I've now made this much progress with Lora Train. It's time for a fresh start. I've already trained on diffusion pipe, not like Kohya or Web Training. Otherwise, go ahead!
readme:
Before you start, this is not the flux.dev, this is a chroma model and with the chroma model you need to prepare what you can create an image with.
Chroma model: ✅
Flux_dev model: ❌
Realistic: ❓
Simulator & 3D: ❓
anime: ✅
Cartoons: ✅
preparation for comfyui & models:
Comfyui:
https://huggingface.co/lodestones/Chroma/blob/main/simple_workflow.json
https://github.com/lodestone-rock/ComfyUI_FluxMod
Models:
Unet: https://huggingface.co/lodestones/Chroma/blob/main/chroma-unlocked-v10.safetensors
https://huggingface.co/lodestones/Chroma/blob/main/chroma-v2.5.safetensors
Or
GGUF:
https://huggingface.co/silveroxides/Chroma-GGUF
text_encoders:
https://huggingface.co/comfyanonymous/flux_text_encoders/tree/main
VAE:
https://huggingface.co/black-forest-labs/FLUX.1-schnell/blob/main/ae.safetensors
that's it, Have fun and good luck!