Found a message this morning where someone shared a link to a 'dead easy LoRA' training for Flux dev with 12/16/20GB VRAM called Fluxgym. It is a 3 step UI for Kohya scripts to train Flux.1-dev and it even has a powerful but simple captioning module. It is a simplified GUI which generates the scripts to run Kohya.
That's it, lol. So all you need are 10-40 images as a start and you are good to go.
I recommend downloading pinokio, it will handle all the hard work and provides fluxgym as a oneclickinstaller. That's the version i am using and it works.
For manual installing and additional information, visit https://github.com/cocktailpeanut/fluxgym.
Actual disscusion seems to take place here: https://www.reddit.com/r/StableDiffusion/comments/1faj88q/fluxgym_dead_simple_flux_lora_training_web_ui_for/
And finally for some parameter insights, there is an excellent article here on Civitai: https://civitai.com/articles/7097/flux-complete-lora-settings-and-dataset-guide-post-mortem-of-two-weeks-of-learning
I tried it and an hour later i ended up with:
Funny thing, isn't it? Now enjoy the new toy...
3rd attempt with 40 images turned out very good. I used 40 images, 10 repeats and 12 epochs. The images were captionen with 'xslx scribble' and the provided Florence-2 captioning. Learning rate was set to 2e-4 and resolution 512x512. You can download the result here: Flux Scribble Art.
With the 20GB VRAM option checked, usage peaks around 65% with my 24GB card, this means real usage is around 16GB. (the author mentioned that somewhere in the links). 
Add on:
Caption or not to caption - Flux LoRA examples
A good read, highly recommended
(I am not the owner or creator of any of these programs and scripts, i am just 'dead happy' that there is now a 'dead simple' method to finally make some nice LoRAs and wanted to share this 3 article combination. That was all i need to make my first Flux LoRA in less than 2 hours)