I saw a post on Reddit, entitled “Llama 3 rocks with taking on a personality!”. A fun experiment! I thought to replicate it, blatantly copying the puzzle presented. Llama 3, released by Meta just before the weekend, is impressive.

I keep posting about Stable Diffusion, but I do experiment with Large Language Models too! I do not have much to contribute in this regard, instead, here is the transcript of a game I played with the open source Mistral 7B model via Ollama.

Happy new year!

Ever needed to extract images from PDFs and found both on-line and off-line tools lacking? Well, I certainly have, and here I present my Python code to extract JPGs/PNGs from PDFs, using PyMuPDF.

In the span of a couple of weeks, we got Crazy fast image generation with LCM LoRA for SDXL, which led me to ask if I could get Faster Stable Diffusion on M-series macs?. A few hours ago, Stability.ai gave us their response in the form of SDXL-Turbo... and now we go even faster!

All my recent Stable Diffusion XL experiments have been on my Windows PC instead of my M2 mac, because it has a faster Nvidia 2060 GPU with more memory. But today, I’m curious to see how much faster diffusion has gotten on a M-series mac (M2 specifically).

In this post, I experiment with latent scaling and latent compositing with SDXL 1.0 using ComfyUI. That is to say, increasing / decreasing the size of the image, and combining multiple images into one à la green screen (chroma key) compositing.

Onward with SDXL and ComfyUI! Sometimes I want to tweak generated images by replacing selected parts that don’t look good while retaining the rest of the image that does look good. Rather than manually creating a mask, I’d like to leverage CLIPSeg to generate a masks from a text prompt.

Yesterday I mentioned in passing that my Nvidia RTX 2060 with 12GB could not run both SDXL 1.0 Base and Refiner models in a single ComfyUI workflow. Today, I show you my workaround and also experiment with adding the SDXL 1.0 Official Offset Example LoRA to the workflow.

Stability.ai has released Stable Diffusion XL (SDXL) 1.0 (26 July 2023)! Time to test it out using a no-code GUI called ComfyUI!

A short post about setting up a mouse on macOS to mimic trackpad gestures.

In my previous post, I managed to get Whisky, with Apple’s Game Porting Toolkit, working on macOS Ventura without Homebrew. Here I manage to get Steam running after jumping through hoops.

A long time ago, I compiled Wine from scratch on 32-bit macOS Sierra. I lost interest because I never used it. Now after WWDC 2023, everyone is talking about the happy side effect of Apple’s Game Porting Toolkit (GPTK) which uses Wine under the hood to run Windows 10 DirectX 12 games on macOS!

I sometimes test randomly downloaded code on macOS in a Sandbox that has limited and network access. I posted about this way back in Jan 2017, Creating a macOS Sandbox to run Kodi, and this is a short refresher for me...

The current macOS “System Settings” continues to confound me. Did you know you can prevent your mac from sleeping with a setting? I previously thought caffeinate was the only way to achieve this.

When I export photos or videos from Photos.app, I want the file’s Created Date (Created Date in Finder) to be the time the photo or video was taken. Alas, this is not the way Photos works, and setting the date turned out to be more challenging than expected.

Want to check your connected bluetooth device battery level on a mac? Here is one way: using xbar (I am aware there are a bunch of other xbar plugins to do something similar, but I’ve been using my own script before I ever used xbar).

Did you know that macOS has a built-in game performance HUD with a frames per second (FPS) counter, similar using RivaTuner Statistics Server (RTSS) to display a GPU performance overlay in Windows? Here’s how to enable the Metal Performance HUD.

Some time ago, I explained how I setup “Global hotkeys for macOS with skhd”. I mentioned running it automatically upon login via the macOS Launch Daemon but did not detail the steps. So here they are...

I have been using UTM on M1/M2 macs for all my virtualization needs for a while, but never posted much because it practically works out of the box. Since around October 2022, UTM v4.0.9 on macOS Ventura allows Apple Hypervisor Virtualization Virtual Machines to use Rosetta 2.

I was recently researching external storage for my Mac, looking for the “fastest” and most “reliable”. There are so many options, and of course, so many different price points! It’s all very confusing: NGFF M.2? NVMe M.2? USB 3.1 Gen 1 (5 Mbps)? USB 3.1 Gen 2 (10 Gbps)? USB 3.1 Gen 2x2 (20 Gbps)? Thunderbolt 3? Thunderbolt 4? USB 4.0?