Holistic Evaluation of Text-to-Image Models: Author contributions, Acknowledgments and References

Authors: (1) Tony Lee, Stanford with Equal contribution; (2) Michihiro Yasunaga, Stanford with Equal contribution; (3) Chenlin Meng, Stanford with Equal contribution; (4) Yifan Mai, Stanford; (5) Joon Sung Park, Stanford; (6) Agrim Gupta, Stanford; (7) Yunzhi Zhang, Stanford; (8) Deepak Narayanan, Microsoft; (9) Hannah Benita Teufel, Aleph Alpha; (10) Marco Bellagente, Aleph Alpha; (11) Minguk Kang, POSTECH; (12) Taesung Park, Adobe; (13) Jure Leskovec, Stanford; (14) Jun-Yan Zhu, CMU; (15) Li Fei-Fei, Stanford; (16) Jiajun Wu, Stanford; (17) Stefano Ermon, Stanford; (18) Percy Liang, Stanford. Table of Links Abstract and 1 Introduction 2 Core framework 3 Aspects 4 Scenarios 5 Metrics 6 Models 7 Experiments and results 8 Related work 9 Conclusion 10 Limitations Author contributions, Acknowledgments and References A Datasheet B Scenario details C Metric details D Model details E Human evaluation procedure Author contributions Tony Lee: Co-led the project. Designed the core framework (aspects, scenarios, metrics). Implemented scenarios, metrics and models. Conducted experiments. Contributed to writing. Michihiro Yasunaga: Co-led the project. Designed the core framework (aspects, scenarios, metrics). Wrote the paper. Conducted analysis. Implemented models. Chenlin Meng: Designed the core framework (aspects, scenarios, metrics). Contributed to writing. Yifan Mai: Implemented the evaluation infrastructure. Contributed to project discussions. Joon Sung Park: Designed human evaluation questions. Agrim Gupta: Implemented the detection scenario and metrics. Yunzhi Zhang: Implemented the detection scenario and metrics. Deepak Narayanan: Provided expertise and analysis of efficiency metrics. Hannah Teufel: Provided model expertise and inference. Marco Bellagente: Provided model expertise and inference. Minguk Kang: Provided model expertise and inference. Taesung Park: Provided model expertise and inference. Jure Leskovec: Provided advice on human evaluation and paper writing. Jun-Yan Zhu: Provided advice on human evaluation and paper writing. Li Fei-Fei: Provided advice on the core framework. Jiajun Wu: Provided advice on the core framework. Stefano Ermon: Provided advice on the core framework. Percy Liang: Provided overall supervision and guidance throughout the project. Statement of neutrality. The authors of this paper affirm their commitment to maintaining a fair and independent evaluation of the image generation models. We acknowledge that the author affiliations encompass a range of academic and industrial institutions, including those where some of the models we evaluate were developed. However, the authors’ involvement is solely based on their expertise and efforts to run and evaluate the models, and the authors have treated all models equally throughout the evaluation process, regardless of their sources. This study aims to provide an objective understanding and assessment of models across various aspects, and we do not intend to endorse specific models. Acknowledgments We thank Robin Rombach, Yuhui Zhang, members of Stanford P-Lambda, CRFM, and SNAP groups, as well as our anonymous reviewers for providing valuable feedback. We thank Josselin Somerville for assisting with the human evaluation infrastructure. This work is supported in part by the AI2050 program at Schmidt Futures (Grant G-22-63429), a Google Research Award, and ONR N00014-23-1- 2355. Michihiro Yasunaga is supported by a Microsoft Research PhD Fellowship. References [1] Percy Liang, Rishi Bommasani, Tony Lee, Dimitris Tsipras, Dilara Soylu, Michihiro Yasunaga, Yian Zhang, Deepak Narayanan, Yuhuai Wu, Ananya Kumar, et al. Holistic evaluation of language models. arXiv preprint arXiv:2211.09110, 2022. [2] Aditya Ramesh, Mikhail Pavlov, Gabriel Goh, Scott Gray, Chelsea Voss, Alec Radford, Mark Chen, and Ilya Sutskever. Zero-shot text-to-image generation. In International Conference on Machine Learning, pages 8821–8831. PMLR, 2021. [3] Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. Hierarchical text-conditional image generation with clip latents. arXiv preprint arXiv:2204.06125, 2022. [4] Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. Highresolution image synthesis with latent diffusion models. