Yapay Zekâ & Teknoloji · Goril Radar’da yayın:
Elorian’s Glimpse model aims to reason by re-examining images
Kaynak: TechCrunch
Kaynak yayın tarihi:
Elorian has launched its first model, Glimpse, as a research preview with API access and a chat interface. According to TechCrunch, the model can re-examine relevant regions of an image during reasoning, select frames from video, and use code for comparison.
The company describes this approach as a “visual workspace”. The aim is to refer back to visual evidence throughout the process rather than merely converting an image into text to interpret it; performance claims should, however, be treated as the company’s own evaluations.
TechCrunch: Orijinal haberi oku
Bu metin, kaynak haberdeki doğrulanabilir bilgilere dayanılarak hazırlanan özgün bir editoryal özettir.