AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
The Differences Between Gaussian Blur and Treeness
When the model is being trained, it's essentially applying Gaussian blur to the image. So at first looks at the image, looks at the words that's associated with the image and then blurs it a little bit. And then it does it again and do it again until it's pure noise. But after looking at millions or billions of images, I guess in this case, that it has learned in a way like what, like, treeness is, let's say. Like tree. Yes, that isn't like the essence of tree. Exactly.