Facebook owner Meta (META.O) released an artificial intelligence model to identify specific items in images and the biggest picture annotation collection ever.
In a blog post, the company’s research division stated its Segment Anything Model, or SAM, could recognize things in photos and videos even if it hadn’t seen them before.
Clicking or typing prompts choose items in SAM. For example, writing “cat” caused the tool to box multiple cats in a picture.
Since Microsoft-backed (MSFT.O) OpenAI’s ChatGPT chatbot went viral in the fall, big tech firms have been touting their artificial intelligence advances.
Meta has teased numerous features that use ChatGPT-style generative Intelligence, which produces new content rather than detecting or categorizing input.
A technology that creates children’s book pictures from writing and a surrealist video generator are examples.
Mark Zuckerberg wants to add generative AI “creative assistance” to Meta’s applications this year.
Meta already uses SAM-like technologies to classify photographs, moderate content, and recommend Facebook and Instagram postings.
The business claimed SAM would increase access to that technology.
Download the non-commercial SAM model and dataset. Anyone adding photographs to a prototype must also consent to use it for study.
Meta Platforms, Inc., previously Facebook, Inc. and TheFacebook, Inc., is a Menlo Park, California-based international technological company. The corporation owns Facebook, Instagram, WhatsApp, and others.
Comment Template