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022. [5] Oran Gafni, Adam Polyak, Oron Ashual, Shelly Sheynin, Devi Parikh, and Yaniv Taigman. Make-a-scene: Scene-based text-to-image generation with human priors. arXiv preprint arXiv:2203.13131, 2022. [6] Chitwan Saharia, William Chan, Saurabh Saxena, Lala Li, Jay Whang, Emily Denton, Seyed Kamyar Seyed Ghasemipour, Burcu Karagol Ayan, S Sara Mahdavi, Rapha Gontijo Lopes, et al. Photorealistic text-to-image diffusion models with deep language understanding. arXiv preprint arXiv:2205.11487, 2022. [7] Jiahui Yu, Yuanzhong Xu, Jing Yu Koh, Thang Luong, Gunjan Baid, Zirui Wang, Vijay Vasudevan, Alexander Ku, Yinfei Yang, Burcu Karagol Ayan, et al. Scaling autoregressive models for contentrich text-to-image generation. arXiv preprint arXiv:2206.10789, 2022. [8] Patrick Schramowski, Manuel Brack, Björn Deiseroth, and Kristian Kersting. Safe latent diffusion: Mitigating inappropriate degeneration in diffusion models. arXiv preprint arXiv:2211.05105, 2022. [9] Michihiro Yasunaga, Armen Aghajanyan, Weijia Shi, Rich James, Jure Leskovec, Percy Liang, Mike Lewis, Luke Zettlemoyer, and Wen-tau Yih. Retrieval-augmented multimodal language modeling. In International Conference on Machine Learning (ICML), 2023. [10] Ming Ding, Wendi Zheng, Wenyi Hong, and Jie Tang. Cogview2: Faster and better text-to-image generation via hierarchical transformers. arXiv preprint arXiv:2204.14217, 2022. [11] Zhida Feng, Zhenyu Zhang, Xintong Yu, Yewei Fang, Lanxin Li, Xuyi Chen, Yuxiang Lu, Jiaxiang Liu, Weichong Yin, Shikun Feng, et al. Ernie-vilg 2.0: Improving text-to-image diffusion model with knowledge-enhanced mixture-of-denoising-experts. arXiv preprint arXiv:2210.15257, 2022. [12] Minguk Kang, Jun-Yan Zhu, Richard Zhang, Jaesik Park, Eli Shechtman, Sylvain Paris, and Taesung Park. Scaling up gans for text-to-image synthesis. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2023. [13] Eva Cetinic and James She. Understanding and creating art with ai: review and outlook. ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM), 18(2): 1–22, 2022. [14] Pierre Chambon, Christian Bluethgen, Curtis P Langlotz, and Akshay Chaudhari. Adapting pretrained vision-language foundational models to medical imaging domains. arXiv preprint arXiv:2210.04133, 2022. [15] Midjourney. https://www.midjourney.com/, . [16] Midjourney statistics. https://photutorial.com/midjourney-statistics/, . [17] Ling Yang, Zhilong Zhang, Yang Song, Shenda Hong, Runsheng Xu, Yue Zhao, Yingxia Shao, Wentao Zhang, Bin Cui, and Ming-Hsuan Yang. Diffusion models: A comprehensive survey of methods and applications. arXiv preprint arXiv:2209.00796, 2022. [18] Chenshuang Zhang, Chaoning Zhang, Mengchun Zhang, and In So Kweon. Text-to-image diffusion model in generative ai: A survey. arXiv preprint arXiv:2303.07909, 2023. [19] Abeba Birhane, Vinay Uday Prabhu, and Emmanuel Kahembwe. Ethical considerations of generative AI. AI for Content Creation Workshop, CVPR, 2021. [20] Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A largescale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009. [21] Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Dollár, and C Lawrence Zitnick. Microsoft coco: Common objects in context. In European conference on computer vision, 2014. [22] Scott Reed, Zeynep Akata, Honglak Lee, and Bernt Schiele. Learning deep representations of fine-grained visual descriptions. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 49–58, 2016. [23] Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. Gans trained by a two time-scale update rule converge to a local nash equilibrium. Advances in neural information processing systems, 30, 2017. [24] Jack Hessel, Ari Holtzman, Maxwell Forbes, Ronan Le Bras, and Yejin Choi. Clipscore: A reference-free evaluation metric for image captioning. arXiv preprint arXiv:2104.08718, 2021. [25] Sharon Zhou, Mitchell Gordon, Ranjay Krishna, Austin Narcomey, Li F Fei-Fei, and Michael Bernstein. Hype: A benchmark for human eye perceptual evaluation of generative models. Advances in neural information processing systems, 32, 2019. [26] Jiazheng Xu, Xiao Liu, Yuchen Wu, Yuxuan Tong, Qinkai Li, Ming Ding, Jie Tang, and Yuxiao Dong. Imagereward: Learning and evaluating human preferences for text-to-image generation. arXiv preprint arXiv:2304.05977, 2023. [27] Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018. [28] Yaru Hao, Zewen Chi, Li Dong, and Furu Wei. Optimizing prompts for text-to-image generation. arXiv preprint arXiv:2212.09611, 2022. [29] Jaemin Cho, Abhay Zala, and Mohit Bansal. Dall-eval: Probing the reasoning skills and social biases of text-to-image generative transformers. arXiv preprint arXiv:2202.04053, 2022. [30] Evelina Leivada, Elliot Murphy, and Gary Marcus. Dall-e 2 fails to reliably capture common syntactic processes. arXiv preprint arXiv:2210.12889, 2022. [31] Colin Conwell and Tomer Ullman. Testing relational understanding in text-guided image generation. arXiv preprint arXiv:2208.00005, 2022. [32] Tristan Thrush, Ryan Jiang, Max Bartolo, Amanpreet Singh, Adina Williams, Douwe Kiela, and Candace Ross. Winoground: Probing vision and language models for visio-linguistic compositionality. