality DiffEdit output or a multi-turn intermediate image that already satisfies the full instruction—and check whether the scorer still follows the heuristic rather than the image content.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that instruction-guided editing data implicitly contains evaluation signal: the original input image is a failed edit, the ground-truth output is a perfect edit, and intermediate images in multi-turn sequences are partial edits whose score is the fraction of instructions completed. The paper augments this with outputs from nine editing models, labels low-quality-model outputs as failures and uses CLIP directional similarity thresholds to flag partial successes, then trains a LLaVA-NeXT-8B with a special [SCORE] token whose embedding is decoded by an MLP into a number. Trained this way, the scorer achieves a 0.4734 Spearman correlation with human ratings on AURORA-Bench point-wise and 59.96 percent pair-wise accuracy on GenAI-Bench, surpassing all open-source VLMs and Gemini-Pro 1.5; on ImagenHub it beats all open-source VLMs and Gemini-Pro 1.5 while trailing only GPT-4o. As a reward model, it lifts the editing model's average ImagenHub score from 5.90 to 6.43, a gain the paper reports as +8.98 percent.

Load-bearing premise

The load-bearing premise is that the heuristic labels—DiffEdit, Pix2Pix-Zero, SDEdit, and Text2LIVE outputs always fail, CLIP-D below 0.2 means partial success, and an intermediate turn's score is the completed fraction of instructions—match what human raters would say; if any of these rules is wrong in the bulk of samples, the scorer's human alignment is an artifact of the labeling scheme.

Editorial extensions

If this is right

  • A VLM judge trained entirely on automatically generated labels can beat strong open-source and proprietary evaluators, so human-annotated evaluation datasets are not a prerequisite for good alignment.
  • The scorer can pick the best edit among candidates, enabling automated model selection and leaderboard ranking without manual inspection.
  • Using the scorer as a reward signal improves the editing model's output quality by a measurable margin (+8.98 percent average score), so automatic judges can substitute for human feedback in fine-tuning.
  • Because the dataset creation recipe relies only on editing datasets and generic CLIP/DINO thresholds, it should transfer to new editing methods and benchmarks without new annotation effort.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • This reader infers that the method's ceiling is set by the heuristics: if CLIP and DINO miss fine-grained failures that humans notice, the scorer will inherit that blind spot.
  • This reader infers the reward-model gain may partly reflect the scorer rewarding the same coarse semantics it was trained to detect, so improvements on human preference could be smaller than the +8.98 percent metric suggests.
  • This reader infers the same recipe could be applied to other instruction-following generative domains, such as video or 3D editing, wherever a ground-truth anchor and a multi-turn sequence exist.
  • A testable extension would be to train the scorer with a small human-annotated correction set on top of the heuristics and measure whether alignment improves further.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 7 minor

Summary. The paper introduces ADIEE, a pipeline that automatically constructs a large training dataset for instruction-guided image editing evaluation by (i) generating edited outputs with nine editing methods on MagicBrush and Emu-Edit images and assigning heuristic scores, and (ii) extracting multi-turn edit sequences from SEED-Data-Edit-Part3. The authors fine-tune LLaVA-NeXT-8B with a special score token and an MLP decoder to predict a 0–10 quality score, and evaluate the resulting scorer on ImagenHub, GenAI-Bench, and AURORA-Bench. They also use the scorer as a reward model to fine-tune MagicBrush, reporting an improvement in the scorer's own evaluation on ImagenHub. The paper claims state-of-the-art alignment with human ratings across all benchmarks.

Significance. If the claims hold, ADIEE offers a scalable, transparent alternative to proprietary VLM judges, and a practical reward model for improving editing models. The paper releases code and models and includes a human preference study, which are strengths. However, the central generalization claim is currently threatened by (a) likely overlap between ADIEE's training sources and AURORA-Bench, (b) inconsistent numbers across the main text and appendix, and (c) a protocol-sensitivity issue where a 1-shot GPT-4o baseline exceeds the proposed scorer on AURORA-Bench point-wise. These issues are load-bearing for the headline claims and require substantial revision.

