LLAVA: The AI That Microsoft Didn't Want You to Know About!

preview_player
Показать описание
Meet LLaVA, the groundbreaking AI model developed by UC Davis and Microsoft Research that can both chat and understand images. This advanced AI outperforms GPT-4 and can help you with a range of tasks, from answering questions to generating images and even offering creative solutions. Discover how LLaVA's vision encoder and language decoder work together to make it the most versatile and powerful AI model yet.

#llava #ai #microsoft
Рекомендации по теме
Комментарии
Автор

Incredible video. If I could make a recommendation on something I would love to see, I would say that having a section or seperate video that teaches a user how to initially set up the tool and use the tool would be incredible. Thank you again for your incredible content.

ericduhon
Автор

I have been working with LLava for a little while now. For some reason, it misses or misinterprets parts of an image. It either does not see the missing areas, hallucinates what it does see, or is blind to major parts. However, when it works, it works very well. It does seem to see colors, but sometimes it gets the colors wrong---it will see a bright red as blue, as an example. When I showed it a black & white picture that Dorothea Lange produced of a poor rural family in the South, LLava spotted the image as one of a series of such monochrome or tinted images taken by Lange. It also gave me a brief history of Dorothea Lange. When I showed it a caricature of Bing Crosby, it recognized the image both as a caricature, and as the image of the well-known singer. LLava included a listing of several of his top song hits such as White Christmas, etc. Very promising!

elysilk
Автор

I can imagine it can interprete electrical distribution box photo and answer to question what is wrong and which swich should I turn on/off

chanpasadopolska
Автор

Why doesnt Microsoft want you to know about it. They released public papers and all over the net.

Also it wasnt to make up for the shortcomings of gpt4 vision, but rather to improvs research technologies. Lets not make it sound different from as it os

Lorentz_Factor
Автор

Sorry had to 2 minutes in, interesting model but your US shock documentary style had my eyes rolling so hard I have a headache

jgcornell
Автор

A load of bamboozling definitions - multimodal - blah blah. For heavens sake, show us it actually working. I am non the wiser about LLAVA after watching this management-team video. They think I am a BMW or HP management-team buying it. Sorry, purchasing it.

user_a