Train image models to keep your character consistent.
What it does
Guides users through preparing character image datasets, configuring LoRA training, and selecting a checkpoint for consistent character generation.
When to use it
Use when a character needs consistent identity across many images or an ongoing series.
How to use it
Give it character reference images; it prepares a captioned dataset, configures training, evaluates checkpoints, and integrates the selected LoRA.
What you provide
- Your files
Uses
What you need · 3
Requires the python binary; no version is specified.
Standard FLUX training calls for 24GB VRAM; the documented SimpleTuner NF4 configuration reduces that to 9GB.
Requires an existing ComfyUI installation for the final checkpoint integration and workflow test.
Access · 1
This skill
{ComfyUI}/models/loras/
Copies a checkpoint into ComfyUI