major comments (5)
  1. [§4.1 and Appendix C] The training set is built from 9,935 MagicBrush samples, 1,100 instructions on Emu-Edit test images, and 21,382 SEED-Data-Edit-Part3 sequences (Section 4.1). Appendix C states that AURORA-Bench point-wise and pair-wise samples are collected from MagicBrush and Emu-Edit, among other sources. The paper reports no deduplication or overlap analysis between ADIEE training data and the AURORA-Bench evaluation sets. If the same images or instructions appear in both, the reported Spearman correlation (0.4734) and pairwise accuracy gains could reflect memorization rather than generalized evaluation ability. The authors should quantify the overlap and re-evaluate on the subset of AURORA-Bench that does not stem from training sources (e.g., video-derived samples), and report the correlation on that subset separately.
  2. [Tables 2, 5, and Appendix Table 7] There are major numerical inconsistencies for the same ADIEE scorer across the manuscript. Table 2 reports GenAI-Bench accuracy 59.96 and AURORA-Bench pair-wise 55.56; Table 5 reports 59.41 and 52.38 for the same conditions; Appendix Table 7 reports 59.41 and 52.88. The abstract's +4.75% gain on AURORA pair-wise is only consistent with the main-text Table 2 value of 55.56, not with the appendix value. Similarly, Section 4.4 reports that reward-conditioned MagicBrush improves the ImagenHub score from 5.90 to 6.43, while Section 4.5 mentions a score of 6.67 for 'our method' and 4.27 for the without-reward-conditioning variant. These discrepancies must be reconciled and the source of variation (e.g., different evaluation subsets, random seeds, or post-processing) must be stated explicitly.
  3. [Appendix Table 6 and Abstract] Appendix Table 6 shows that GPT-4o with a 1-shot in-context example reaches 0.4779 Spearman correlation on AURORA-Bench point-wise, which is higher than the proposed ADIEE scorer's 0.4734. The main text, however, compares against zero-shot baselines and claims a +17.24% improvement over state-of-the-art. This makes the SOTA claim protocol-sensitive. The authors should either include the 1-shot baselines in the main tables, or justify why zero-shot is the appropriate comparison and temper the abstract and conclusion accordingly.
  4. [§3.4 and §4.4, Tables 3–4] The reward-model improvement on ImagenHub (+8.98%, from 5.90 to 6.43) is measured by the ADIEE scorer itself, which is the very model being proposed. The human pairwise study in Table 4 provides some external validation, but it is not directly linked to the ImagenHub score improvement and lacks any significance test. In addition, the count column in Table 4 (914 + 752 + 1190 = 2856) does not match the stated protocol of 56 participants x 50 comparisons (2800). The authors should report human-rated quality scores for the baseline and fine-tuned models, or at least provide a statistical test on the preference counts, and correct the count mismatch.
  5. [§3.1 and Eq. (4)] The heuristic score assignments are strong and untested assumptions. Section 3.1 assigns score 0 to all outputs from DiffEdit, Pix2Pix-Zero, SDEdit, and Text2LIVE, and uses a CLIP-D threshold of tau = 0.2 to identify partial edits. Equation (4) assigns intermediate multi-turn edits a score proportional to the fraction of completed instructions, and assigns 0.5 to over-edited outputs. These rules are not validated against human perception. Since ImagenHub and AURORA-Bench include outputs from some of these same editing methods, the scorer could partially align with human ratings by learning method-specific biases. A small human-annotation study on a sample of the generated training data, or an ablation showing that the results are stable to changes in these thresholds, is needed to support the claim that the labels track human judgment.
minor comments (7)
  1. [Abstract and §2.3] The introduction states that 'no public training datasets exist to fine-tune open-source VLMs' for image-editing evaluation, but Related Work Section 2.3 cites OmniEdit and RewardEdit20K, which are public training datasets built with proprietary models. The claim should be qualified to refer to human-annotated or open-label datasets.
  2. [§3.1] The sentence 'we apply a CLIP-D threshold of tau_CLIP-D = 0.2, which is used in IP2P to keep successful edits, and assign this score to samples below the threshold' is ambiguous: in IP2P, a higher CLIP-D typically indicates a more successful edit, so it is unclear whether score 0.5 is assigned to samples below or above the threshold. Please clarify the intended direction.
  3. [Table 4] The total number of human preference comparisons (2856) does not match the stated 56 participants each comparing 50 pairs (2800). Please verify the numbers or the procedure.
  4. [§4.5] The section title 'Alation Study' should be 'Ablation Study'.
  5. [Figure 8 caption] The caption reads 'face are blocked'; this should be 'faces are blocked'.
  6. [Appendix Table 7] Several entries report 0.0 for Qwen2.5-VL (1-shot) on GenAI-Bench and for Fuyu, CogVLM, and OpenFlamingo on AURORA-Bench pair-wise; these likely reflect parsing or formatting failures rather than true accuracies and should be marked as unavailable instead of 0.0.
  7. [§4.4] The reward-feedback-learning result (score 6.30) is only mentioned in the text; presenting it in a table alongside the reward-conditioning result would improve reproducibility.

Circularity Check

2 steps flagged · score 5.0 of 10

Partial circularity: reward-model gain is self-scored, and AURORA-Bench overlaps ADIEE training sources.

  1. fitted input called prediction [Section 3.4 (Eq. 5) and Section 4.4]
    "we generate the pristine edited output and its evaluation score s (normalized to 0-10) using the proposed scorer. The reward feedback learning loss is formulated as: Lreward = 10− s ... According to our evaluation scorer, the MagicBrush baseline [83] achieves an average score of 5.90 on ImagenHub samples [38], while our model fine-tuned with reward conditioning scores 6.43 (+8.98%)."

    The same ADIEE scorer supplies score s in the training loss Lreward = 10 - s, so the editing model is explicitly optimized to maximize ADIEE's scores. Section 4.4 then measures the claimed improvement with that same scorer, so the +8.98% and +6.78% gains are partly self-consistency results: the model was trained to please the scorer and is then judged by the same scorer. The independent human pairwise preference study (Table 4, 41.67% vs 32.00%) provides external evidence, so the circularity is partial rather than total, but the headline numeric gains are not independent evidence of improved editing quality.

  2. other [Section 4.1 and Appendix C]
    "We use 9,935 samples from MagicBrush [83] and 1,100 additional instructions for global/style edits on Emu-Edit test images [60], generating 93,915 outputs via nine text-guided editing methods. ... Samples from AURORA-Bench are collected from multiple sources: MagicBrush [83], Action-Genome [31], Something-Something [24], Epic-Kitchen [17], Kubric [25], CLEVR [33], WhatsUp [34], and Emu-Edit [60]."

    Training data is built from MagicBrush samples and Emu-Edit test images, and AURORA-Bench is explicitly collected from MagicBrush and Emu-Edit among other sources. With no reported deduplication or source-based train/test split, AURORA-Bench samples can coincide with ADIEE training samples, letting the scorer memorize expected scores for seen images/instructions rather than learn general evaluation. This affects the point-wise correlation (0.4734) and pairwise accuracy (55.56) on AURORA-Bench, which are central to the claim of outperforming all baselines. Because GenAI-Bench/ImagenHub results and the human preference study are external, this is a contamination risk rather than a definitional circularity, but it is load-bearing.

full rationale

The main scorer claim is not definitionally circular: ADIEE is trained on heuristically labeled data and tested against human-annotated benchmarks (ImagenHub, GenAI-Bench, and the human preference study), and the scorer's gains over its own CLIP/DINO label sources show it learned beyond the heuristics. However, two load-bearing steps weaken the strongest claims. First, the reward-model improvement is measured by the same scorer used as the training reward, so the headline +8.98% is partly self-measurement, though the external human preference study mitigates this. Second, AURORA-Bench is assembled from the same source datasets (MagicBrush, Emu-Edit) used to construct ADIEE training data, and no overlap analysis is reported, so the AURORA-Bench numbers may be inflated by memorization. These are not self-citation or uniqueness-theorem issues; no load-bearing self-citations appear. Overall partial circularity, score 5.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The central claim rests on a set of labeling heuristics and assumptions about the relationship between automated similarity metrics and human judgment.

