decompressiemodels
Decompressiemodels, often referred to as diffusion models, are a class of generative models in machine learning that learn to create data, such as images, by reversing a gradual corruption process. The core idea is to start with random noise and iteratively denoise it until a coherent piece of data is formed. This denoising process is guided by a neural network that has been trained to predict the noise that was added at each step.
The training of a decompressiemodel involves two main phases. First, a forward diffusion process is applied
During inference, the model starts with a sample of pure noise and iteratively applies the learned denoising