Oriol Vinyals: Flamingo, a drastic VLM that achieve...| Robocrunch AI

Flamingo, a drastic VLM that achieves SOTA on many vision + language tasks! Fun fact: (real) flamingos change colors depending on what they eat. Our Flamingo model behaves differently depending on the inputs it's fed 🍕🦩https://t.co/1FuCbSGe2r pic.twitter.com/WuMYL8KOzX
— Oriol Vinyals (@OriolVinyalsML) April 28, 2022

Introducing Flamingo 🦩: a generalist visual language model that can rapidly adapt its behaviour given just a handful of examples. Out of the box, it's also capable of rich visual dialog. Read more: https://t.co/xEzqTizoJQ 1/

Shared by DeepMind at 4/28/2022

In case you missed it...Flamingo 🦩 a new SOTA visual language model. Read more below ⬇️ Paper: https://t.co/uv8LKkSal7 Blog: https://t.co/BPHWfGI1yi

Shared by DeepMind at 5/9/2022

Flamingo seamlessly handles input sequences of images, videos, & text by bridging powerful pretrained LMs with visual encoders. With only simple few-shot learning (no fine-tuning) a single 🦩 achieves SotA results on several competitive computer vision benchmarks in 32 shots. 2/

Shared by DeepMind at 4/28/2022

Deepmind: Flamingo combines language and vision https://t.co/QXREk7TcZd

Shared by KDnuggets at 5/9/2022

Flamingo (🦩) is amazing. But it's really important to be honest about the limitations of the systems we build, and not to fuel hype 2/9

Shared by Murray Shanahan at 5/10/2022

Took some examples from the "ImageNet-trained CNNs are biased towards texture" paper. Nice to see @DeepMind Flamingo 🦩🦩 showing the desired shape bias! 😍 (without explicitly being trained to do so!) Glad to be a part of this! We've come a long way with deep learning 🧵👇 1/

Shared by Aleksa Gordić at 5/10/2022

768