Prepare your dataset and train a custom Flux LoRA face model on Replicate
Press play on the video. It'll jump straight to the section that answers the
title above — no need to watch the full video.
Replicate
Flux
Image Generation
Customization
A complete guide on how to select and name image files for your dataset, upload them to Replicate, and train a custom AI model using the Flux LoRA trainer.
The 'Trash In, Trash Out' Principle
The quality of your AI model depends entirely on the quality of your dataset. If you're aiming for 'high fashion' results, don't use basic selfies. Provide images that match the quality of the output you envision.
The Importance of Filenames & Trigger Words
Naming files with the subject's name helps the model identify you. Use that same name as the 'Trigger Word' during training so the AI knows to generate your likeness when that name is called in a prompt.
Settings Tips (Hyperparameters)
Based on testing, the 'Default' settings on Replicate are already well-optimized for most general use cases. There's no need to tweak complex technical settings if you're just starting out.
More from Generate & Edit Professional AI Images
View All
None
Midjourney
Creating consistent vector graphics and icon sets with Recraft
Recraft
Generate and Edit Realistic Images and Text with Recraft AI
Recraft
Downloading and Setting Up Flux Models and Required Encoders
ComfyUI
Flux
Generating 4K Images with DyPE Workflow Settings in ComfyUI
ComfyUI
DyPE
Adding LoRA to the DyPE Workflow for Specific Styles
ComfyUI
Flux