AlekseyCalvin commited on
Commit
dca590e
·
verified ·
1 Parent(s): 511ddbe

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +34 -5
README.md CHANGED
@@ -1,5 +1,7 @@
1
  ---
2
  license: other
 
 
3
  base_model:
4
  - black-forest-labs/FLUX.1-dev
5
  library_name: diffusers
@@ -36,14 +38,37 @@ tags:
36
 
37
  widget:
38
  - text: >-
39
- -
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
40
  output:
41
- url: MangledMergeV1.jpeg
42
  - text: Seed:595570113703122 2steps
43
  output:
44
  url: Pixel9.png
45
- language:
46
- - en
47
  ---
48
 
49
  ## MANGLED MERGE V1.0
@@ -54,7 +79,11 @@ Herein lives a **HF/Diffusers-ported** version of the [Mangled Merge Flux - v1.0
54
 
55
  **Mangled Merge V1** is a merge of [Mangled Merge Flux Matrix](https://civitai.com/models/893430/mangled-merge-flux-matrix), [Mangled Merge Flux Magic](https://civitai.com/models/891563/mangled-merge-flux-magic), [PixelWave 03](https://civitai.com/models/141592?modelVersionId=992642) (or [HF/Diffusers port here](https://huggingface.co/AlekseyCalvin/PixelWave_Schnell_03_by_humblemikey_Diffusers_fp8_T4bf16)), [FluxBooru v0.3](https://civitai.com/models/859032?modelVersionId=961597), and [Flux-dev-de-distill](https://huggingface.co/nyanko7/flux-dev-de-distill), to create a model that offers *the aesthetics of 808 merged LoRAs*, with *the styles of **PixelWave***, *the knowledge of the **Booru** dataset* and *the functionality of a dedistilled model*! <br>
56
 
57
- LoRAs work fine with this model, CFG is best between 1 and 3.5, Flux Guidance doesn't work, but negative prompts work great, and I HIGHLY recommend using dynamic thresholding. <br>
 
 
 
 
58
 
59
  For a model comparison between **Mangled Merge V1** vs. **Flux.1 Dev**, check out this [post](https://civitai.com/posts/8666955). <br>
60
 
 
1
  ---
2
  license: other
3
+ language:
4
+ - en
5
  base_model:
6
  - black-forest-labs/FLUX.1-dev
7
  library_name: diffusers
 
38
 
39
  widget:
40
  - text: >-
41
+ Seed:595570113703157 2steps
42
+ output:
43
+ url: Pixel2.png
44
+ - text: >-
45
+ Seed:595570113703486 4steps
46
+ output:
47
+ url: Pixel6.png
48
+ - text: >-
49
+ Seed:595570113703445 2steps Vibrant Graffiti Art, digital illustration, text /PEACE/, graffiti, colorful, flowers, purple, green, blue hues, splatter effect, shadowing, 3D effect, gradient, urban art, street style, bold outlines, bright background
50
+ output:
51
+ url: Pixel5.png
52
+ - text: >-
53
+ Seed:595570113703140 2steps
54
+ output:
55
+ url: Pixel1.png
56
+ - text: >-
57
+ Seed:595570113703255 2steps Vivid digital portrait, contemporary surrealism. Young woman in intricate yellow and blue lace-patterned attire. Background bright yellow with subtle dots, left side crowned by elaborate green lace-like halo. Right side dynamic turquoise and aqua brushstrokes, resembling ethereal fire. Focus on intricate textures and striking color contrasts
58
+ output:
59
+ url: Pixel8.png
60
+ - text: >-
61
+ Seed:595570113703505 2steps
62
+ output:
63
+ url: Pixel7.png
64
+ - text: >-
65
+ Seed:595570113703362 3steps a robot priest in torn robes, dark industrial, melancholic
66
  output:
67
+ url: Pixel3.png
68
  - text: Seed:595570113703122 2steps
69
  output:
70
  url: Pixel9.png
71
+
 
72
  ---
73
 
74
  ## MANGLED MERGE V1.0
 
79
 
80
  **Mangled Merge V1** is a merge of [Mangled Merge Flux Matrix](https://civitai.com/models/893430/mangled-merge-flux-matrix), [Mangled Merge Flux Magic](https://civitai.com/models/891563/mangled-merge-flux-magic), [PixelWave 03](https://civitai.com/models/141592?modelVersionId=992642) (or [HF/Diffusers port here](https://huggingface.co/AlekseyCalvin/PixelWave_Schnell_03_by_humblemikey_Diffusers_fp8_T4bf16)), [FluxBooru v0.3](https://civitai.com/models/859032?modelVersionId=961597), and [Flux-dev-de-distill](https://huggingface.co/nyanko7/flux-dev-de-distill), to create a model that offers *the aesthetics of 808 merged LoRAs*, with *the styles of **PixelWave***, *the knowledge of the **Booru** dataset* and *the functionality of a dedistilled model*! <br>
81
 
82
+ <Gallery />
83
+
84
+ **LoRAs** work fine with this model,
85
+ **CFG** is best between 1 and 3.5,
86
+ **Flux Guidance** doesn't work, but **negative prompts** work great, and I HIGHLY recommend using **dynamic thresholding**. <br>
87
 
88
  For a model comparison between **Mangled Merge V1** vs. **Flux.1 Dev**, check out this [post](https://civitai.com/posts/8666955). <br>
89