Following the debut of its beta version in March, Adobe Firefly was met with positive feedback from users with its suite of AI-powered services.
This time, the second generation of the beta software is introducing a new set of models with over 100 unique features and improvements, but a select few stand out.
Here are some of the new features users are raving about from Adobe Firefly 2.
Generative Match allows Adobe Firefly 2 to produce multiple images that match the certain feel and style of uploaded reference images.
X user Faisal Shariff calls it the "killer feature" in the latest software update, as he shares his creations with Adobe Firefly 2 in comparison to images he produced with DALL-E, noting Firefly's ability to mimic style accurately.
1 / Here’s an image that I created in DALL-E 3. London’s Canary Wharf. pic.twitter.com/Dt1372OciY— Faisal Shariff (@thetechfai) October 11, 2023
When it comes to AI-generated images, one of the key issues includes AI's inability to produce realistic facial features, often smoothening skin, exaggerating a subject's focus, and more.
But with Firefly's new Photo Settings, users can toggle an image's aperture, depth, level of blur, and field of view to produce a desired outcome.
Through it, users can also tweak details like pores, skin tone, and more to create a more hyperrealistic image.
Oftentimes, one word can change the whole outcome of an AI-generated image.
This is why Adobe Firefly 2's new Prompt Suggestions help the user by auto-filling words to complete their prompts, producing different images depending on the suggestions they choose.
Amazing work from the Firefly team 👏— Flemming Miguel (@flemmingmiguel) October 10, 2023
The model is very capable and greatly avoids producing copyrighted materials
The output quality is impressive and the tooling around it is very mature, specially prompt suggestions and styles 🤌
Excited for the next models to come