HiDream Foundation Model 1.0 — это масштабная мультимодальная система, объединяющая генерацию текста, изображений, видео и 3D-контента. Модель построена на архитектуре Diffusion Transformer (DiT) и содержит более миллиарда параметров, открывая новые горизонты для визуального творчества.
The large image multi-model model is the large model developed independently by the future of (HiDream.ai). It is the first image and video generated by the world to be open for online use to generate Diffusion Transformer (DiT) architecture model. The scale of the model parameters has exceeded 1 billion, realizing the joint modeling of text, image, video, 3D It has been filed through both the model and the algorithm. Based on the ” Imagine AI“ series of products constructed by the large image model ”, the functions of image generation editing, 4K HD picture, global/local controllable, script multi-lens video generation, etc. have been realized in AIGC technology and digital creativity. Has a leading advantage in commercialization.