Meet Microsoft's Phi-3 Vision, a small multimodal AI model capable of processing images and videos
Phi-3 Vision has 4.2 billion parameters.
1 min. read
Published on
Share this article
Read our disclosure page to find out how can you help Windows Report sustain the editorial team. Read more
Earlier this year, Microsoft released Phi-3 Mini, an AI model with 3.8 billion parameters, which can run locally, rather than on-cloud, and it’s quite impressive.
Today, at the Microsoft Build 2024, the Redmond-based tech giant announced the new Phi-3 Vision, which can accept images and videos as prompts, and it can generate answers based on them.
Microsoft made Phi-3 Vision bigger than its Mini version: it has 4.2 billion parameters, and it’s now available in a private preview.
However, there is no word when it will become generally available, but the Redmond-based tech giant will most likely release it later this year.
Speaking of AI, Microsoft Teams will also get a Team Copilot model, and similar to Google’s AI Teammate, this one will help users access information easily and faster.
Microsoft wasn’t joking when it said this year’s Build is all about AI.
User forum
0 messages