Update README.md
Browse files
README.md
CHANGED
@@ -7,6 +7,15 @@ license: openrail
|
|
7 |
The embedding were trained using A1111 TI for the 768px Stable Diffusion v2.0 model.
|
8 |
The embedding should work on any model that uses SD v2.0 as a base.
|
9 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
<div align="center">
|
11 |
<img src="https://huggingface.co/fzbuzz/TungstenDispo-embedding-sd-v2-1/resolve/main/00000.png">
|
12 |
</div>
|
@@ -23,18 +32,9 @@ The embedding should work on any model that uses SD v2.0 as a base.
|
|
23 |
<img src="https://huggingface.co/fzbuzz/TungstenDispo-embedding-sd-v2-1/resolve/main/00003.png">
|
24 |
</div>
|
25 |
|
26 |
-
|
27 |
-
Download the TungstenDispo.pt file and put in embeddings/. Prepend "TungstenDispo" at start of prompt.
|
28 |
-
|
29 |
-
**TungstenDispo (v1)**
|
30 |
-
|
31 |
-
The TungstenDispo embedding were trained for 1000 epochs with a gradient batch size of 50.
|
32 |
-
A total of ~100 training images of tungsten photographs taken with CineStill 800T were used. The split was around 50/50 people landscapes.
|
33 |
-
The effect isn't quite the tungsten photo effect I was going for, but creates very nice, artistic portraits of people.
|
34 |
-
Landscapes haven't been experimented with much. For some of the people, I used SoCalGuitarist's [Negative FaceLift](https://civitai.com/models/2385/socalguitarists-magic-facelift-negative-embedding-for-model-2x-fix-yo-ugly-faces) as a negative embedding.
|
35 |
-
I used it on 0.3 strength, and it seems like it makes the eyes slightly less wonky. Unclear extent of effect.
|
36 |
-
|
37 |
**Workflow for Above Pictures**
|
|
|
38 |
Sampler: Euler-A, 20 Steps, CFG: 7.0. Slightly cherry-picked for best pictures.
|
39 |
900x768 -> 4x LDSR upscaled
|
40 |
|
@@ -43,7 +43,13 @@ Negative Prompt for all images (Not entirely sure if all of them matter, but doe
|
|
43 |
|
44 |
TungstenDispo, photoshoot of a old man, highly detailed
|
45 |
|
|
|
|
|
|
|
|
|
|
|
46 |
**More People Samples w/out exact workflow**
|
|
|
47 |
Pretty much the same, only changed up subject a little + weights.
|
48 |
|
49 |
<div align="center">
|
|
|
7 |
The embedding were trained using A1111 TI for the 768px Stable Diffusion v2.0 model.
|
8 |
The embedding should work on any model that uses SD v2.0 as a base.
|
9 |
|
10 |
+
**TungstenDispo (v1)**
|
11 |
+
|
12 |
+
The TungstenDispo embedding were trained for 1000 epochs with a gradient batch size of 50.
|
13 |
+
A total of ~100 training images of tungsten photographs taken with CineStill 800T were used. The split was around 50/50 people landscapes.
|
14 |
+
The effect isn't quite the tungsten photo effect I was going for, but creates very nice, artistic portraits of people.
|
15 |
+
Landscapes haven't been experimented with much. For some of the people, I used SoCalGuitarist's [Negative FaceLift](https://civitai.com/models/2385/socalguitarists-magic-facelift-negative-embedding-for-model-2x-fix-yo-ugly-faces) as a negative embedding.
|
16 |
+
I used it on 0.3 strength, and it seems like it makes the eyes slightly less wonky. Unclear extent of effect.
|
17 |
+
|
18 |
+
|
19 |
<div align="center">
|
20 |
<img src="https://huggingface.co/fzbuzz/TungstenDispo-embedding-sd-v2-1/resolve/main/00000.png">
|
21 |
</div>
|
|
|
32 |
<img src="https://huggingface.co/fzbuzz/TungstenDispo-embedding-sd-v2-1/resolve/main/00003.png">
|
33 |
</div>
|
34 |
|
35 |
+
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
36 |
**Workflow for Above Pictures**
|
37 |
+
|
38 |
Sampler: Euler-A, 20 Steps, CFG: 7.0. Slightly cherry-picked for best pictures.
|
39 |
900x768 -> 4x LDSR upscaled
|
40 |
|
|
|
43 |
|
44 |
TungstenDispo, photoshoot of a old man, highly detailed
|
45 |
|
46 |
+
**Usage for A1111 WebUI**
|
47 |
+
|
48 |
+
Download the TungstenDispo.pt file and put in embeddings/. Prepend "TungstenDispo" at start of prompt.
|
49 |
+
|
50 |
+
|
51 |
**More People Samples w/out exact workflow**
|
52 |
+
|
53 |
Pretty much the same, only changed up subject a little + weights.
|
54 |
|
55 |
<div align="center">
|