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5238–5248, 2022. [33] Federico Bianchi, Pratyusha Kalluri, Esin Durmus, Faisal Ladhak, Myra Cheng, Debora Nozza, Tatsunori Hashimoto, Dan Jurafsky, James Zou, and Aylin Caliskan. Easily accessible text-to-image generation amplifies demographic stereotypes at large scale. arXiv preprint arXiv:2211.03759, 2022. [34] Morgan King. Harmful biases in artificial intelligence. The Lancet Psychiatry, 9(11):e48, 2022. [35] Mayu Otani, Riku Togashi, Yu Sawai, Ryosuke Ishigami, Yuta Nakashima, Esa Rahtu, Janne Heikkilä, and Shin’ichi Satoh. Toward verifiable and reproducible human evaluation for text-toimage generation. arXiv preprint arXiv:2304.01816, 2023. [36] Tim Salimans, Ian Goodfellow, Wojciech Zaremba, Vicki Cheung, Alec Radford, and Xi Chen. Improved techniques for training gans. Advances in neural information processing systems, 29, 2016. [37] Nicholas Carlini, Jamie Hayes, Milad Nasr, Matthew Jagielski, Vikash Sehwag, Florian Tramer, Borja Balle, Daphne Ippolito, and Eric Wallace. Extracting training data from diffusion models. arXiv preprint arXiv:2301.13188, 2023. [38] Nupur Kumari, Bingliang Zhang, Sheng-Yu Wang, Eli Shechtman, Richard Zhang, and Jun-Yan Zhu. Ablating concepts in text-to-image diffusion models. In International Conference on Computer Vision (ICCV), 2023. [39] Rohit Gandikota, Joanna Materzynska, Jaden Fiotto-Kaufman, and David Bau. Erasing concepts from diffusion models. 2023. [40] Christoph Schuhmann, Romain Beaumont, Richard Vencu, Cade Gordon, Ross Wightman, Mehdi Cherti, Theo Coombes, Aarush Katta, Clayton Mullis, Mitchell Wortsman, et al. Laion-5b: An open large-scale dataset for training next generation image-text models. arXiv preprint arXiv:2210.08402, 2022. [41] Richard P Taylor, Adam P Micolich, and David Jonas. Fractal analysis of pollock’s drip paintings. Nature, 399(6735):422–422, 1999. [42] Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to-end object detection with transformers. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part I 16, pages 213–229. Springer, 2020. [43] Yanghao Li, Hanzi Mao, Ross Girshick, and Kaiming He. Exploring plain vision transformer backbones for object detection. arXiv preprint arXiv:2203.16527, 2022. [44] Nudenet. https://github.com/notAI-tech/NudeNet. [45] Dreamlike diffusion 1.0. https://huggingface.co/dreamlike-art/ dreamlike-diffusion-1.0, . [46] Dreamlike photoreal 2.0. https://huggingface.co/dreamlike-art/ dreamlike-photoreal-2.0, . [47] Openjourney. https://huggingface.co/prompthero/openjourney, . [48] Openjourney-v4. https://huggingface.co/prompthero/openjourney-v4, . [49] Redshift diffusion. https://huggingface.co/nitrosocke/redshift-diffusion. [50] Vintedois (22h) diffusion. https://huggingface.co/22h/vintedois-diffusion-v0-1. [51] Lexica. https://lexica.art/docs. [52] Dall-e mini. https://github.com/borisdayma/dalle-mini. [53] Saehoon Kim, Sanghun Cho, Chiheon Kim, Doyup Lee, and Woonhyuk Baek. mindall-e on conceptual captions. https://github.com/kakaobrain/minDALL-E, 2021. [54] Marco Bellagente, Manuel Brack, Hannah Teufel, Felix Friedrich, Björn Deiseroth, Constantin Eichenberg, Andrew Dai, Robert Baldock, Souradeep Nanda, Koen Oostermeijer, Andres Felipe Cruz-Salinas, Patrick Schramowski, Kristian Kersting, and Samuel Weinbach. Multifusion: Fusing pre-trained models for multi-lingual, multi-modal image generation, 2023. [55] deep-floyd. https://github.com/deep-floyd/IF. [56] Pranav Rajpurkar, Jian Zhang, Konstantin Lopyrev, and Percy Liang. Squad: 100,000+ questions for machine comprehension of text. In Empirical Methods in Natural Language Processing (EMNLP), 2016. [57] Kawin Ethayarajh and Dan Jurafsky. Utility is in the eye of the user: A critique of nlp leaderboards. arXiv preprint arXiv:2009.13888, 2020. [58] Inioluwa Deborah Raji, Emily M Bender, Amandalynne Paullada, Emily Denton, and Alex Hanna. Ai and the everything in the whole wide world benchmark. arXiv preprint arXiv:2111.15366, 2021. [59] Alex Wang, Amanpreet Singh, Julian Michael, Felix Hill, Omer Levy, and Samuel R Bowman. Glue: A multi-task benchmark and analysis platform for natural language understanding. arXiv preprint arXiv:1804.07461, 2018. [60] Pang Wei Koh, Shiori Sagawa, Henrik Marklund, Sang Michael Xie, Marvin Zhang, Akshay Balsubramani, Weihua Hu, Michihiro Yasunaga, Richard Lanas Phillips, Irena Gao, et al. Wilds: A benchmark of in-the-wild distribution shifts. In International Conference on Machine Learning, pages 5637–5664. PMLR, 2021. [61] Aarohi Srivastava, Abhinav Rastogi, Abhishek Rao, Abu Awal Md Shoeb, Abubakar Abid, Adam Fisch, Adam R Brown, Adam Santoro, Aditya Gupta, Adrià Garriga-Alonso, et al. Beyond the imitation game: Quantifying and extrapolating the capabilities of language models. arXiv preprint arXiv:2206.04615, 2022. [62] Chengwei Qin, Aston Zhang, Zhuosheng Zhang, Jiaao Chen, Michihiro Yasunaga, and Diyi Yang. Is chatgpt a general-purpose natural language processing task solver? arXiv preprint arXiv:2302.06476, 2023. [63] Tero Karras, Samuli Laine, and Timo Aila. A style-based generator architecture for generative adversarial networks. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019. [64] Suman Ravuri and Oriol Vinyals. Classification accuracy score for conditional generative models. Advances in neural information processing systems, 32, 2019. [65] Tuomas Kynkäänniemi, Tero Karras, Samuli Laine, Jaakko Lehtinen, and Timo Aila. Improved precision and recall metric for assessing generative models. Advances in Neural Information Processing Systems, 32, 2019. [66] Gaurav Parmar, Richard Zhang, and Jun-Yan Zhu. On aliased resizing and surprising subtleties in gan evaluation. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022. [67] 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 generation. arXiv preprint arXiv:2305.01569, 2023. [68] Rudolf Arnheim. Art and visual perception: A psychology of the creative eye. Univ of California Press, 1969. [69] Philip Galanter. Computational aesthetic evaluation: past and future. Computers and creativity, pages 255–293, 2012. [70] William Lidwell, Kritina Holden, and Jill Butler. Universal principles of design, revised and updated: 125 ways to enhance usability, influence perception, increase appeal, make better design decisions, and teach through design. Rockport Pub, 2010. [71] Catherine Wah, Steve Branson, Peter Welinder, Pietro Perona, and Serge J. Belongie. The caltech-ucsd birds-200-2011 dataset. 2011. [72] Caleb Ziems, Jiaao Chen, Camille Harris, Jessica Anderson, and Diyi Yang. Value: Understanding dialect disparity in nlu, 2022. [73] 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 supervision. In International conference on machine learning, pages 8748–8763. PMLR, 2021. [74] Yogesh Balaji, Seungjun Nah, Xun Huang, Arash Vahdat, Jiaming Song, Karsten Kreis, Miika Aittala, Timo Aila, Samuli Laine, Bryan Catanzaro, et al. ediffi: Text-to-image diffusion models with an ensemble of expert denoisers. arXiv preprint arXiv:2211.01324, 2022. [75] Christian Szegedy, Wei Liu, Yangqing Jia, Pierre Sermanet, Scott Reed, Dragomir Anguelov, Dumitru Erhan, Vincent Vanhoucke, and Andrew Rabinovich. Going deeper with convolutions. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 1–9, 2015. [76] Maximilian Seitzer. pytorch-fid: FID Score for PyTorch. https://github.com/mseitzer/pytorch-fid, August 2020. Version 0.3.0. [77] Jonathan Ho and Tim Salimans. Classifier-free diffusion guidance. arXiv preprint arXiv:2207.12598, 2022. [78] Anton Obukhov, Maximilian Seitzer, Po-Wei Wu, Semen Zhydenko, Jonathan Kyl, and Elvis YuJing Lin. High-fidelity performance metrics for generative models in pytorch, 2020. URL https: //github.com/toshas/torch-fidelity. Version: 0.3.0, DOI: 10.5281/zenodo.4957738. [79] Deepak Narayanan, Keshav Santhanam, Peter Henderson, Rishi Bommasani, Tony Lee, and Percy Liang. Cheaply Evaluating Inference Efficiency Metrics for Autoregressive Transformer APIs. arXiv preprint arXiv:2305.02440, 2023. [80] Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, et al. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. [81] Yuxin Wu, Alexander Kirillov, Francisco Massa, Wan-Yen Lo, and Ross Girshick. Detectron2. https://github.com/facebookresearch/detectron2, 2019. [82] Tim Salimans and Jonathan Ho. Progressive distillation for fast sampling of diffusion models. arXiv preprint arXiv:2202.00512, 2022. [83] Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. Language models are unsupervised multitask learners. OpenAI blog, 1(8):9, 2019. [84] Minwoo Byeon, Beomhee Park, Haecheon Kim, Sungjun Lee, Woonhyuk Baek, and Saehoon Kim. Coyo-700m: Image-text pair dataset. https://github.com/kakaobrain/ coyo-dataset, 2022. This paper is available on arxiv under CC BY 4.0 DEED license. Authors: (1) Tony Lee, Stanford with Equal contribution; (2) Michihiro Yasunaga, Stanford with Equal contribution; (3) Chenlin Meng, Stanford with Equal contribution; (4) Yifan Mai, Stanford; (5) Joon Sung Park, Stanford; (6) Agrim Gupta, Stanford; (7) Yunzhi Zhang, Stanford; (8) Deepak Narayanan, Microsoft; (9) Hannah Benita Teufel, Aleph Alpha; (10) Marco Bellagente, Aleph Alpha; (11) Minguk Kang, POSTECH; (12) Taesung Park, Adobe; (13) Jure Leskovec, Stanford; (14) Jun-Yan Zhu, CMU; (15) Li Fei-Fei, Stanford; (16) Jiajun Wu, Stanford; (17) Stefano