Chaos
When I was in my 20s, my long-term partner / husband (who was 10 years older) at the time gave me a card for my birthday. It was very simple. It’s just a quote from Friedrich Nietzsche:
You must have chaos within you to give birth to a dancing star.
I didn’t think a lot about it at the time. In truth, when I first saw it, my first reaction was “WTF” but this quote would continue to resonate with me for the rest of my life. Now, after nearly two decades (Sept 25 is my birthday), I can say that I have a much better understanding of what it means, and I appreciate how thoughtful he was to have sent this to me on my birthday.
If you know me IRL, you’d know that while I meticulously organize my files and things in the virtual space, I’m rather messy. I’m not even sure why but somehow the boyfriends who ended up with me in a relationship tends to be neat freaks. When I’m not in a relationship, I usually would need to hire a cleaning lady to take care of my place, but when I am in one, usually my boyfriends would just keep everything in order (talk about benefits, LOL).
It doesn’t mean that I don’t like an orderly home — it’s more that I’m simply too busy to want to care about it. I’m always working on something, and I don’t want to spend my time cleaning. I’d rather spend my time creating.
In a way, this quote being sent to me on my birthday was perhaps his way of saying that despite my mess, he appreciates everything that I do. And it shows the extent of our love together. I miss that about him. We lived together for a couple of years. We ended things amicably, but the circumstances were a bit complicated. I still think about him from time to time.
Tech
I made these renders using a custom SDXL Lora that I trained on Replicate, using 30 paintings from my Construction Series as the training dataset. I don’t usually share my prompts, but since this one is special, I’ll share a link to my model as well as the full prompt, so you that you play with them if you like.
I’ll write a longer explanation of this model in a later post. The actual training process is quite simple with Replicate, but it is quite involved to explain that process — so I’ll try to collect my thoughts in a way that’s understandable to people who are not technically inclined nor understand the AI technology with drives Stable Diffusion.
Text prompts in Stable Diffusion, SDXL
- SDXL, KarrasDPM, 1024x1024
- Inference steps: 50
- Refiner steps: 50
- Custom SDXL LoRA: gdmjp4
- SDXL Base Model
- SDXL Base Refiner Model
- High Noise Frac: 0.8
- Post Upscale: Topaz Gigapixel AI HQ 4x, 4096x4096