free parameters (4)
  • CLIP-D threshold tau = 0.2
    Used to assign score 0.5 for MagicBrush and AURORA outputs below the threshold (Sec 3.1).
  • 5th percentile CLIP-I and DINO-I thresholds = data-dependent percentiles
    Used to assign score 0 to generated outputs that are visually far from ground truth (Sec 3.1).
  • lambda_text and lambda_score = 1 and 10
    Scorer training loss weights (Appendix B).
  • lambda_reward = 0.001
    Weight for reward feedback learning loss (Appendix B).
assumptions (6)
  • domain assumption CLIP and DINO similarity metrics capture semantic and visual similarity relevant to edit success.
    Used in Sec 3.1 to identify negative samples and thresholds.
  • domain assumption Ground-truth edited images are successful, input images are failed edits.
    Core assumption of Sec 3.1.
  • ad hoc to paper All outputs from DiffEdit, P2P-0, SDEdit, and T2L are failures.
    Blanket score 0 assignment, acknowledged to be imperfect (Sec 3.1).
  • ad hoc to paper Intermediate edit success is proportional to fraction of instructions completed.
    Eq 4 in Sec 3.2.
  • ad hoc to paper Over-edited outputs receive score 0.5.
    Eq 4 in Sec 3.2.
  • domain assumption The fine-tuned captioning VLM generates input/target prompts aligned with edit intent.
    Required for prompt-guided editing models; validated only qualitatively (Fig 4).

how reviews work

0 comments
Cite this review

Pith. "Pith review of ADIEE: Automatic Dataset Creation and Scorer for Instruction-Guided Image Editing Evaluation." pith.science (2026). https://pith.science/paper/NJ4WUWYH

@misc{pith2026250707317,
  author       = {Pith},
  title        = {Pith review of: ADIEE: Automatic Dataset Creation and Scorer for Instruction-Guided Image Editing Evaluation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NJ4WUWYH}},
  note         = {Machine review of arXiv:2507.07317}
}
read the original abstract

Recent advances in instruction-guided image editing underscore the need for effective automated evaluation. While Vision-Language Models (VLMs) have been explored as judges, open-source models struggle with alignment, and proprietary models lack transparency and cost efficiency. Additionally, no public training datasets exist to fine-tune open-source VLMs, only small benchmarks with diverse evaluation schemes. To address this, we introduce ADIEE, an automated dataset creation approach which is then used to train a scoring model for instruction-guided image editing evaluation. We generate a large-scale dataset with over 100K samples and use it to fine-tune a LLaVA-NeXT-8B model modified to decode a numeric score from a custom token. The resulting scorer outperforms all open-source VLMs and Gemini-Pro 1.5 across all benchmarks, achieving a 0.0696 (+17.24%) gain in score correlation with human ratings on AURORA-Bench, and improving pair-wise comparison accuracy by 4.03% (+7.21%) on GenAI-Bench and 4.75% (+9.35%) on AURORA-Bench, respectively, compared to the state-of-the-art. The scorer can act as a reward model, enabling automated best edit selection and model fine-tuning. Notably, the proposed scorer can boost MagicBrush model's average evaluation score on ImagenHub from 5.90 to 6.43 (+8.98%). Our code and models are available at https://github.com/SherryXTChen/ADIEE.git.

Figures

Figures reproduced from arXiv: 2507.07317 by the authors.

Figure 1
Figure 1. Top: We fine-tune LLaVA-Next-8B [46] on an image editing evaluation dataset automatically generated using our proposed methods to achieve good alignment with human judgment (faces are blocked due to privacy con￾cerns). Bottom: The trained image editing quality scorer can also function as a reward model to fine-tune and en￾hance the editing capabilities of MagicBrush [83]. the growing need for effective automated eva… view at source ↗
Figure 2
Figure 2. ADIEE consists of two image editing evaluation training data generation approaches using text-guided image [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. We use 9 text-guided image editing methods - CycleDiffusion (CycleDiff) [ [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Compared to the input/target prompt generated by a baseline VLM [ [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: For each sequence of edited images in a multi [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: To train a VLM [47] as image editing evalua￾tor, we concatenate the input and the output image and send them to the model along with the text query. The embed￾ding corresponding to the special token [SCORE] is de￾coded into the final image editing quality evaluation sc…
Figure 7
Figure 7. Figure 7: Evaluation examples (faces are blocked for privacy concern) from GPT-4o [ [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Image editing examples (face are blocked for privacy concern) from the MagicBrush editing model [ [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 9
Figure 9. Figure 9: We use 9 text-guided image editing methods - CycleDiffusion (CycleDiff) [ [PITH_FULL_IMAGE:figures/full_fig_p014_9.png]
Figure 10
Figure 10. Figure 10: More evaluation examples (faces are blocked due to privacy concerns) from GPT-4o [ [PITH_FULL_IMAGE:figures/full_fig_p016_10.png]
Figure 11
Figure 11. Figure 11: More mage editing examples (faces are blocked due to privacy concerns) from the MagicBrush editing model [ [PITH_FULL_IMAGE:figures/full_fig_p017_11.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

88 extracted references · 40 canonical work pages

  1. [1]

    Gpt-4 technical report

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ah- mad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774 ,

    arXiv

  2. [2]

    Cos stable diffusion xl 1.0 and cos stable dif- fusion xl 1.0 edit, 2024

    Stability AI. Cos stable diffusion xl 1.0 and cos stable dif- fusion xl 1.0 edit, 2024. https://huggingface.co/ stabilityai/cosxl. 15

    2024

  3. [3]

    Claude 3.5 sonnet model card addendum

    Anthropic. Claude 3.5 sonnet model card addendum. Tech- nical report, June 2024. 3

    2024

  4. [4]

    Openflamingo: An open-source frame- work for training large autoregressive vision-language mod- els

