In this video, I'll be diving into the world of Flux LoRA training and showing you how I've been training my own custom LoRAs.
We'll cover:
- The basics of Flux LoRA training: How to set up a cloud GPU instance on RunPod and get started with AI Toolkit.
- Experimental LoRa techniques: Learn how to train LoRAs with a single image and target specific layers for faster, more efficient results.
- Budget-friendly training: Discover how you can train LoRAs on the cloud without breaking the bank.
- Step-by-step guide: Follow along as I train LoRAs using a real-world example.
Want to support me? You can buy me a ko-fi here:
RunPod (affiliate link): https://runpod.io?ref=u0pqdiej
AI Toolkit: https://github.com/ostris/ai-toolkit
AI Toolkit for RunPod Walkthrough: https://github.com/ostris/ai-toolkit/issues/90
Config Files:
- basic: https://pastebin.com/cKVtywkR
- single image: https://pastebin.com/ZBbLM30z
- single layer targets: https://pastebin.com/GA07PySr
Timestamps:
00:00 - Intro
00:53 - The Basics of LoRA Training: Why RunPod, Local, AI-Toolkit
02:02 - Setting Up RunPod
05:11 - Setting Up AI-Toolkit on RunPod
07:38 - Setting Up AI-Toolkit Locally
08:21 - Troubleshooting Requirements Installation on startup
09:50 - Setting Up the Huggingface Token
11:28 - Setting Up the Dataset
14:00 - Setting Up the Config File
19:15 - Starting Training
20:19 - First Training (Basic LoRA Flux Training)
22:35 - Second Training (One Image Training)
25:20 - Third Training (Target Single Layers Training)
29:30 - Outro
#flux #lora #loratraining #stablediffusion #runpod #aitoolkit #ai #generativeai