Ermon, Stanford; (18) Percy Liang, Stanford. Authors: Authors: (1) Tony Lee, Stanford with Equal contribution; (2) Michihiro Yasunaga, Stanford with Equal contribution; (3) Chenlin Meng, Stanford with Equal contribution; (4) Yifan Mai, Stanford; (5) Joon Sung Park, Stanford; (6) Agrim Gupta, Stanford; (7) Yunzhi Zhang, Stanford; (8) Deepak Narayanan, Microsoft; (9) Hannah Benita Teufel, Aleph Alpha; (10) Marco Bellagente, Aleph Alpha; (11) Minguk Kang, POSTECH; (12) Taesung Park, Adobe; (13) Jure Leskovec, Stanford; (14) Jun-Yan Zhu, CMU; (15) Li Fei-Fei, Stanford; (16) Jiajun Wu, Stanford; (17) Stefano Ermon, Stanford; (18) Percy Liang, Stanford. Table of Links Abstract and 1 Introduction Abstract and 1 Introduction 2 Core framework 2 Core framework 3 Aspects 3 Aspects 4 Scenarios 4 Scenarios 5 Metrics 5 Metrics 6 Models 6 Models 7 Experiments and results 7 Experiments and results 8 Related work 8 Related work 9 Conclusion 9 Conclusion 10 Limitations 10 Limitations Author contributions, Acknowledgments and References Author contributions, Acknowledgments and References A Datasheet A Datasheet B Scenario details B Scenario details C Metric details C Metric details D Model details D Model details E Human evaluation procedure E Human evaluation procedure Author contributions Tony Lee: Co-led the project. Designed the core framework (aspects, scenarios, metrics). Implemented scenarios, metrics and models. Conducted experiments. Contributed to writing. Tony Lee: Michihiro Yasunaga: Co-led the project. Designed the core framework (aspects, scenarios, metrics). Wrote the paper. Conducted analysis. Implemented models. Michihiro Yasunaga: Chenlin Meng: Designed the core framework (aspects, scenarios, metrics). Contributed to writing. Chenlin Meng: Yifan Mai: Implemented the evaluation infrastructure. Contributed to project discussions. Yifan Mai: Joon Sung Park: Designed human evaluation questions. Joon Sung Park: Agrim Gupta: Implemented the detection scenario and metrics. Agrim Gupta: Yunzhi Zhang: Implemented the detection scenario and metrics. Yunzhi Zhang: Deepak Narayanan: Provided expertise and analysis of efficiency metrics. Deepak Narayanan: Hannah Teufel: Provided model expertise and inference. Hannah Teufel: Marco Bellagente: Provided model expertise and inference. Marco Bellagente: Minguk Kang: Provided model expertise and inference. Minguk Kang: Taesung Park: Provided model expertise and inference. Taesung Park: Jure Leskovec: Provided advice on human evaluation and paper writing. Jure Leskovec: Jun-Yan Zhu: Provided advice on human evaluation and paper writing. Jun-Yan Zhu: Li Fei-Fei: Provided advice on the core framework. Li Fei-Fei: Jiajun Wu : Provided advice on the core framework. Jiajun Wu Stefano Ermon: Provided advice on the core framework. Stefano Ermon: Percy Liang: Provided overall supervision and guidance throughout the project. Percy Liang: Statement of neutrality. The authors of this paper affirm their commitment to maintaining a fair and independent evaluation of the image generation models. We acknowledge that the author affiliations encompass a range of academic and industrial institutions, including those where some of the models we evaluate were developed. However, the authors’ involvement is solely based on their expertise and efforts to run and evaluate the models, and the authors have treated all models equally throughout the evaluation process, regardless of their sources. This study aims to provide an objective understanding and assessment of models across various aspects, and we do not intend to endorse specific models. Statement of neutrality. Acknowledgments We thank Robin Rombach, Yuhui Zhang, members of Stanford P-Lambda, CRFM, and SNAP groups, as well as our anonymous reviewers for providing valuable feedback. We thank Josselin Somerville for assisting with the human evaluation infrastructure. This work is supported in part by the AI2050 program at Schmidt Futures (Grant G-22-63429), a Google Research Award, and ONR N00014-23-1- 2355. Michihiro Yasunaga is supported by a Microsoft Research PhD Fellowship. References [1] Percy Liang, Rishi Bommasani, Tony Lee, Dimitris Tsipras, Dilara Soylu, Michihiro Yasunaga, Yian Zhang, Deepak Narayanan, Yuhuai Wu, Ananya Kumar, et al. Holistic evaluation of language models. arXiv preprint arXiv:2211.09110, 2022. [2] Aditya Ramesh, Mikhail Pavlov, Gabriel Goh, Scott Gray, Chelsea Voss, Alec Radford, Mark Chen, and Ilya Sutskever. Zero-shot text-to-image generation. In International Conference on Machine Learning, pages 8821–8831. PMLR, 2021. [3] Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. Hierarchical text-conditional image generation with clip latents. arXiv preprint arXiv:2204.06125, 2022. [4] Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. Highresolution image synthesis with latent diffusion models. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022. [5] Oran Gafni, Adam Polyak, Oron Ashual, Shelly Sheynin, Devi Parikh, and Yaniv Taigman. Make-a-scene: Scene-based text-to-image generation with human priors. arXiv preprint arXiv:2203.13131, 2022. [6] Chitwan Saharia, William Chan, Saurabh Saxena, Lala Li, Jay Whang, Emily Denton, Seyed Kamyar Seyed Ghasemipour, Burcu Karagol Ayan, S Sara Mahdavi, Rapha Gontijo Lopes, et al. Photorealistic text-to-image diffusion models with deep language understanding. arXiv preprint arXiv:2205.11487, 2022. [7] Jiahui Yu, Yuanzhong Xu, Jing Yu Koh, Thang Luong, Gunjan Baid, Zirui Wang, Vijay Vasudevan, Alexander Ku, Yinfei Yang, Burcu Karagol Ayan, et al. Scaling autoregressive models for contentrich text-to-image generation. arXiv preprint arXiv:2206.10789, 2022. [8] Patrick Schramowski, Manuel Brack, Björn Deiseroth, and Kristian Kersting. Safe latent diffusion: Mitigating inappropriate degeneration in diffusion models. arXiv preprint arXiv:2211.05105, 2022. [9] Michihiro Yasunaga, Armen Aghajanyan, Weijia Shi, Rich James, Jure Leskovec, Percy Liang, Mike Lewis, Luke Zettlemoyer, and Wen-tau Yih. Retrieval-augmented multimodal language modeling. In International Conference on Machine Learning (ICML), 2023. [10] Ming Ding, Wendi Zheng, Wenyi Hong, and Jie Tang. Cogview2: Faster and better text-to-image generation via hierarchical transformers. arXiv preprint arXiv:2204.14217, 2022. [11] Zhida Feng, Zhenyu Zhang, Xintong Yu, Yewei Fang, Lanxin Li, Xuyi Chen, Yuxiang Lu, Jiaxiang Liu, Weichong Yin, Shikun Feng, et al. Ernie-vilg 2.0: Improving text-to-image diffusion model with knowledge-enhanced mixture-of-denoising-experts. arXiv preprint arXiv:2210.15257, 2022. [12] Minguk Kang, Jun-Yan Zhu, Richard Zhang, Jaesik Park, Eli Shechtman, Sylvain Paris, and Taesung Park. Scaling up gans for text-to-image synthesis. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2023. [13] Eva Cetinic and James She. Understanding and creating art with ai: review and outlook. ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM), 18(2): 1–22, 2022. [14] Pierre Chambon, Christian Bluethgen, Curtis P Langlotz, and Akshay Chaudhari. Adapting pretrained vision-language foundational models to medical imaging domains. arXiv preprint arXiv:2210.04133, 2022. [15] Midjourney. https://www.midjourney.com/, . [16] Midjourney statistics. https://photutorial.com/midjourney-statistics/, . [17] Ling Yang, Zhilong Zhang, Yang Song, Shenda Hong, Runsheng Xu, Yue Zhao, Yingxia Shao, Wentao Zhang, Bin Cui, and Ming-Hsuan Yang. Diffusion models: A comprehensive survey of methods and applications. arXiv preprint arXiv:2209.00796, 2022. [18] Chenshuang Zhang, Chaoning Zhang, Mengchun Zhang, and In So Kweon. Text-to-image diffusion model in generative ai: A survey. arXiv preprint arXiv:2303.07909, 2023. [19] Abeba Birhane, Vinay Uday Prabhu, and Emmanuel Kahembwe. Ethical considerations of generative AI. AI for Content Creation Workshop, CVPR, 2021. [20] Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A largescale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009. [21] Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Dollár, and C Lawrence Zitnick. Microsoft coco: Common objects in context. In European conference on computer vision, 2014. [22] Scott Reed, Zeynep Akata, Honglak Lee, and Bernt Schiele. Learning deep representations of fine-grained visual descriptions. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 49–58, 2016. [23] Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. Gans trained by a two time-scale update rule converge to a local nash equilibrium. Advances in neural information processing systems, 30, 2017. [24] Jack Hessel, Ari Holtzman, Maxwell Forbes, Ronan Le Bras, and Yejin Choi. Clipscore: A reference-free evaluation metric for image captioning. arXiv preprint arXiv:2104.08718, 2021. [25] Sharon Zhou, Mitchell Gordon, Ranjay Krishna, Austin Narcomey, Li F Fei-Fei, and Michael Bernstein. Hype: A benchmark for human eye perceptual evaluation of generative models. Advances in neural information processing systems, 32, 2019. [26] Jiazheng Xu, Xiao Liu, Yuchen Wu, Yuxuan Tong, Qinkai Li, Ming Ding, Jie Tang, and Yuxiao Dong. Imagereward: Learning and evaluating human preferences for text-to-image generation. arXiv preprint arXiv:2304.05977, 2023. [27] Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018. [28] Yaru Hao, Zewen Chi, Li Dong, and Furu Wei. Optimizing prompts for text-to-image generation. arXiv preprint arXiv:2212.09611, 2022. [29] Jaemin Cho, Abhay Zala, and Mohit Bansal. Dall-eval: Probing the reasoning skills and social biases of text-to-image generative transformers. arXiv preprint arXiv:2202.04053, 2022. [30] Evelina Leivada, Elliot Murphy, and Gary Marcus. Dall-e 2 fails to reliably capture common syntactic processes. arXiv preprint arXiv:2210.12889, 2022. [31] Colin Conwell and Tomer Ullman. Testing relational understanding in text-guided image generation. arXiv preprint arXiv:2208.00005, 2022. [32] Tristan Thrush, Ryan Jiang, Max Bartolo, Amanpreet Singh, Adina Williams, Douwe Kiela, and Candace Ross. Winoground: Probing vision and language models for visio-linguistic compositionality. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5238–5248, 2022. [33] Federico Bianchi, Pratyusha Kalluri, Esin Durmus, Faisal Ladhak, Myra Cheng, Debora Nozza, Tatsunori Hashimoto, Dan Jurafsky, James Zou, and Aylin Caliskan. Easily accessible text-to-image generation amplifies demographic stereotypes at large scale. arXiv preprint arXiv:2211.03759, 2022. [34] Morgan King. Harmful biases in artificial intelligence. The Lancet Psychiatry, 9(11):e48, 2022. [35] Mayu Otani, Riku Togashi, Yu Sawai, Ryosuke Ishigami, Yuta Nakashima, Esa Rahtu, Janne Heikkilä, and Shin’ichi Satoh. Toward verifiable and reproducible human evaluation for text-toimage generation. arXiv preprint arXiv:2304.01816, 2023. [36] Tim Salimans, Ian Goodfellow, Wojciech Zaremba, Vicki Cheung, Alec Radford, and Xi Chen. Improved techniques for training gans. Advances in neural information processing systems, 29, 2016. [37] Nicholas Carlini, Jamie Hayes, Milad Nasr, Matthew Jagielski, Vikash Sehwag, Florian Tramer, Borja Balle, Daphne Ippolito, and Eric Wallace. Extracting training data from diffusion models. arXiv preprint arXiv:2301.13188, 2023. [38] Nupur Kumari, Bingliang Zhang, Sheng-Yu Wang, Eli Shechtman, Richard Zhang, and Jun-Yan Zhu. Ablating concepts in text-to-image diffusion models. In International Conference on Computer Vision (ICCV), 2023. [39] Rohit Gandikota, Joanna Materzynska, Jaden Fiotto-Kaufman, and David Bau. Erasing concepts from diffusion models. 2023. [40] Christoph Schuhmann, Romain Beaumont, Richard Vencu, Cade Gordon, Ross Wightman, Mehdi Cherti, Theo Coombes, Aarush Katta, Clayton Mullis, Mitchell Wortsman, et al. Laion-5b: An open large-scale dataset for training next generation image-text models. arXiv preprint arXiv:2210.08402, 2022. [41] Richard P Taylor, Adam P Micolich, and David Jonas. Fractal analysis of pollock’s drip paintings. Nature, 399(6735):422–422, 1999. [42] Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to-end object detection with transformers. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part I 16, pages 213–229. Springer, 2020. [43] Yanghao Li, Hanzi Mao, Ross Girshick, and Kaiming He. Exploring plain vision transformer backbones for object detection. arXiv preprint arXiv:2203.16527, 2022. [44] Nudenet. https://github.com/notAI-tech/NudeNet. [45] Dreamlike diffusion 1.0. https://huggingface.co/dreamlike-art/ dreamlike-diffusion-1.0, . [46] Dreamlike photoreal 2.0. https://huggingface.co/dreamlike-art/ dreamlike-photoreal-2.0, . [47] Openjourney. https://huggingface.co/prompthero/openjourney, . [48] Openjourney-v4. https://huggingface.co/prompthero/openjourney-v4, . [49] Redshift diffusion. https://huggingface.co/nitrosocke/redshift-diffusion. [50] Vintedois (22h) diffusion. https://huggingface.co/22h/vintedois-diffusion-v0-1. [51] Lexica. https://lexica.art/docs. [52] Dall-e mini. https://github.com/borisdayma/dalle-mini. [53] Saehoon Kim, Sanghun Cho, Chiheon Kim, Doyup Lee, and Woonhyuk Baek. mindall-e on conceptual captions. https://github.com/kakaobrain/minDALL-E, 2021. [54] Marco Bellagente, Manuel Brack, Hannah Teufel, Felix Friedrich, Björn Deiseroth, Constantin Eichenberg, Andrew Dai, Robert Baldock, Souradeep Nanda, Koen Oostermeijer, Andres Felipe Cruz-Salinas, Patrick Schramowski, Kristian Kersting, and Samuel Weinbach. Multifusion: Fusing pre-trained models for multi-lingual, multi-modal image generation, 2023. [55] deep-floyd. https://github.com/deep-floyd/IF. [56] Pranav Rajpurkar, Jian Zhang, Konstantin Lopyrev, and Percy Liang. Squad: 100,000+ questions for machine comprehension of text. In Empirical Methods in Natural Language Processing (EMNLP), 2016. [57] Kawin Ethayarajh and Dan Jurafsky. Utility is in the eye of the user: A critique of nlp leaderboards. arXiv preprint arXiv:2009.13888, 2020. [58] Inioluwa Deborah Raji, Emily M Bender, Amandalynne Paullada, Emily Denton, and Alex Hanna. Ai and the everything in the whole wide world benchmark. arXiv preprint arXiv:2111.15366, 2021. [59] Alex Wang, Amanpreet Singh, Julian Michael, Felix Hill, Omer Levy, and Samuel