    Anas Awadalla, Irena Gao, Josh Gardner, Jack Hessel, Yusuf Hanafy, Wanrong Zhu, Kalyani Marathe, Yonatan Bitton, Samir Gadre, Shiori Sagawa, Jenia Jitsev, Simon Kornblith, Pang Wei Koh, Gabriel Ilharco, Mitchell Wortsman, and Ludwig Schmidt. Openflamingo: An open-source frame- work for training large autoregressive vision-language mod- els. arXiv preprint...

    arXiv 2023

  5. [5]

    Qwen-vl: A versatile vision-language model for un- derstanding, localization, text reading, and beyond

    Jinze Bai, Shuai Bai, Shusheng Yang, Shijie Wang, Sinan Tan, Peng Wang, Junyang Lin, Chang Zhou, and Jingren Zhou. Qwen-vl: A versatile vision-language model for un- derstanding, localization, text reading, and beyond. arXiv preprint arXiv:2308.12966, 2023. 2

    arXiv 2023

  6. [6]

    Text2live: Text-driven layered image and video editing

    Omer Bar-Tal, Dolev Ofri-Amar, Rafail Fridman, Yoni Kas- ten, and Tali Dekel. Text2live: Text-driven layered image and video editing. In European Conference on Computer Vision, pages 707–723. Springer, 2022. 3, 4, 5, 14

    2022

  7. [7]

    Introducing our multimodal models, 2023

    Rohan Bavishi, Erich Elsen, Curtis Hawthorne, Maxwell Nye, Augustus Odena, Arushi Somani, and Sa ˘gnak Tas ¸ırlar. Introducing our multimodal models, 2023. 2

    2023

  8. [8]

    Is clip the main roadblock for fine-grained open-world perception? In 2024 International Conference on Content-Based Multimedia Indexing (CBMI), pages 1–8

    Lorenzo Bianchi, Fabio Carrara, Nicola Messina, and Fab- rizio Falchi. Is clip the main roadblock for fine-grained open-world perception? In 2024 International Conference on Content-Based Multimedia Indexing (CBMI), pages 1–8. IEEE, 2024. 5

    2024

Show all 88 references
  1. [9]

    In- structpix2pix: Learning to follow image editing instructions

    Tim Brooks, Aleksander Holynski, and Alexei A Efros. In- structpix2pix: Learning to follow image editing instructions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 18392–18402, 2023. 1, 2, 3, 4, 5, 6, 14

    2023

  2. [10]

    Lan- guage models are few-shot learners

    Tom Brown, Benjamin Mann, Nick Ryder, Melanie Sub- biah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakan- tan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Lan- guage models are few-shot learners. Advances in neural in- formation processing systems, 33:1877–1901, 2020. 2

    1901

  3. [11]

    Emerg- ing properties in self-supervised vision transformers

    Mathilde Caron, Hugo Touvron, Ishan Misra, Herv ´e J´egou, Julien Mairal, Piotr Bojanowski, and Armand Joulin. Emerg- ing properties in self-supervised vision transformers. In Pro- ceedings of the IEEE/CVF international conference on com- puter vision, pages 9650–9660, 2021. 1...

    2021

  4. [12]

    Mega-bench: Scaling multimodal evaluation to over 500 real-world tasks

    Jiacheng Chen, Tianhao Liang, Sherman Siu, Zhengqing Wang, Kai Wang, Yubo Wang, Yuansheng Ni, Wang Zhu, Ziyan Jiang, Bohan Lyu, Dongfu Jiang, Xuan He, Yuan Liu, Hexiang Hu, Xiang Yue, and Wenhu Chen. Mega-bench: Scaling multimodal evaluation to over 500 real-world tasks. arXiv...

    arXiv

  5. [13]

    Expanding performance boundaries of open-source multimodal models with model, data, and test- time scaling

    Zhe Chen, Weiyun Wang, Yue Cao, Yangzhou Liu, Zhang- wei Gao, Erfei Cui, Jinguo Zhu, Shenglong Ye, Hao Tian, Zhaoyang Liu, et al. Expanding performance boundaries of open-source multimodal models with model, data, and test- time scaling. arXiv preprint arXiv:2412.05271, 2024. 3

    2024 arXiv

  6. [14]

    Internvl: Scaling up vision foundation mod- els and aligning for generic visual-linguistic tasks

    Zhe Chen, Jiannan Wu, Wenhai Wang, Weijie Su, Guo Chen, Sen Xing, Muyan Zhong, Qinglong Zhang, Xizhou Zhu, Lewei Lu, et al. Internvl: Scaling up vision foundation mod- els and aligning for generic visual-linguistic tasks. In Pro- ceedings of the IEEE/CVF Conference on Computer...

    2024

  7. [15]

    Diffedit: Diffusion-based seman- tic image editing with mask guidance

    Guillaume Couairon, Jakob Verbeek, Holger Schwenk, and Matthieu Cord. Diffedit: Diffusion-based seman- tic image editing with mask guidance. arXiv preprint arXiv:2210.11427, 2022. 2, 3, 4, 5, 14

    2022 arXiv

  8. [16]

    Instructblip: To- wards general-purpose vision-language models with instruc- tion tuning

    Wenliang Dai, Junnan Li, D Li, AMH Tiong, J Zhao, W Wang, B Li, P Fung, and S Hoi. Instructblip: To- wards general-purpose vision-language models with instruc- tion tuning. arxiv 2023. arXiv preprint arXiv:2305.06500, 2,

    2023 arXiv

  9. [17]

    The epic-kitchens dataset: Collection, chal- lenges and baselines

    Dima Damen, Hazel Doughty, Giovanni Maria Farinella, Sanja Fidler, Antonino Furnari, Evangelos Kazakos, Davide Moltisanti, Jonathan Munro, Toby Perrett, Will Price, and Michael Wray. The epic-kitchens dataset: Collection, chal- lenges and baselines. IEEE Transactions on Patter...

