Sign In

Miltank

89

648

2.4k

34

Verified:

SafeTensor

Type

LoRA

Stats

648

2.4k

22

Reviews

Published

Oct 24, 2023

Base Model

SD 1.5

Training

Steps: 2,000
Epochs: 8

Usage Tips

Clip Skip: 1

Trigger Words

Miltank

Hash

AutoV2
ABDA80D328

Miltank LoRA trained on NovelAI with Kohya training scripts using the DAdaptAdam optimizer at 512x512 resolution. Can have issues getting the skin patterns right, may require adding (pink skin) or an equivalent to your prompt. I also had issues getting consistent bovine traits with YiffyMix, so I had to add (bovine) to my prompt to fix that, alongside (cloven hooves) whenever her feet were in the picture. Works quite well once you add those reinforcement tags. Should have wide compatibility with model mixes based on NovelAI, which is most of them. Trained on an all female dataset.