Sentinel7 commited on
Commit
b8cb1e8
·
verified ·
1 Parent(s): aff0917

Upload 1476909/2510724/README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. 1476909/2510724/README.md +122 -3
1476909/2510724/README.md CHANGED
@@ -1,10 +1,129 @@
1
  ---
2
  license: other
3
  tags:
4
- - civitai
5
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6
  Author: [tazmannner379](https://civitai.com/user/tazmannner379)
7
 
8
- Model: [https://civitai.com/models/1476909?modelVersionId=2510724](https://civitai.com/models/1476909?modelVersionId=2510724)
 
 
9
 
10
- Mirror: [https://civarchive.com/models/1476909?modelVersionId=2510724](https://civarchive.com/models/1476909?modelVersionId=2510724)
 
1
  ---
2
  license: other
3
  tags:
4
+ - style
5
  ---
6
+
7
+ # Penis Lora (+Blowjob, +Cumshot) [Taz] - WAN 2.2 14b / 5B / 1.3b T2V & I2V (Wan 2.1 & 2.2) + Qwen + Zimage Turbo-Qwen V2 - 2510724
8
+
9
+ **Model Type**: LORA
10
+
11
+ **Base Model**: Qwen
12
+
13
+ **Trigger Words**: PENISLORA
14
+
15
+ **Tags**: style
16
+
17
+ ## Gallery
18
+
19
+ <table>
20
+ <tr>
21
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC84YjA5YWRhMi1lY2Y2LTQ5YjAtOTI3ZS05NWE2MjA4Yzc1YzA=" width="200"></td>
22
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC9mZTIxN2UxYi05ZWVmLTQ1ODYtYTdiMy1hYTA2M2U0MWQyMDI=" width="200"></td>
23
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC9hMWQ0MWJhYS1jZjI0LTQ2YTktYjQwZC1mNTg3NTI5MjRjZWM=" width="200"></td>
24
+ </tr>
25
+ <tr>
26
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC80YjVlNTFlYS1kY2M1LTRjMjUtOGJkZi05ZDhhN2E0NDZmZjY=" width="200"></td>
27
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC83Y2MwMzBiOS0yNjc5LTQyYmItYWE0ZS0zYWMzZWYxMjExOTA=" width="200"></td>
28
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC9kOTIyOWM5My0yMTk3LTQxOGQtOWY0Zi1iOGE0Mzg4ZDVlNGI=" width="200"></td>
29
+ </tr>
30
+ <tr>
31
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC8xOGFhNjI2MS02NGU5LTRmOTgtOGZkYy00OGFiY2YyMjdmODY=" width="200"></td>
32
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC8zMzVhMTQ1MS1jNTMxLTQ3NWEtYThlMC05MmRkM2JlOGNjYWI=" width="200"></td>
33
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC82MTVjNmZlZC1jNzM4LTRlNmUtOTBhZS01MDhkMWM2NGJjODA=" width="200"></td>
34
+ </tr>
35
+ <tr>
36
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC9jNjUyZDdmMy1mNzRmLTQ0M2UtOWJhMy1iNTkxNmFkNGMyMmI=" width="200"></td>
37
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC83ZTZkMDY2MS02MGMwLTQ3NTMtYTFkMC04MzZmNWFhNmQxZDM=" width="200"></td>
38
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC8xZGY1NmY2Yy00NDkwLTQzMTUtOGVmNy1lNWJjNmQ4ZGEyOWQ=" width="200"></td>
39
+ </tr>
40
+ <tr>
41
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC80ZDE3MDM5OS1jYWVhLTRiZDktYjgwZC1kNjUzYTZhMTdjOTY=" width="200"></td>
42
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC80YTgwY2RhYi03OGMxLTRhODUtYWUxYS02OTllMmFmODliYjI=" width="200"></td>
43
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC8yZGViYzdlOS02ZGJiLTQzMDUtYTRmMy1jOTljMjQxNmRiYzg=" width="200"></td>
44
+ </tr>
45
+ <tr>
46
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC9mYTAzMTIzYy0xYWYzLTRkYTUtOTg4ZC0yMmE3NmJkMjQxMTM=" width="200"></td>
47
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC9iMTcwMmQwMS1mODk4LTRjODktYTg2Zi05MWJjODk3YzYwNzI=" width="200"></td>
48
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC8yMGZmOTRlOC00YzRhLTRiM2UtYmRhZC1lNWNmM2IxMjFkNzE=" width="200"></td>
49
+ </tr>
50
+ <tr>
51
