ellibsЭлектронные библиотекиRussian Digital Libraries Journal1562-5419Казанский (Приволжский) федеральный университет10.26907/1562-5419-2023-26-4-437-455ellibs-382Research ArticleСтатьиНейронная сеть для генерации изображений на основе текста песен с применением моделей OpenAI и CLIPNeural Network for Generating Images Based on Song Lyrics using OpenAI and CLIP ModelsДавлетгарееваА. Р.DavletgareevaA. R.alsudavletgareeva@gmail.comЕдковаК. А.EdkovaK. A.ksushka.e21@gmail.comКазанский (Приволжский) Федеральный университетKazan (Volga region) Federal University202328082023264437–455Copyright © Давлетгареева А.Р., Едкова К.А., 20232023Давлетгареева А.Р., Едкова К.А.Davletgareeva A.R., Edkova K.A.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://ellibs.elpub.ru/jour/article/view/382

Исследована эффективность моделей ImageNet diffusion model и CLIP для генерации изображений по текстовому описанию. С использованием различных текстовых вводов на разных параметрах проведены два эксперимента для определения лучших параметров при генерации изображений на основе текстового описания. Результаты показали, что, хотя ImageNet хорошо справляется с созданием изображений, CLIP лучше обеспечивает соединение текстовых подсказок с релевантными изображениями. Полученные результаты характеризуют высокий потенциал объединения названных моделей для создания высококачественных и контекстно релевантных изображений на основе текстового описания.

The effectiveness of the ImageNet diffusion model and CLIP models for image generation based on textual descriptions was investigated. Two experiments were conducted using various textual inputs and different parameters to determine the optimal settings for generating images from text descriptions. The results showed that while ImageNet performed well in generating images, CLIP demonstrated better alignment between textual prompts and relevant images. The obtained results highlight the high potential of combining these mentioned models for creating high-quality and contextually relevant images based on textual descriptions.

генерация изображенийглубокое обучениенейронные сетиобработка естественного языкаImageNet diffusion modelCLIPReferencesElasri M., Elharrouss O., Al-Maadeed S., Tairi H. Image Generation: A Review // Neural Processing Letters. 2022. Vol. 54. No. 5. P. 4609–4646.Elasri M., Elharrouss O., Al-Maadeed S., Tairi H. Image Generation: A Review // Neural Processing Letters. 2022. Vol. 54. No. 5. P. 4609–4646.Zhang H., Song H., Li S., Zhou M., Song D. A survey of controllable text generation using transformer-based pre-trained language models // arXiv preprint arXiv:2201.05337. 2022Zhang H., Song H., Li S., Zhou M., Song D. A survey of controllable text generation using transformer-based pre-trained language models // arXiv preprint arXiv:2201.05337. 2022Основы генеративно-состязательных сетей. URL: https://habr.com/ru/articles/726254/Основы генеративно-состязательных сетей. URL: https://habr.com/ru/articles/726254/Brown T., Mann B., Ryder N., Subbiah M., Kaplan J. D., Dhariwal P., Dhariwal P., Neelakantan A., Shyam P., Sastry G., Askell. A, Agarwal S., Herbert-Voss A., Krueger G., Henighan T., Child R., Ramesh A., Ziegler D.M., Wu J., Winter C., Hesse C., Chen M., Sigler E., Litwin M., Gray S., Chess B., Clark J., Berner C., McCandlish S., Radford A., Sutskever I., Amodei D. Language models are few-shot learners // Advances in neural information processing systems. 2020. Vol. 33. P. 1877–1901.Brown T., Mann B., Ryder N., Subbiah M., Kaplan J. D., Dhariwal P., Dhariwal P., Neelakantan A., Shyam P., Sastry G., Askell. A, Agarwal S., Herbert-Voss A., Krueger G., Henighan T., Child R., Ramesh A., Ziegler D.M., Wu J., Winter C., Hesse C., Chen M., Sigler E., Litwin M., Gray S., Chess B., Clark J., Berner C., McCandlish S., Radford A., Sutskever I., Amodei D. Language models are few-shot learners // Advances in neural information processing systems. 2020. Vol. 33. P. 1877–1901.DALL⋅E 2. URL:https://openai.com/product/dall-e-2.DALL⋅E 2. URL:https://openai.com/product/dall-e-2.How AI is Transforming Text-to-Image Generation. URL: https://nesesho.com/index.php/2023/04/12/how-ai-is-transforming-text-to- image-generation/How AI is Transforming Text-to-Image Generation. URL: https://nesesho.com/index.php/2023/04/12/how-ai-is-transforming-text-to- image-generation/OpenAI⋅GitHub. URL: https://github.com/openai.OpenAI⋅GitHub. URL: https://github.com/openai.Gulrajani I., Ahmed F., Arjovsky M., Dumoulin V., Courville A.C. Improved training of wasserstein GANs // Advances in neural information processing systems. 2017. Vol. 30. P. 5767–5777.Gulrajani I., Ahmed F., Arjovsky M., Dumoulin V., Courville A.C. Improved training of wasserstein GANs // Advances in neural information processing systems. 2017. Vol. 30. P. 5767–5777.Indolia S., Goswami A.K., Mishra S.P., Asopa P. Conceptual understanding of convolutional neural network-a deep learning approach // Procedia computer science. 2018. Vol. 132. P. 679–688.Indolia S., Goswami A.K., Mishra S.P., Asopa P. Conceptual understanding of convolutional neural network-a deep learning approach // Procedia computer science. 2018. Vol. 132. P. 679–688.Laudani A., Lozito G.M., Fulginei F.R., Salvini A. On training efficiency and computational costs of a feed forward neural network: a review // Computational intelligence and neuroscience. 2015. P. 83–83.Laudani A., Lozito G.M., Fulginei F.R., Salvini A. On training efficiency and computational costs of a feed forward neural network: a review // Computational intelligence and neuroscience. 2015. P. 83–83.CLIP. URL: https://github.com/openai/CLIP.CLIP. URL: https://github.com/openai/CLIP.Dhariwal P., Nichol A. Diffusion models beat gans on image synthesis // Advances in Neural Information Processing Systems. 2021. Vol. 34. P. 8780–8794.Dhariwal P., Nichol A. Diffusion models beat gans on image synthesis // Advances in Neural Information Processing Systems. 2021. Vol. 34. P. 8780–8794.Kim G., Kwon T., Ye J.C. Diffusionclip: Text-guided diffusion models for robust image manipulation // In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022. P. 2426–2435.Kim G., Kwon T., Ye J.C. Diffusionclip: Text-guided diffusion models for robust image manipulation // In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022. P. 2426–2435.

The authors declare that there are no conflicts of interest present.