LightRicks Introduce Open-Source Ltx Video Ai Model With Real-Time Video Generation Capability

Lightricks, a software company focused on image and video editing, released an open-source Artificial intelligence (Ai) Video model in preview last week. Dubbed ltx video, the AI ​​Model Can Generate Medium-Resolution Videos in Real Time. While Real-Time Video Generation Capability Exists in a Few Large Language Models (LLMS), this is the first one to be open-sourced. The company also stated that once the full version is released, it will be free for both personal and commercial use and can be integrated into the ltx studio.

Lightricks Introduces Open-Source Ai Video Model

In a series of posts On X (Formerly Known as Twitter), LightRicks Detailed Its Open-Source Ai Model. The ltx video accepts both text and images as input and can generate five-second-live videos in 768 x 512p resolution. While the preview model caps the video quality at Medium Resolution, It offers Near Real-Time Generation with Four Seconds of Wait Time. However, this generation time is possible on devices equipped with the Nvidia H100 Chip.

The company claims that the ai model can generate dynamic videos with high Prompt adherence and does not require high-end resources to run. To locally run the ltx video, users will need a gpu similer to the level of RTX 4090. Lightrick Billion parameters to keep its size small.

Ltx video is currently available on github, hugging face, and comfyui to download. To test the model’s capabilityes before downloading it from the hosting websites, users can go to the model page on fal. here,

The AI ​​model can also be integrated with a wide range of external editing tools to further fin-tune the generated videos. The company also plans to release the full version of the video model and make it open source for both personal use cases and commercial usage. The full version of the tool will also be integrated with the ltx studio, the company’s AI-Powered Storyboard Platform.