LoRA is trained on 74 screenshots from Spider-Man: Across the Spider-Verse (2023)
Preview generated without using ADetailer & Hires. fix
Recommended LoRA Strength (Weight): 0.8-0.9
v1.0 F (HQ) High Quality - Almost the same as v1.0 F, but compression artifacts have been removed and the file size is half as small (There are no more changes)
v1.0 F (Fixed) - This version has a different Trigger Word, spidey2099 is changed to sm2099moh and covered eyes token is added, this token is needed for the mask, the costume and mask look much better (This version used to be called v1.1, but I decided to rename it)
v1.0 - This is the version I don't recommend using.
Description
Repeats: 10
Epoch: 5
Image Count: 74
Network Rank (Dimension): 16
Network Alpha: 8
Batch size: 1
Clip skip: 1
Training steps: 3700
Resolution: 1024x768
Train Resolution: 1024x1024
LR Scheduler: cosine_with_restarts
Optimizer: AdamW8bit
Learning rate: 0,0005
Text Encoder learning rate: 0,00005
Unet learning rate: 0,0005
Captioning: WD14 Captioning (SmilingWolf/wd-v1-4-moat-tagger-v2) + Edit
Trigger Words: miguelohara, spidey2099suit, sm2099moh,
Recommended LoRA Strength (Weight): 0.7-0.8