  10. [18]

    Diffusion models beat GANs on image synthesis

    Prafulla Dhariwal and Alexander Nichol. Diffusion models beat GANs on image synthesis. 34:8780–8794, 2021. 2

    2021

  11. [19]

    Dreamlike photoreal 2.5, 2023

    Dreamlike.art. Dreamlike photoreal 2.5, 2023. 3

    2023

  12. [20]

    Scaling recti- fied flow transformers for high-resolution image synthesis, march 2024

    Patrick Esser, Sumith Kulal, Andreas Blattmann, Rahim Entezari, Jonas M ¨uller, Harry Saini, Yam Levi, Dominik Lorenz, Axel Sauer, Frederic Boesel, et al. Scaling recti- fied flow transformers for high-resolution image synthesis, march 2024. URL http://arxiv. org/abs/2403.0320...

    2024 arXiv

  13. [21]

    Dream- sim: Learning new dimensions of human visual similar- ity using synthetic data

    Stephanie Fu, Netanel Tamir, Shobhita Sundaram, Lucy Chai, Richard Zhang, Tali Dekel, and Phillip Isola. Dream- sim: Learning new dimensions of human visual similar- ity using synthetic data. arXiv preprint arXiv:2306.09344 ,

    arXiv

  14. [22]

    Seed-data-edit technical report: A hybrid dataset for in- structional image editing

    Yuying Ge, Sijie Zhao, Chen Li, Yixiao Ge, and Ying Shan. Seed-data-edit technical report: A hybrid dataset for in- structional image editing. arXiv preprint arXiv:2405.04007,

    arXiv

  15. [23]

    Seed-x: Mul- timodal models with unified multi-granularity comprehen- sion and generation

    Yuying Ge, Sijie Zhao, Jinguo Zhu, Yixiao Ge, Kun Yi, Lin Song, Chen Li, Xiaohan Ding, and Ying Shan. Seed-x: Mul- timodal models with unified multi-granularity comprehen- sion and generation. arXiv preprint arXiv:2404.14396, 2024. 5

    2024 arXiv

  16. [24]

    The” something something” video database for learning and evaluating visual common sense

    Raghav Goyal, Samira Ebrahimi Kahou, Vincent Michal- ski, Joanna Materzynska, Susanne Westphal, Heuna Kim, Valentin Haenel, Ingo Fruend, Peter Yianilos, Moritz Mueller-Freitag, et al. The” something something” video database for learning and evaluating visual common sense. In ...

    2017

  17. [25]

    Klaus Greff, Francois Belletti, Lucas Beyer, Carl Doersch, Yilun Du, Daniel Duckworth, David J Fleet, Dan Gnanapra- gasam, Florian Golemo, Charles Herrmann, Thomas Kipf, Abhijit Kundu, Dmitry Lagun, Issam Laradji, Hsueh- Ti (Derek) Liu, Henning Meyer, Yishu Miao, Derek Nowrouz...

    2022

  18. [26]

    Multi-reward as condi- tion for instruction-based image editing

    Xin Gu, Ming Li, Libo Zhang, Fan Chen, Longyin Wen, Tiejian Luo, and Sijie Zhu. Multi-reward as condi- tion for instruction-based image editing. arXiv preprint arXiv:2411.04713, 2024. 3

    2024 arXiv

  19. [27]

    Videoscore: Building automatic metrics to simulate fine-grained human feedback for video genera- tion

    Xuan He, Dongfu Jiang, Ge Zhang, Max Ku, Achint Soni, Sherman Siu, Haonan Chen, Abhranil Chandra, Ziyan Jiang, Aaran Arulraj, et al. Videoscore: Building automatic metrics to simulate fine-grained human feedback for video genera- tion. arXiv preprint arXiv:2406.15252, 2024. 2

    2024 arXiv

  20. [28]

    Prompt-to-Prompt im- age editing with cross attention control

    Amir Hertz, Ron Mokady, Jay Tenenbaum, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. Prompt-to-Prompt im- age editing with cross attention control. arXiv preprint arXiv:2208.01626, 2022. 2, 3, 4, 14

    2022 arXiv

  21. [29]

    Clipscore: A reference-free evaluation met- ric for image captioning

    Jack Hessel, Ari Holtzman, Maxwell Forbes, Ronan Le Bras, and Yejin Choi. Clipscore: A reference-free evaluation met- ric for image captioning. arXiv preprint arXiv:2104.08718,

    arXiv

  22. [30]

    Lora: Low-rank adaptation of large language models

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen- Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Weizhu Chen, et al. Lora: Low-rank adaptation of large language models. ICLR, 1(2):3, 2022. 6

    2022

  23. [31]

    Action genome: Actions as compositions of spatio- temporal scene graphs

    Jingwei Ji, Ranjay Krishna, Li Fei-Fei, and Juan Carlos Niebles. Action genome: Actions as compositions of spatio- temporal scene graphs. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 10236–10247, 2020. 14

    2020

  24. [32]

    Genai arena: An open evaluation platform for generative models

    Dongfu Jiang, Max Ku, Tianle Li, Yuansheng Ni, Shizhuo Sun, Rongqi Fan, and Wenhu Chen. Genai arena: An open evaluation platform for generative models. arXiv preprint arXiv:2406.04485, 2024. 2, 3, 6, 15

    2024 arXiv

  25. [33]

    Clevr: A diagnostic dataset for compositional language and elementary visual reasoning

    Justin Johnson, Bharath Hariharan, Laurens Van Der Maaten, Li Fei-Fei, C Lawrence Zitnick, and Ross Girshick. Clevr: A diagnostic dataset for compositional language and elementary visual reasoning. In Proceedings of the IEEE conference on computer vision and pattern recognitio...

