ChatGPT and Midjourney are now making news not only in the tech media but also in other areas. This might change soon since the next big AI tech is on the way. NVIDIA has demoed a new tech that might soon elevate your GIFs to new heights.
Gizchina News of the week
Use NVIDIA’s Text-To-Video Tech To Create Videos
Not long ago, NVIDIA unleashed an interesting product dubbed VideoLDM. That’s an AI model for text-to-video conversion. By the way, NVIDIA has created this tech by joining hands with Cornell University researchers. In simple terms, this AI model can generate videos with a resolution of up to 2048 x 1280 pixels, a frame rate of 24 frames per second, and a runtime of up to 4.7 seconds based on text.
The model uses advances in stable diffusion neural networks. The NVIDIA solution has up to 4.1 billion parameters. But only 2.7 billion were trained using video. Though you might think this is a huge number, it is small by today’s AI standards.
Also, the model can capture videos with driving scenes. They can last up to 5 minutes and have a 1024 x 512 pixels resolution. This is because driving sequences are tough to shoot in real life. As a result, when it comes to such moments, this technology can assist film makers. The model can forecast multimodal motion possibilities by creating a large number of possible missions from a single source image.
At the moment, text-to-video techs such as the one introduced by NVIDIA are good to use for creating thumbnails and GIFs. But we know that NVIDIA is growing fast in the AI field. So we guess it won’t take long for NVIDIA to bring more advanced techs for creating longer text-to-video clips.
This study will be presented at the Machine Vision and Pattern Recognition Conference, which will be held in Vancouver from June 18 to 22.