On October 22, Genmo, the AI based video generating platform, released Mochi 1, a new state-of-the-art open-source text-to-video generation model. It can generate high quality videos with mere text prompts.
In addition to Mochi 1, the company also unveiled its hosted playground, where users can try Mochi 1 for free. The weights and architecture for Mochi 1 are open and available on HuggingFace.
However, the company gained traction soon after its release as Genmo acknowledged the website crash, attributing it to the rise in traffic. “We’re seeing extremely high load. Some users are seeing errors, please bear with us as we scale up capacity,” the company announced on X.
We are so back!
We added a ton of GPUs to the playground and it's healthy again. Use it now: https://t.co/NWjmchepSs
Everyone hugged our servers so hard they melted. Thank you to all who waited for things to scale up.
Please create and share your best prompts and generations! pic.twitter.com/T7bOjsjseP— Genmo (@genmoai) October 23, 2024
The company claims Mochi 1 excels in realistic motion dynamics and follows text instructions.
Trained with a 10 billion parameter diffusion model, Mochi 1 creates videos at 30 frames per second. Genmo has released the model under a permissive Apache 2.0 license.
Mochi 1 is similar to proprietary models, including Runway’s Gen3 Alpha, Luma AI’s Dream Machine , Kuaishou’s Kling, and Minimax’s Hailuo , among others.
“At Genmo, our mission is to unlock the right brain of artificial general intelligence. Mochi 1 is the first step toward building world simulators that can imagine anything, whether possible or impossible,” says the Genmo team.
In the research preview of Mochi 1, it unlocked multiple possibilities across different fields. Mochi 1 advances video generation techniques and fosters exploration of novel methodologies in research and development.
For product development, it supports innovative applications in entertainment, advertising, and education. It also enables artists and creators to express their visions through AI-generated videos, expanding creative capabilities.
Additionally, in robotics, it aids in generating synthetic data for training AI models in autonomous vehicles, robotics, and virtual environments, catering to various industries.
Genmo has partnered with top platforms to ensure Mochi 1 is easily accessible to developers. Additionally, it can integrate Mochi 1 into applications seamlessly using APIs from partner platforms, offering a straightforward way to leverage its capabilities in projects.
The post Genmo launches Mochi 1 Text-to-Video Generation Model, But Server Crashes Within Hours appeared first on AIM.
