In a previous post I briefly covered Diffusion Models. And as I pointed out, they have an important drawback that makes them slow at creating new images. We often have to wait minutes for an image ...
Some thoughts on the usage of AI for developing software: Is the end near?
– Beware of the following to be quite philosophycal, if you're pragmatic, end the reading here! – In the last post I wrote about how I used AI as an assistant to build a guitar chord website. ...
The AI Coder: Building a website using AI
Last week I decided to tackle a quite silly side-project I left stopped about 2 years ago: building my own (guitar) chord / tab site so I can store all my music sheets. The main motivation for doi...
How I generated professional headshots for my LinkedIn profile
The other day I was bored, wandering through LinkedIn and realized that I probably needed some new profile pics. You know, the ones very corporate-guys share on their posts. Something cool with a s...
LoRA: How can I tune my model without an expensive GPU?
Training neural networks can be really tedious. Both in terms of methodological complexity – tuning hyperparameters, syncing training runs, scheduling learning rates, etc. – and also hardware requi...
Intro to Diffusion Models
On Diffusion Models In this post I’m going to cover an introduction to Diffusion Models (as I don’t consider myself an expert on them) by summarizing Ho et al (2020). “Denoising Diffusion Probabi...
The magic behind (forward mode) Automatic Differentiation
Demystifying Automatic Differentiation If you recall from highschool you’ll probably remember the definition of a derivative: [\begin{equation} f’(x) = \lim_{\epsilon \rightarrow 0} \frac{f(...
Recently Updated