On Thursday, Adobe launched the latest iteration of redesigned web apps from competitors, in addition to the Firefly family of image-generating AI models, models for generating vectors, and a redesigned web app that houses all AI models. The work also has a mobile app for fireflies.
According to Adobe, the new Firefly Image Model 4 improves its predecessor in terms of quality, speed and power, camera angle, and zoom structure and style control. Images can be generated at a resolution of up to 2k. There is also a tweaked, more capable version of this model called Image Model 4 Ultra, which allows you to render complex scenes that are made up of small structures and many details.
Alexandru Costin, vice president of Adobe’s Generation AI, said that the company trained the models at computing size to allow it to generate more detailed images. He added that compared to previous generation models, the new model has the ability to improve text generation for images and to obtain the model using the image of their choice and generate photos in that style.
The company is also creating a Firefly video model that was launched last year at Limited Beta. You can generate video clips from a text prompt or image, use camera angles, control shots by specifying start and end frames, generate atmospheric elements, and customize motion design elements. This model can generate video clips from text at resolutions up to 1080p.
Firefly Vector models, on the other hand, can create editable vector-based artwork and generate iteratively variations of logos, product packaging, icons, scenes and patterns.
In particular, Firefly’s web apps have access to all these models as well as several image and video generation models from OpenAI (GPT Image Generation), Google (Imagen 3 and Veo 2), and Flux (Flux 1.1 Pro). Users can toggle between any of these models at any time, and images generated from any model will come with content credentials. The company has suggested that other AI models could be added to web apps in the future.
Adobe is also releasing a new product called Firefly Boards, a canvas for ideas and mood boards. Users can generate or import images, remix them, and collaborate with others, just like what they can do with Visual Electric, Cove and Kosmik’s other AI-based idea boards. The board is available through the Firefly web app.
The company said these new models will soon be integrated into their product portfolio, but did not provide a rollout timeline.
Adobe also said that it has made text-to-image and Avatar APIs generally available, and that the new text-to-video APIs are now available in beta. These APIs are available through the company’s Firefly Services collection of APIs, tools and services.
Adobe also allows users to test a web app called Adobe Content Authenticity to attach their credentials to their work and show ownership and belongings via metadata. Users can also indicate whether AI companies can use any image for AI model training.
Source link
