OpenAI May Use 3D Engine/Unreal Engine to Train Video Model Sora
-
OpenAI recently released several videos generated by Sora, showcasing stunning 3D effects. The content of these videos is rich, including diving to explore shipwrecks and dioramas depicting tranquil scenes from Japan's Edo period.
In addition, Sora has demonstrated some new features, such as seamless transitions. This means gradual interpolation can be performed between two input videos, resulting in smoother transition effects. Some netizens have begun discussing the source of Sora's training data, with the mainstream view being that Sora may have been trained using a 3D engine or Unreal Engine.
Image source note: The image is AI-generated, licensed by Midjourney. These newly released videos once again highlight OpenAI's leading position in the field of artificial intelligence and their continuous efforts in developing cutting-edge technologies. Through constant optimization and updates, Sora's video generation technology has reached astonishing levels, providing users with more vivid and engaging visual experiences.
OpenAI's Sora video generation technology not only demonstrates its innovation capabilities in the AI field but also reveals the potential future of digital media and visual effects. These videos are not only visually stunning but also spark unlimited imagination about AI's applications in creative and entertainment industries.