    2017

  26. [34]

    What’s “up” with vision-language models? investigating their strug- gle with spatial reasoning

    Amita Kamath, Jack Hessel, and Kai-Wei Chang. What’s “up” with vision-language models? investigating their strug- gle with spatial reasoning. In EMNLP, 2023. 14

    2023

  27. [35]

    Pick-a-pic: An open dataset of user preferences for text-to-image genera- tion

    Yuval Kirstain, Adam Polyak, Uriel Singer, Shahbuland Matiana, Joe Penna, and Omer Levy. Pick-a-pic: An open dataset of user preferences for text-to-image genera- tion. Advances in Neural Information Processing Systems , 36:36652–36663, 2023. 2

    2023

  28. [36]

    Learning Action and Reasoning-Centric Image Editing from Videos and Simulations

    Benno Krojer, Dheeraj Vattikonda, Luis Lara, Varun Jam- pani, Eva Portelance, Christopher Pal, and Siva Reddy. Learning Action and Reasoning-Centric Image Editing from Videos and Simulations. In NeurIPS, 2024. Spotlight Paper. 1, 2, 3, 4, 5, 6, 14

    2024

  29. [37]

    Viescore: Towards explainable metrics for conditional image synthesis evaluation

    Max Ku, Dongfu Jiang, Cong Wei, Xiang Yue, and Wenhu Chen. Viescore: Towards explainable metrics for conditional image synthesis evaluation. arXiv preprint arXiv:2312.14867, 2023. 2, 3, 6, 8, 14, 15

    2023 arXiv

  30. [38]

    Imagenhub: Standardizing the evaluation of conditional image generation models

    Max Ku, Tianle Li, Kai Zhang, Yujie Lu, Xingyu Fu, Wen- wen Zhuang, and Wenhu Chen. Imagenhub: Standardizing the evaluation of conditional image generation models. In The Twelfth International Conference on Learning Represen- tations, 2024. 2, 3, 6, 7, 8, 13, 15, 16

    2024

  31. [39]

    Lisa: Reasoning segmentation via large language model

    Xin Lai, Zhuotao Tian, Yukang Chen, Yanwei Li, Yuhui Yuan, Shu Liu, and Jiaya Jia. Lisa: Reasoning segmentation via large language model. arXiv preprint arXiv:2308.00692,

    arXiv

  32. [40]

    Introduc- ing idefics2: A powerful 8b vision-language model for the community

    Hugo Laurenc ¸on, L´eo Tronchon, and Victor Sanh. Introduc- ing idefics2: A powerful 8b vision-language model for the community. Hugging Face Blog, April 2024. 3

    2024

  33. [41]

    Llava-onevision: Easy visual task transfer

    Bo Li, Yuanhan Zhang, Dong Guo, Renrui Zhang, Feng Li, Hao Zhang, Kaichen Zhang, Peiyuan Zhang, Yanwei Li, Zi- wei Liu, et al. Llava-onevision: Easy visual task transfer. arXiv preprint arXiv:2408.03326, 2024. 3

    2024 arXiv

  34. [42]

    Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models

    Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi. Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models. In In- ternational conference on machine learning , pages 19730– 19742. PMLR, 2023. 2, 3

    2023

  35. [43]

    ZONE: Zero-shot instruction-guided local editing

    Shanglin Li, Bohan Zeng, Yutang Feng, Sicheng Gao, Xi- uhui Liu, Jiaming Liu, Lin Li, Xu Tang, Yao Hu, Jianzhuang Liu, et al. ZONE: Zero-shot instruction-guided local editing. pages 6254–6263, 2024. 1

    2024

  36. [44]

    Rich hu- man feedback for text-to-image generation

    Youwei Liang, Junfeng He, Gang Li, Peizhao Li, Arseniy Klimovskiy, Nicholas Carolan, Jiao Sun, Jordi Pont-Tuset, Sarah Young, Feng Yang, Junjie Ke, Krishnamurthy Dj Dvi- jotham, Katie Collins, Yiwen Luo, Yang Li, Kai J Kohlhoff, Deepak Ramachandran, and Vidhya Navalpakkam. Ric...

    2024

  37. [45]

    Evaluating text-to-visual generation with image-to-text gen- eration

    Zhiqiu Lin, Deepak Pathak, Baiqi Li, Jiayao Li, Xide Xia, Graham Neubig, Pengchuan Zhang, and Deva Ramanan. Evaluating text-to-visual generation with image-to-text gen- eration. In European Conference on Computer Vision, pages 366–384. Springer, 2024. 2

    2024

  38. [46]

    Llava-next: Im- proved reasoning, ocr, and world knowledge, January 2024

    Haotian Liu, Chunyuan Li, Yuheng Li, Bo Li, Yuanhan Zhang, Sheng Shen, and Yong Jae Lee. Llava-next: Im- proved reasoning, ocr, and world knowledge, January 2024. 1, 2, 3, 6

    2024

  39. [47]

    Visual instruction tuning

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. Visual instruction tuning. In NeurIPS, 2023. 3, 6

    2023

  40. [48]

    Visual instruction tuning

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. Visual instruction tuning. Advances in neural information processing systems, 36, 2024. 2

    2024

  41. [49]

    Latent Consistency Models: Synthesizing high- resolution images with few-step inference

    Simian Luo, Yiqin Tan, Longbo Huang, Jian Li, and Hang Zhao. Latent Consistency Models: Synthesizing high- resolution images with few-step inference. arXiv preprint arXiv:2310.04378, 2023. 2

    2023 arXiv

  42. [50]

    I2ebench: A comprehensive benchmark for instruction- based image editing

    Yiwei Ma, Jiayi Ji, Ke Ye, Weihuang Lin, Zhibin Wang, Yonghan Zheng, Qiang Zhou, Xiaoshuai Sun, and Rongrong Ji. I2ebench: A comprehensive benchmark for instruction- based image editing. arXiv preprint arXiv:2408.14180 ,

    arXiv

  43. [51]

    Sdedit: Guided image synthesis and editing with stochastic differential equa- tions

    Chenlin Meng, Yutong He, Yang Song, Jiaming Song, Jia- jun Wu, Jun-Yan Zhu, and Stefano Ermon. Sdedit: Guided image synthesis and editing with stochastic differential equa- tions. arXiv preprint arXiv:2108.01073, 2021. 3, 4, 5, 14

    2021 arXiv

  44. [52]