R Bowman. Glue: A multi-task benchmark and analysis platform for natural language understanding. arXiv preprint arXiv:1804.07461, 2018. [60] Pang Wei Koh, Shiori Sagawa, Henrik Marklund, Sang Michael Xie, Marvin Zhang, Akshay Balsubramani, Weihua Hu, Michihiro Yasunaga, Richard Lanas Phillips, Irena Gao, et al. Wilds: A benchmark of in-the-wild distribution shifts. In International Conference on Machine Learning, pages 5637–5664. PMLR, 2021. [61] Aarohi Srivastava, Abhinav Rastogi, Abhishek Rao, Abu Awal Md Shoeb, Abubakar Abid, Adam Fisch, Adam R Brown, Adam Santoro, Aditya Gupta, Adrià Garriga-Alonso, et al. Beyond the imitation game: Quantifying and extrapolating the capabilities of language models. arXiv preprint arXiv:2206.04615, 2022. [62] Chengwei Qin, Aston Zhang, Zhuosheng Zhang, Jiaao Chen, Michihiro Yasunaga, and Diyi Yang. Is chatgpt a general-purpose natural language processing task solver? arXiv preprint arXiv:2302.06476, 2023. [63] Tero Karras, Samuli Laine, and Timo Aila. A style-based generator architecture for generative adversarial networks. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019. [64] Suman Ravuri and Oriol Vinyals. Classification accuracy score for conditional generative models. Advances in neural information processing systems, 32, 2019. [65] Tuomas Kynkäänniemi, Tero Karras, Samuli Laine, Jaakko Lehtinen, and Timo Aila. Improved precision and recall metric for assessing generative models. Advances in Neural Information Processing Systems, 32, 2019. [66] Gaurav Parmar, Richard Zhang, and Jun-Yan Zhu. On aliased resizing and surprising subtleties in gan evaluation. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022. [67] 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 generation. arXiv preprint arXiv:2305.01569, 2023. [68] Rudolf Arnheim. Art and visual perception: A psychology of the creative eye. Univ of California Press, 1969. [69] Philip Galanter. Computational aesthetic evaluation: past and future. Computers and creativity, pages 255–293, 2012. [70] William Lidwell, Kritina Holden, and Jill Butler. Universal principles of design, revised and updated: 125 ways to enhance usability, influence perception, increase appeal, make better design decisions, and teach through design. Rockport Pub, 2010. [71] Catherine Wah, Steve Branson, Peter Welinder, Pietro Perona, and Serge J. Belongie. The caltech-ucsd birds-200-2011 dataset. 2011. [72] Caleb Ziems, Jiaao Chen, Camille Harris, Jessica Anderson, and Diyi Yang. Value: Understanding dialect disparity in nlu, 2022. [73] 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 supervision. In International conference on machine learning, pages 8748–8763. PMLR, 2021. [74] Yogesh Balaji, Seungjun Nah, Xun Huang, Arash Vahdat, Jiaming Song, Karsten Kreis, Miika Aittala, Timo Aila, Samuli Laine, Bryan Catanzaro, et al. ediffi: Text-to-image diffusion models with an ensemble of expert denoisers. arXiv preprint arXiv:2211.01324, 2022. [75] Christian Szegedy, Wei Liu, Yangqing Jia, Pierre Sermanet, Scott Reed, Dragomir Anguelov, Dumitru Erhan, Vincent Vanhoucke, and Andrew Rabinovich. Going deeper with convolutions. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 1–9, 2015. [76] Maximilian Seitzer. pytorch-fid: FID Score for PyTorch. https://github.com/mseitzer/pytorch-fid, August 2020. Version 0.3.0. [77] Jonathan Ho and Tim Salimans. Classifier-free diffusion guidance. arXiv preprint arXiv:2207.12598, 2022. [78] Anton Obukhov, Maximilian Seitzer, Po-Wei Wu, Semen Zhydenko, Jonathan Kyl, and Elvis YuJing Lin. High-fidelity performance metrics for generative models in pytorch, 2020. URL https: //github.com/toshas/torch-fidelity. Version: 0.3.0, DOI: 10.5281/zenodo.4957738. [79] Deepak Narayanan, Keshav Santhanam, Peter Henderson, Rishi Bommasani, Tony Lee, and Percy Liang. Cheaply Evaluating Inference Efficiency Metrics for Autoregressive Transformer APIs. arXiv preprint arXiv:2305.02440, 2023. [80] Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, et al. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. [81] Yuxin Wu, Alexander Kirillov, Francisco Massa, Wan-Yen Lo, and Ross Girshick. Detectron2. https://github.com/facebookresearch/detectron2, 2019. [82] Tim Salimans and Jonathan Ho. Progressive distillation for fast sampling of diffusion models. arXiv preprint arXiv:2202.00512, 2022. [83] Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. Language models are unsupervised multitask learners. OpenAI blog, 1(8):9, 2019. [84] Minwoo Byeon, Beomhee Park, Haecheon Kim, Sungjun Lee, Woonhyuk Baek, and Saehoon Kim. Coyo-700m: Image-text pair dataset. https://github.com/kakaobrain/ coyo-dataset, 2022. This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv