Vines' Log

A unit test exercises a unit of functionality to test its behaviour. We cover the framework unittest and coverage.py

Some keypoints and details jot from CVPR 2022 tutorial - Tutorial on Denoising Diffusion-based Generative Modeling - Foundations and Applications

Some keypoints and details jot from CVPR 2022 tutorial - Tutorial on Denoising Diffusion-based Generative Modeling - Foundations and Applications

Topics covered Introduction to Team Collaboration, Requirements Engineering, Requirements Engineering Process

Topics covered Foundations of Software Engineering, Process Models, Agile Methods

We quickly revisit how Local segmentation, FCN (Fully Convolutional Network), and UNet can achieve image segmentation, and talk about their strength and drawbacks.

Some tips that I use when my ubuntu stuck in boot screen. It also works when we edited some file and messed up and cant boot into ubuntu by going to recovery mode.

You need to first understand DDPM before understanding RePaint. Using a pretrained unconditional DDPM as the generative prior, we can perform inpainting task.

You need to first understand VAE before understanding DDPM. The key idea of Diffusion is that it learn a reverse diffusion process that restores structure in data, yielding a highly flexible and tractable generative model of the data.

The idea of Variational AutoEncoder is to map a distribution into the latent space and then randomly sample it, which is stochastic.