    Watch Your Steps: Local image and scene editing by text instructions

    Ashkan Mirzaei, Tristan Aumentado-Armstrong, Marcus A Brubaker, Jonathan Kelly, Alex Levinshtein, Konstantinos G Derpanis, and Igor Gilitschenski. Watch Your Steps: Local image and scene editing by text instructions. pages 111–129. Springer, 2025. 1

    2025

  45. [53]

    Null-text Inversion for editing real images using guided diffusion models

    Ron Mokady, Amir Hertz, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. Null-text Inversion for editing real images using guided diffusion models. pages 6038–6047, 2023. 2

    2023

  46. [54]

    Dinov2: Learning robust visual features without supervision

    Maxime Oquab, Timoth ´ee Darcet, Th ´eo Moutakanni, Huy V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, et al. Dinov2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193, 2023. 1, 3, 4

    2023 arXiv

  47. [55]

    Zero-shot image-to-image translation

    Gaurav Parmar, Krishna Kumar Singh, Richard Zhang, Yijun Li, Jingwan Lu, and Jun-Yan Zhu. Zero-shot image-to-image translation. In ACM SIGGRAPH 2023 Conference Proceed- ings, pages 1–11, 2023. 2, 3, 4, 5, 14

    2023

  48. [56]

    SDXL: Improving latent diffusion mod- els for high-resolution image synthesis

    Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas M ¨uller, Joe Penna, and Robin Rombach. SDXL: Improving latent diffusion mod- els for high-resolution image synthesis. arXiv preprint arXiv:2307.01952, 2023. 2

    2023 arXiv

  49. [57]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. In International conference on machine learning, ...

    2021

  50. [58]

    Hierarchical text-conditional image gen- eration with clip latents

    Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. Hierarchical text-conditional image gen- eration with clip latents. arXiv preprint arXiv:2204.06125, 1(2):3, 2022. 2

    2022 arXiv

  51. [59]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. pages 10684–10695,

  52. [60]

    Emu edit: Precise image editing via recognition and gen- eration tasks

    Shelly Sheynin, Adam Polyak, Uriel Singer, Yuval Kirstain, Amit Zohar, Oron Ashual, Devi Parikh, and Yaniv Taigman. Emu edit: Precise image editing via recognition and gen- eration tasks. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pa...

    2024

  53. [61]

    Aria: Advancing multimodal ai

    John Smith, Jane Doe, and Michael Lee. Aria: Advancing multimodal ai. arXiv preprint arXiv:2310.67890, 2023. 3

    2023

  54. [62]

    Denoising diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models. arXiv preprint arXiv:2010.02502, 2020. 2

    2010 arXiv

  55. [63]

    Ie-bench: Advancing the measurement of text- driven image editing for human perception alignment

    Shangkun Sun, Bowen Qu, Xiaoyu Liang, Songlin Fan, and Wei Gao. Ie-bench: Advancing the measurement of text- driven image editing for human perception alignment. arXiv preprint arXiv:2501.09927, 2025. 3

    2025 arXiv

  56. [64]

    Gemini 1.5: Unlocking multimodal understanding across millions of tokens of con- text

    Gemini Team, Petko Georgiev, Ving Ian Lei, Ryan Burnell, Libin Bai, Anmol Gulati, Garrett Tanzer, Damien Vincent, Zhufeng Pan, Shibo Wang, et al. Gemini 1.5: Unlocking multimodal understanding across millions of tokens of con- text. arXiv preprint arXiv:2403.05530, 2024. 2, 3,...

    2024 arXiv

  57. [65]

    Plug-and-Play diffusion features for text-driven image-to-image translation

    Narek Tumanyan, Michal Geyer, Shai Bagon, and Tali Dekel. Plug-and-Play diffusion features for text-driven image-to-image translation. pages 1921–1930, 2023. 2, 15

    1921

  58. [66]

    EDICT: Ex- act diffusion inversion via coupled transformations

    Bram Wallace, Akash Gokul, and Nikhil Naik. EDICT: Ex- act diffusion inversion via coupled transformations. pages 22532–22541, 2023. 2

    2023

  59. [67]

    Qwen2-vl: Enhancing vision-language model’s perception of the world at any resolution

    Peng Wang, Shuai Bai, Sinan Tan, Shijie Wang, Zhihao Fan, Jinze Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, et al. Qwen2-vl: Enhancing vision-language model’s perception of the world at any resolution. arXiv preprint arXiv:2409.12191, 2024. 3, 4, 6, 13

    2024 arXiv

  60. [68]

    Cogvlm: Visual expert for pretrained language models, 2023

    Weihan Wang, Qingsong Lv, Wenmeng Yu, Wenyi Hong, Ji Qi, Yan Wang, Junhui Ji, Zhuoyi Yang, Lei Zhao, Xixuan Song, Jiazheng Xu, Bin Xu, Juanzi Li, Yuxiao Dong, Ming Ding, and Jie Tang. Cogvlm: Visual expert for pretrained language models, 2023. 2

    2023

  61. [69]

    Image quality assessment: from error visibility to structural similarity

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Si- moncelli. Image quality assessment: from error visibility to structural similarity. IEEE transactions on image processing, 13(4):600–612, 2004. 1, 2

    2004

  62. [70]

    Omniedit: Building image edit- ing generalist models through specialist supervision

    Cong Wei, Zheyang Xiong, Weiming Ren, Xinrun Du, Ge Zhang, and Wenhu Chen. Omniedit: Building image edit- ing generalist models through specialist supervision. arXiv preprint arXiv:2411.07199, 2024. 2, 3

    2024 arXiv

  63. [71]

    A latent space of stochastic diffusion models for zero-shot image editing and guidance

    Chen Henry Wu and Fernando De la Torre. A latent space of stochastic diffusion models for zero-shot image editing and guidance. In ICCV, 2023. 2, 3, 4, 14

    2023

  64. [72]

    Uncovering the disentanglement capability in text- to-image diffusion models

    Qiucheng Wu, Yujian Liu, Handong Zhao, Ajinkya Kale, Trung Bui, Tong Yu, Zhe Lin, Yang Zhang, and Shiyu Chang. Uncovering the disentanglement capability in text- to-image diffusion models. pages 1900–1910, 2023. 2

    1900

  65. [73]

    Human preference score v2: A solid benchmark for evaluating human preferences of text-to-image synthesis

    Xiaoshi Wu, Yiming Hao, Keqiang Sun, Yixiong Chen, Feng Zhu, Rui Zhao, and Hongsheng Li. Human preference score v2: A solid benchmark for evaluating human preferences of text-to-image synthesis. arXiv preprint arXiv:2306.09341 ,

    arXiv

  66. [74]

    Multimodal large language models make text-to- image generative models align better

    Xun Wu, Shaohan Huang, Guolong Wang, Jing Xiong, and Furu Wei. Multimodal large language models make text-to- image generative models align better. Advances in Neural Information Processing Systems, 37:81287–81323, 2025. 2

    2025

  67. [75]

    Multimodal large language model is a human-aligned annotator for text-to- image generation

    Xun Wu, Shaohan Huang, and Furu Wei. Multimodal large language model is a human-aligned annotator for text-to- image generation. arXiv preprint arXiv:2404.15100, 2024. 3

    2024 arXiv

  68. [76]

    Human preference score: Better aligning text- to-image models with human preference

    Xiaoshi Wu, Keqiang Sun, Feng Zhu, Rui Zhao, and Hong- sheng Li. Human preference score: Better aligning text- to-image models with human preference. In Proceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 2096–2105, 2023. 2, 3

    2023

  69. [77]

    Imagere- ward: Learning and evaluating human preferences for text- to-image generation

    Jiazheng Xu, Xiao Liu, Yuchen Wu, Yuxuan Tong, Qinkai Li, Ming Ding, Jie Tang, and Yuxiao Dong. Imagere- ward: Learning and evaluating human preferences for text- to-image generation. Advances in Neural Information Pro- cessing Systems, 36, 2024. 2, 3, 6, 14

    2024

  70. [78]

    Inversion-free image editing with natural language

    Sihan Xu, Yidong Huang, Jiayi Pan, Ziqiao Ma, and Joyce Chai. Inversion-free image editing with natural language

  71. [79]

    Minicpm-v: A gpt-4v level mllm on your phone

    Yuan Yao, Tianyu Yu, Ao Zhang, Chongyi Wang, Junbo Cui, Hongji Zhu, Tianchi Cai, Haoyu Li, Weilin Zhao, Zhihui He, et al. Minicpm-v: A gpt-4v level mllm on your phone. arXiv preprint arXiv:2408.01800, 2024. 3

    2024 arXiv

  72. [80]

    Mmmu: A massive multi-discipline multimodal understanding and reasoning benchmark for ex- pert agi

    Xiang Yue, Yuansheng Ni, Kai Zhang, Tianyu Zheng, Ruoqi Liu, Ge Zhang, Samuel Stevens, Dongfu Jiang, Weiming Ren, Yuxuan Sun, et al. Mmmu: A massive multi-discipline multimodal understanding and reasoning benchmark for ex- pert agi. In Proceedings of CVPR, 2024. 3

    2024

  73. [81]

    When and why vision- language models behave like bags-of-words, and what to do about it? arXiv preprint arXiv:2210.01936, 2022

    Mert Yuksekgonul, Federico Bianchi, Pratyusha Kalluri, Dan Jurafsky, and James Zou. When and why vision- language models behave like bags-of-words, and what to do about it? arXiv preprint arXiv:2210.01936, 2022. 2, 3

    2022 arXiv

  74. [82]

    Long-clip: Unlocking the long-text capability of clip

    Beichen Zhang, Pan Zhang, Xiaoyi Dong, Yuhang Zang, and Jiaqi Wang. Long-clip: Unlocking the long-text capability of clip. In European Conference on Computer Vision, pages 310–325. Springer, 2024. 2, 3

    2024

  75. [83]

    Magicbrush: A manually annotated dataset for instruction- guided image editing

    Kai Zhang, Lingbo Mo, Wenhu Chen, Huan Sun, and Yu Su. Magicbrush: A manually annotated dataset for instruction- guided image editing. Advances in Neural Information Pro- cessing Systems, 36, 2024. 1, 2, 3, 4, 5, 6, 7, 8, 13, 14, 15, 17

    2024

  76. [84]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shecht- man, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 586–595, 2018. 1, 2

    2018

  77. [85]

    Learning multi- dimensional human preference for text-to-image generation

    Sixian Zhang, Bohan Wang, Junqiang Wu, Yan Li, Tingt- ing Gao, Di Zhang, and Zhongyuan Wang. Learning multi- dimensional human preference for text-to-image generation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8018–8027, 2024. 2

    2024

  78. [86]

    Hive: Harnessing human feedback for instructional visual editing

    Shu Zhang, Xinyi Yang, Yihao Feng, Can Qin, Chia-Chih Chen, Ning Yu, Zeyuan Chen, Huan Wang, Silvio Savarese, Stefano Ermon, et al. Hive: Harnessing human feedback for instructional visual editing. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogn...

    2024

  79. [87]

    UltraEdit: Instruction-based fine-grained image editing at scale

    Haozhe Zhao, Xiaojian Ma, Liang Chen, Shuzheng Si, Ru- jie Wu, Kaikai An, Peiyu Yu, Minjia Zhang, Qing Li, and Baobao Chang. UltraEdit: Instruction-based fine-grained image editing at scale. arXiv preprint arXiv:2407.05282 ,

    arXiv

  80. [88]

    Can you rate how successful the edit instruction [IN- STRUCTION] has been executed from the first image to the second image with a score from 0 to 10?

    Yaowei Zheng, Richong Zhang, Junhao Zhang, Yanhan Ye, Zheyan Luo, Zhangchi Feng, and Yongqiang Ma. Llamafac- tory: Unified efficient fine-tuning of 100+ language mod- els. In Proceedings of the 62nd Annual Meeting of the As- sociation for Computational Linguistics (Volume 3: S...

    2024

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.