Kohya Trainer
LoRA XL Trainer colab from Linarquf with some updates.
Updates
2024
- removed WD Tagger V2 (use V3 instead)
- added NoobAI-XL-EarlyAccess Diffusers Models
- added Pony Diffusion v6 XL and Illustrious XL Models
- support WD14 v3 Tagger Large Models
- added WD14 v3 Tagger
- added
Prodigy
optimizer.
- added
text_encoder_lr
cache_latents
cache_latents_to_disk
.
removed Scraping stuff.
- added
optimizer_args
recomended values.