Facebook brand company Meta has launched a new generative AI model CM3leon. This platform is the best performance experience for text-to-image and image-to-text generation. According to the news of IANS, this platform has come when AI (Artificial Intelligence) run image generator has become very popular and easy. Today big companies are using it everyday.
Image creation gets even better
Meta explained in a blogpost that CM3leon is the first multimodal model, which has been trained from a text-only language model with a customized recipe. Image creation is made even better by the features present in CM3leon. According to Meta, CM3leon requires only five times the computing power and a smaller training dataset than previous transformer-based methods. The tech giant said that the CM3leon achieves the best results across a wide range of visual language tasks, such as visual question answering and long-form captioning.
Meta’s initiative to create great images
Meta said that with the goal of creating a high-quality generator model, we believe that the CM3leon is a step towards creating an image of strong performance in different tasks. The CM3Leon can also understand instructions for editing an existing photo. Although Meta hasn’t said whether it plans to release the CM3Leon.
The company writes in its blog that as the AI industry continues to evolve, generator models such as the CM3Leon are becoming increasingly refined. While the industry is still in the early stages of understanding and addressing these challenges, we believe transparency will be key to accelerating development.