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC9lMGJiYTRhMC0zOGY4LTRiYTUtYmEyZS1iMWJlNjE0MWQyYjI=" width="200"></td>
52
+ <td><img src="https://img.genur.art/sig/width:450/quality:85/aHR0cHM6Ly9jLmdlbnVyLmFydC82Zjk0N2I4Yy04MTE1LTQ1YzUtYTZlYS1hYmM4OTk2MTU0OGE=" width="200"></td>
53
+ </tr>
54
+ </table>
55
+
56
+ ## Description
57
+
58
+ About this version (qwen v2)
59
+
60
+ A brushup of the dataset, with better result than V1 for qwen. I will use the same dataset again for zturbo when the base model is released.
61
+
62
+ **Trigger word:** PENISLORA
63
+
64
+ **What can this lora do?**
65
+
66
+ This lora can add erect penises to both men or women viewed from the front/side. Other angles such as POV may have a backwards penis head.
67
+
68
+ Other things it can now do:
69
+ Side view of the penis
70
+
71
+ Cumming / Cumshots
72
+
73
+ Blowjobs (its captioned for the words "blowjob" and "deepthroat" )
74
+
75
+ What can't it do?
76
+
77
+ No penetration in the training data. Also nothing from POV angle, though there is a few images from above and 1 POV video in the training data.
78
+
79
+ Sometimes blowjobs with cumming have the penis slip out the closed mouth.
80
+
81
+ **Recommended Settings**
82
+
83
+ It works pretty good with the new lightning dyno high model. I'll link to it in my example workflow. I like to use dyno high model (no lightning lora), then for low I use the lightning v2 lora on the regular 2.2 low base model.
84
+
85
+ **Dataset**
86
+
87
+ 84 images at 512x resolution
88
+
89
+ 43 videos at 256x resolution
90
+
91
+ (I let DP pick the aspect ratio automatically)
92
+
93
+ This is the same exact dataset as the 2.2 5B model. I made no changes.
94
+
95
+ **Training**
96
+
97
+ I used the default diffusion pipe settings.
98
+
99
+ [optimizer]
100
+
101
+ type = 'adamw\_optimi'
102
+
103
+ lr = 2e-5
104
+
105
+ betas = [0.9, 0.99]
106
+
107
+ weight\_decay = 0.01
108
+
109
+ eps = 1e-8
110
+
111
+ I was baffled why it was taking so long to train the high until I realized after over 60 hours of training that I had put my videos in the images directory which resulted in the high being trained ONLY only on videos and twice (once with a very high resolution). Once I fixed this, I went back and trained from 11K steps up to around 13K with the images in the training data. The high model was fine without to be honest.
112
+
113
+ For the low, I trained it properly with videos and images the whole way, around 6K steps in I upped the image resolution from 512 to 1024 actually and didn't get an OOM (it fit around 24GB exactly). I trained it to around 10.5K steps. Also I trained the low on the full timestep range (0 to 1 instead of 0 to 0.85) from some advice, it may switch better over from high to low on the speed up lora with low steps.
114
+
115
+ I think I might do another version with more angles such as POV and from the behind to make this work for any situation. In that case I don't think it needs 10K steps per training session, epochs around 5K steps looked fine.
116
+
117
+ **The results**
118
+
119
+ I think it was a combination of improved captioning and 2.2 base model being better. But this lora turned out really well.
120
+
121
+ ---
122
+
123
  Author: [tazmannner379](https://civitai.com/user/tazmannner379)
124
 
125
+ Model: [CivitAI Model Page](https://civitai.com/models/1476909?modelVersionId=2510724)
126
+
127
+ Archive: [CivArchive Page](https://civarchive.com/models/1476909?modelVersionId=2510724)
128
 
129
+ <!-- Version: 20260425_upload -->