PicsArt Unveils StreamingT2V: A New AI Model for Generating LongVideos
San Francisco, CA – PicsArt, the popular photo andvideo editing app, has announced the release of StreamingT2V, a groundbreaking AI model capable of generating videos up to two minutes long. This new technology addressesa major challenge in the field of text-to-video generation, where existing models often struggle to produce high-quality videos beyond a few seconds.
StreamingT2V utilizes a novel approach that combines conditional attention modules (CAM) and appearance preservation modules (APM) with a random mixing method. This allows the model to generate long videos with seamless transitions and consistent visual elements, ensuring that thegenerated content remains coherent and closely aligned with the input text description.
We are excited to introduce StreamingT2V, a significant advancement in text-to-video generation, said [Name of PicsArt representative], [
【source】https://ai-bot.cn/streamingt2v/
Views: 1