Sentinel7 commited on
Commit
8fa1b0a
·
verified ·
1 Parent(s): ff3efaf

Upload 2379205/2879113/README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. 2379205/2879113/README.md +219 -0
2379205/2879113/README.md ADDED
@@ -0,0 +1,219 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ tags:
4
+ - augmented
5
+ - bimbo
6
+ - boobs
7
+ - breast
8
+ - concept
9
+ - fake
10
+ - huge
11
+ - implants
12
+ - large
13
+ - plastic surgery
14
+ - tits
15
+ ---
16
+
17
+ # Breast Implants - v5.0 (experimental) - 2879113
18
+
19
+ **Model Type**: LORA
20
+
21
+ **Base Model**: Flux.2 Klein 9B
22
+
23
+ **Trigger Words**: None
24
+
25
+ **Tags**: augmented, bimbo, boobs, breast, concept, fake, huge, implants, large, plastic surgery, tits
26
+
27
+ ## Gallery
28
+
29
+ <table>
30
+ <tr>
31
+ <td><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/e2f991ab-6af2-4866-bb50-c33d3705afea/original=true/128588911.jpeg" width="200"></td>
32
+ <td><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/5a85d307-a4d2-4991-9b76-832d30e00da5/original=true/128589054.jpeg" width="200"></td>
33
+ <td><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/796266bc-b5d2-4c43-a225-787cb38db8b6/original=true/128589208.jpeg" width="200"></td>
34
+ </tr>
35
+ <tr>
36
+ <td><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/bd7c3a43-4f83-4fb1-89e6-dcdee96d0ce1/original=true/128589455.jpeg" width="200"></td>
37
+ <td><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/6038a12b-fe4f-4f60-ac43-94b007736d19/original=true/128589519.jpeg" width="200"></td>
38
+ <td><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/a8e2728a-5e82-4f6a-9ee7-f95e32e77fb5/original=true/128589522.jpeg" width="200"></td>
39
+ </tr>
40
+ <tr>
41
+ <td><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/2ebed486-50e9-4cf7-9ee7-1de12a206456/original=true/128589527.jpeg" width="200"></td>
42
+ <td><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/848ca581-5323-4909-9f46-953992a23405/original=true/128589553.jpeg" width="200"></td>
43
+ <td><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/b2f56826-4f81-40bc-90c0-1f62386ea370/original=true/128589592.jpeg" width="200"></td>
44
+ </tr>
45
+ </table>
46
+
47
+ ## Description
48
+
49
+ Feedback welcome!
50
+ -----------------
51
+
52
+ Support is appreciated, but never expected:
53
+ [ko-fi.com/rickdog](https://ko-fi.com/rickdog)
54
+
55
+ **About 4.0**
56
+ =============
57
+
58
+ * Improved stability and delineation between sizes.
59
+ * Pruned dataset and using all new training settings.
60
+ * This is also a LoKr instead of a LoRA now.
61
+ * Should be better in pretty much all situations to previous versions.
62
+ * Used with 4-step and base with turbo LoRA with good results. (Images I posted were with 4step Klein 9B, not base with turbo LoRA)
63
+ * Prompt adherence is better with more flexibility.
64
+ * Used euler with sgm uniform or beta for all testing
65
+ * Strength: 1.0 (I've used up to 1.5 but likely to get more images with bad anatomy)
66
+ * *Read "About Version 3.0" below for more info regarding settings and prompting tips.*
67
+
68
+ About Version 3.0
69
+ =================
70
+
71
+ * New dataset should allow for much better results across the board.
72
+ * No trigger word like previous versions, so use natural language
73
+ * If you want a certain, look describe it ad nauseum and the image usually responds well. (larger, perkier, etc.)
74
+ * Useful terms:
75
+
76
+ + Size: small, medium, large, huge, gigantic, massive, etc.
77
+ + Shape: round, perky, "bolted on" (these aren't as good as I wanted them to be)
78
+ + General Modifiers: very, extremely, significant, prominent, etc.
79
+ + Other: busty, bust, bustline, voluptuous, curvy, etc.
80
+ * Dataset captions follow this general layout:
81
+
82
+ ```
83
+ A medium shot of a girl with [size] breast(s) (implants) facing to the right. She is wearing [a tight, thin shirt] and [smiling]. Her breasts are [very perky] and [round]. The outline of her nipples peaks are visible.... [rest of prompt]
84
+ ```
85
+
86
+ * Some new features that sometimes work when prompted:
87
+
88
+ + Covering nipples with one hand/both hands
89
+ + Middle buttons unbuttons revealing cleavage
90
+ + Outlines of nipples peaks under clothing
91
+ + Better side profile images
92
+ + Better nipple quality and sizes
93
+ + Better areola quality and color
94
+
95
+ Settings:
96
+ ---------
97
+
98
+ * Sampler: Tested with euler and res2m (other likely work well too).
99
+ * Scheduler: simple, SGM uniform (most stable), beta, beta57
100
+ * Strength: I think the general use sweet spot is (1.15)-(1.25)
101
+
102
+ + Safe 1.00
103
+ + General use: 1.25
104
+ + Bigger but more glitchy: 1.50+
105
+
106
+ Feedback welcome!
