Adobe comes out with a killer weapon! Bringing 'Firefly' into the game, full of professional buffs, or Immaculate Conception AI

AI’s iPhone moment has arrived.

#At the GTC conference, Lao Huang officially announced that he and Adobe would release the artificial intelligence image generation tool Firefly.

The announcement stated that Firefly has been integrated into Adobe’s enterprise-level creative tool Adobe Express. In closed beta, only the word art function is enabled.

#Currently, it is not open to users, but you can apply for a waitlist.

In the future, this tool will be added to Adobe’s Photoshop, Illustrator and Premiere.

# "With Firefly, everyone can create the content of their dreams in their own words."

Address: https://firefly.adobe.com/

"Firefly" flies into the AIGC battle

Currently, Firefly provides two modes, one is the function of using text to generate images, and the other is font effect editing.

In terms of drawing, Firefly is similar to the currently popular AI drawing tools. It can generate images based on prompts, provide hundreds of styles and adjust the results.

In addition, Firefly can also generate images based on surrounding content, which allows users to more effectively to test and refine image effects.

#In graphic design and illustration work, custom vectors, brushes, and textures can be generated from a few words or even sketches, with the ability to edit and create using familiar tools Content.

#In addition to basic text-to-image generation, the Firefly model can "transfer" different styles to existing images like Prisma. It can also apply styles or textures to letters and fonts, based on user-provided descriptions.

#On the creative text tool, people can create text that looks furry, scaly, and other shapes and effects, similar to the "Word Art" function in Office .

##Firefly can also change 3D modeling, such as quickly generating model styles and materials:

The maximum resolution image that Firefly can currently produce is 2000*2000 pixels.

In addition, Firefly is still exploring other features, including text generation vectors/brushes/templates, generating images according to your own style, and specifying aspect ratio expansion. images, generate images from 3D models, etc.

"Immaculate Conception" model

Here, Adobe said that works created by the Firefly model will include metadata to prove that it is part Or all generated by artificial intelligence.

#The reason for doing this is mainly to protect the artist's copyright issues.

In terms of technology, Firefly is no different from DALL-E2 and text-to-image AI like Stable Diffusion.

These AI models are trained by learning to generate new images from existing images, often from data sets that are searched through Cobbled together by public image hosting sites

These AI models can generate new ones by "learning" from textual prompts by "training" on existing images. image.

#These existing images typically come from datasets obtained by searching public image hosting sites.

# Some experts believe that using public images (even copyrighted ones) to train models would be protected by the fair use doctrine in the United States. However, this is an issue that is unlikely to be resolved anytime soon.

Midjourney and Stability AI, previously accused of violating the rights of millions of artists by using their tools on web-scraped images, are facing Legal action.

Stock photo provider Getty Images has reportedly taken Stability AI to court separately , accused it of using millions of images from its website without permission to train the art generative model Stable Diffusion.

In addition to unresolved questions about artist and platform compensation, there is a more pressing issue facing generative AI:

It tends to use copied images, text and other content, including copyrighted content, from the data used to train the model.

# Some image hosting platforms have banned the use of AI-generated content because they fear legal disputes.

Experts warn that businesses could be put at risk if they inadvertently incorporate copyrighted content generated by generated AI tools into any products they sell. Will face risks.

Adobe’s solution is to train Firefly using only Adobe Stock, the company’s royalty-free media library, and publicly licensed and public domain content. Model.

Adobe said that in the future users will be able to train and fine-tune Firefly models with their own content so that their output has a specific style and design language.

The above is the detailed content of Adobe comes out with a killer weapon! Bringing 'Firefly' into the game, full of professional buffs, or Immaculate Conception AI. For more information, please follow other related articles on the PHP Chinese website!