nitrosocke
commited on
Commit
•
4027643
1
Parent(s):
5b2dbff
updated the readme file with training settings
Browse files
README.md
CHANGED
@@ -1,3 +1,14 @@
|
|
1 |
---
|
2 |
license: creativeml-openrail-m
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: creativeml-openrail-m
|
3 |
---
|
4 |
+
# Arcane Diffusion Dataset
|
5 |
+
Dataset containing the 75 images used to train the [Arcane Diffusion](https://huggingface.co/nitrosocke/Arcane-Diffusion) model.
|
6 |
+
|
7 |
+
Settings for training:
|
8 |
+
```class prompt: illustration style
|
9 |
+
instance prompt: illustration arcane style
|
10 |
+
learning rate: 5e-6
|
11 |
+
lr scheduler: constant
|
12 |
+
num class images: 1000
|
13 |
+
max train steps: 5000
|
14 |
+
```
|