AGI is hard to imagine and it's scary, it's a superstition, nothing more.

#SDXL #SD #LoRA

AI in 2024 is Bitcoin in 2013, whatever you're doing keep doing it and you'll have a chance to stand with the giants in the future, ups and downs are just noise, bitcoin crashing from 1$ to 75 cents was just noise, same with AI, don't listen to the noise.

It works on character models too. Workflow: 1. #Train a model (15 min + 1 hr processing time) 2. #Sketch (5 min in the Live Canvas) 3. #Enhance to polish the render (1 min) You can get ultra-consistent characters, exactly in the pose you want, in minutes, for less than a dollar.

Building an optimized flux LoRA trainer, quality is a priority

Davy Jones, No LoRA vs 350 steps LoRA, Flux. Extremely low quality 7 pics dataset resized (not upscaled) up to 1024 for more stable training.

flux papercut, will publish various LoRAs once the format is supported by ComfyUI

Flux LoRA, 700 steps, 10 images

Existing tokens greatly help reduce the training time, the joker and the hound were the training tokens for these flux LoRAs, between 500-700 steps, and about 10-20 minutes training time on A100-80G

King of the north, Flux LoRA

Game Of Thrones Characters eating Spaghetti, Flux LoRA

For powerful models like Flux, training LoRAs shouldn't be straightforward, taking advantage of the power of the model gives a lot of room for optimization and fine-grained LoRA training, that includes targeting specific blocks at varied strengths, it saves time, memory & quality