Spider-Woman (Julia Carpenter) [LoRA]

v1.0 SD v1.5 AnimeFull
prgfrg23
11 months ago

This LoRA is trained on images from comics and some fan art and synthetic images created by me using IMG2IMG

v1.0 - Images without a mask were additionally marked with the swjcsuit token, This token is optional and does not change the image much, the swjcsuit token creates a black outline and an effect more like a comic book, and does not greatly affect the creation of images with a mask. Hair color (red, orange, blonde, yellow) The color of the bodysuit can be black or blue, the color can be any, but these colors do not change the costume much.

If there are no eyes on the mask, add white eyes, no pupils, to prompt

Preview generated without using ADetailer & Hires. fix (Except for the SD v1.5 version, Kohya Hires.fix)

Recommended LoRA Strength: 0.8-0.9

γ…€

LoRAs which I used to create synthetic images

https://civitai.com/models/522562/jessica-drew-spider-clone-ultimate-universe

https://civitai.com/models/512197/venom-eddie-brock-spider-man-the-animated-series-lora

Read more...
Download (70.4 MB) Download available on desktop only

Popularity

470 ~10

Info

Latest version (v1.0 SD v1.5 AnimeFull): 1 File

To download these files, please visit this page from a desktop computer.

About this version: v1.0 SD v1.5 AnimeFull

Repeats: 5

Epoch: 10

Image Count: 60

Network Rank (Dimension): 64

Network Alpha: 32

Batch size: 1

Clip skip: 1

Model: animefull-final-pruned-fp16

Training steps: 3000

Resolution: 640x960

Train Resolution: 640,960

LR Scheduler: cosine_with_restarts

Optimizer: AdamW8bit

Learning rate: 0,0005

Text Encoder learning rate: 0,0001

Unet learning rate: 0,0002

Captioning: WD14 Captioning (SmilingWolf/wd-v1-4-moat-tagger-v2) + Edit

Trigger Words: juliacarpenter,

Recommended LoRA Strength (Weight): 0.8-0.9

3 Versions

πŸ˜₯ There are no Spider-Woman (Julia Carpenter) [LoRA] v1.0 SD v1.5 AnimeFull prompts yet!

Go ahead and upload yours!

No results

Your query returned no results – please try removing some filters or trying a different term.