Topics
late
AI
Amazon
Image Credits:Ron Miller/TechCrunch
Apps
Biotech & Health
Climate
Image Credits:Ron Miller/TechCrunch
Cloud Computing
Commerce
Crypto
Image Credits:Amazon
endeavor
EVs
Fintech
Image Credits:Amazon
fund raise
Gadgets
stake
Government & Policy
Hardware
layoff
Media & Entertainment
Meta
Microsoft
Privacy
Robotics
security measure
societal
blank space
startup
TikTok
Transportation
speculation
More from TechCrunch
Events
Startup Battlefield
StrictlyVC
Podcasts
video
Partner Content
TechCrunch Brand Studio
Crunchboard
Contact Us
Amazon has released an upgraded version of its in - house image - sire mannikin , Titan Image Generator , for AWS customers using its Bedrock procreative AI platform .
Simply called Titan Image Generator v2 , the new model brings with it several new capabilities , AWS principal developer urge Channy Yun explainsin a blog mail . drug user can “ channelize ” the images they get using reference images , edit existing visuals , remove backgrounds and engender edition of image , state Yun .
“ Titan Image Generator v2 can intelligently detect and segment multiple foreground objects , ” Yun write . “ With the Titan Image Generator v2 , you’re able to generate colour - conditioned images based on a colour palette . [ And ] you could utilise the image conditioning feature to shape your creations . ”
Titan Image Generator v2 brook image conditioning , optionally taking in a reference range and concenter on specific visual characteristics in that range , like border , object outlines and structural element . The model can also be finely - tune up using acknowledgment images like a Cartesian product or company logo , so that return image maintain a ordered aesthetical .
AWS continues to remain vague about which data point , exactly , it uses to train its Titan Image Generator example . The company antecedently told TechCrunch only that it ’s a combination of proprietary and accredited data .
Few vendors readily unveil such information ; they see preparation data as a free-enterprise vantage and thus keep it and info relating to it a closely guarded enigma . preparation data detail are also a potential source ofIP - have-to doe with lawsuits , another disincentive to let out much .
In stead of transparency , AWS offers anindemnification policythat covers client in the event a Titan model like Titan Image Generator v2regurgitates(i.e . spits out a mirror copy of ) a potentially copyright training instance .
Join us at TechCrunch Sessions: AI
Exhibit at TechCrunch Sessions: AI
In the company ’s recent second - fourth net income call , Amazon CEO Andy Jassy tell he ’s still “ very bullish ” on generative AI tech like AWS ’ Titan model , despite mansion of secondly - guess from the enterprisingness and the mounting cost related to preparation , fine - tuning and serving modelling .
“ In the generative AI quad , it ’s going to get big tight , ” he say , “ and it ’s largely all belong to be built from the get - go in the cloud . ”
