Completed
Required resources
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Stable Diffusion - How to Train LoRA for Style - LoRA Part 1
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 Why is LoRA a new standard?
- 3 Required resources
- 4 Overview of video parts 1. local GUI and 2. Google Colab Notebook
- 5 Part 1. local GUI training
- 6 PC hardware specs
- 7 Dataset preparation
- 8 Batch renaming filetypes to a new filetype
- 9 Batch renaming attempt #2
- 10 BIRME cropping
- 11 BLIP captioning
- 12 How to write good captions
- 13 Captioning example #2
- 14 Creating project folder structure using kohya_ss GUI
- 15 Training parameters
- 16 Why the training steps in the console are different from max train steps
- 17 Console output after training commences
- 18 What you get after training finishes
- 19 Testing the style LoRA in AUTOMATIC1111 webui
- 20 Using x/y/z plot in scripts to vary parameters
- 21 Checking x/y/z plot in photoshop
- 22 Testing the final epoch LoRA
- 23 Verifying LoRA's flexibility with prompting
- 24 Part 2. Anime style LoRA
- 25 Dataset preparation on redjuice/shiru artwork
- 26 Using AUTOMATIC1111 webui extension tagger
- 27 Modifying Danbooru style tags
- 28 Using the kohya LoRA DreamBooth colab notebook
- 29 Fixing the error that occurred
- 30 Preliminary testing of redjuice style LoRA
- 31 Checking x/y/z plot in photoshop
- 32 Combining character and style LoRAs
