Man, it was mad.
I just trained a LoRa model on my own photos to see what kind of results it could produce. I wanted to explore what’s possible—and what might actually help the solopreneurs I serve.
Because here’s the thing: content creation is heavy.
You need thumbnails. Blog banners. A decent LinkedIn photo. And while professional photographers are worth every dollar, not every solopreneur has hundreds to spare at the start.
So I wanted to test a simple question:
Can AI help solopreneurs generate professional-looking assets from just a handful of self-taken photos?
What I Did (It Was Easier Than I Thought)
Step 1: Took About 100 Photos of Myself
I used my phone. Nothing fancy. I smiled in every photo—some close-ups, some half-body, some full-body, all from different angles and distances. The idea was to train the model on a specific expression first (in this case, smiling), so I could test output quality before expanding.
Step 2: Uploaded and Trained the LoRa
The training process was super simple. Zip up the images, upload to an online tool, hit go. That was it. No code. Not too many setups. Just upload, setup, and wait.
Step 3: Generation Time
This is where things got interesting.
I entered some prompts and started generating. And then—
Boom. Magic.
Studio-like portraits. Natural lighting. Clear smiles. Teeth that didn’t look weird.
Photos that I’d actually use if I were a solopreneur trying to brand myself online.
I even generated a polished headshot that looked LinkedIn-ready. The kind of thing you’d usually need a good photographer and a clean backdrop to get. And here it was—created in minutes.
What Surprised Me
😲 How Good the Quality Was
Yes, some outputs were off. But a surprising number were ready to use as-is.
With just 100 training photos? That’s impressive.
If you’re trying to generate normal lifestyle or branding photos with a consistent aesthetic, this is a real option. Want to ride a unicorn while eating ice cream? Maybe not there yet. But for day-to-day brand assets? Definitely usable.
😬 The Privacy Instinct Kicked In
As the model trained and started working, I felt oddly protective.
This is my face, after all. I made sure to keep the model private and double-checked settings.
For solopreneurs, this means thinking about data privacy when using online tools—especially ones that retain your face.
⚡️ The Potential Is Wild
As I generated more photos, my mind went to new places:
- What if this could replace stock photos on vision boards—with images of yourself doing the thing you’re working toward?
- What if you could show up more consistently online because photo generation became easier than choosing an outfit?
- What if your visual identity could grow with your brand—without rebooking a new photoshoot every 6 months?
What’s Next?
If I’m serious about helping solopreneurs with this, I’ll need to dig deeper:
- Better prompting skills (the model’s good—but only as good as your instructions).
- Testing img2img with LoRa, where you use existing photos for pose reference and swap in your own look.
- Creating a pipeline to automate thumbnails, book covers, Instagram covers, etc.—all generated on-brand, with your face, and in your style.
This isn’t about replacing photographers.
It’s about expanding what’s possible for people who are just starting out—those trying to look professional without spending thousands.
If you’re a solopreneur, would you try something like this?
And if you’re already using AI to build your brand, I’d love to hear how.
Leave a Reply