107
+ -----------------
108
+
109
+ Support is appreciated, but never expected:
110
+ [ko-fi.com/rickdog](https://ko-fi.com/rickdog)
111
+
112
+ ---
113
+
114
+ **WARNING:**
115
+ ============
116
+
117
+ **If you are having issues with the clothing not covering the breasts then try the following:**
118
+
119
+ * Turn down the strength. You can achieve very large sizes with prompting alone, albeit it's inconsistent.
120
+ * Do not use the word "breast(s)" as this will trigger more skin to show.
121
+ * Use terms like bust, busty, bustline, etc. (example phrases):
122
+
123
+ ```
124
+ "A very busty girl..."
125
+ "girl with a large bust"
126
+ "she has a massive bust"
127
+ ```
128
+
129
+ * Describe clothing as large or well-fitted (examples phrases):
130
+
131
+ ```
132
+ "form-fitting dress"
133
+ "large tube top"
134
+ "large tight t-shirt with a crewneck"
135
+ ```
136
+
137
+ * Briefly tested with a breast slider and it seems to keep the clothes on better while still keeping the "fake" look
138
+
139
+ ---
140
+
141
+ About v2.5:
142
+ ===========
143
+
144
+ This version (v2.5) was supposed to be v3.0, but it's honestly not consistent or refined enough. After multiple training attempts, I decided to release v2.5 while I try to improve the LoRA for a proper v3.0. It should be more flexible with sizes, angles, and posing as well as how clothing works with breasts. Also, breast shape should be more flexible too.
145
+
146
+ **Usage**:
147
+
148
+ * I do most all of my testing with simple workflows so I can know the images I get will be similar to others.
149
+ * Primary used euler/simple or euler/sgm uniform at 4-8 steps
150
+
151
+ + The RES4LYF samplers work well
152
+ + Some of the DPMPP samplers work well.
153
+ + Let me know if you find something else...
154
+ * Strength of 0.5-1.5 seems to do ok. you can push it to 1.75 or even 2.0 at times.
155
+ * I will admit this version will expose breasts more readily at larger sizes/strengths (1.00+)
156
+ * Use natural language. Images were captioned with phrases like:
157
+
158
+ ```
159
+ -Medium shot of a girl with large breasts wearing...
160
+ -Close up selfie of a girl with huge breast implants.
161
+ -Wide shot of a girl facing away. She has very large breasts and is wearing...
162
+ ```
163
+
164
+ * Be wary of samplers/scheduler combos and step count. A distilled 4-step model like Klein is very sensitive to these parameters so adjust them accordingly.
165
+ * Fun to use with a breast slider!
166
+
167
+ ---
168
+
169
+ About V1 and V2:
170
+ ================
171
+
172
+ Used 70-80 photographs depicting women with breast implants. This LoRA can do both clothed and topless women. If topless, mention nipples or areolae. This LoRA was not trained on genitalia so add your preferred LoRA for that. It seems to play decently well with other LoRAs and does not appear to mess with faces all that much if at all.
173
+
174
+ There is no trigger word as all of the photos were captioned using natural language, but mention breast implants somewhere in the prompt. That being said:
175
+
176
+ Training images were subdivided into "small", "medium", "large", and "huge" breast implants. Using the terms has a much more minimal effect that anticipated, although some size change can be seen.
177
+
178
+ v2.0 changes
179
+ ============
180
+
181
+ * Now has a range of 1.0-2.0 although 1.0-1.5 is the sweet spot.
182
+ * Improved size differentiation.
183
+ * Some poses will break anatomy
184
+
185
+ v1.0 changes
186
+ ============
187
+
188
+ * Use a strength of 1.0 with the term "small" or "medium" breast implants to achieve the smaller size.
189
+
190
+ * Use a strength of 1.5 with the term "large" or "huge" breast implants to achieve the larger size.
191
+ * Adding "very" before the term can further influence size
192
+ * If you are wanting more control you can use a breast slider lora for smaller or much much larger sizes.
193
+
194
+ I use it with the distilled model. Haven't tried with the base.
195
+
196
+ CFG:1
197
+
198
+ Steps: 4
199
+
200
+ Sampler/scheduler combos I have found success with:
201
+
202
+ * euler/beta
203
+ * res\_2s/beta
204
+ * euler cfg pp/beta57
205
+ * whatever you like best....
206
+
207
+ (I'm sure there are better sampler/scheduler combos out there, so let me know if you have suggestions)
208
+
209
+ Feedback welcome.
210
+
211
+ ---
212
+
213
+ Author: [RickDog](https://civitai.com/user/RickDog)
214
+
215
+ Model: [CivitAI Model Page](https://civitai.com/models/2379205?modelVersionId=2879113)
216
+
217
+ Archive: [CivArchive Page](https://civarchive.com/models/2379205?modelVersionId=2879113)
218
+
219
+ <!-- Version: 20260428_upload -->