Sentinel7 commited on
Commit
8282722
·
verified ·
1 Parent(s): 77637c4

Upload 2909069/3404835/README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. 2909069/3404835/README.md +80 -0
2909069/3404835/README.md ADDED
@@ -0,0 +1,80 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ tags:
4
+ - style
5
+ ---
6
+
7
+ # Retro 90's Anime / Golden Boy Style Lora - H3 - v2.0 - 3404835
8
+
9
+ **Model Type**: LORA
10
+
11
+ **Base Model**: MiniMax H3
12
+
13
+ **Trigger Words**: retro 90s 2d anime style, golden boy style,
14
+
15
+ **Tags**: style
16
+
17
+ ## Gallery
18
+
19
+ <table>
20
+ <tr>
21
+ <td><video src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/28081fa4-48d4-4a43-bafd-176c215964fa/original=true/145643214.mp4" width="200" controls muted autoplay loop></video></td>
22
+ <td><video src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/1bb3558b-b8c2-4fab-a1fc-74d4dd8bbebb/original=true/145643279.mp4" width="200" controls muted autoplay loop></video></td>
23
+ <td><video src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/af4a85c0-76c6-4701-a985-cf51c901d74c/original=true/145643471.mp4" width="200" controls muted autoplay loop></video></td>
24
+ </tr>
25
+ <tr>
26
+ <td><video src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/c26a0b58-2855-4c5c-af8d-62bfa4a4eedf/original=true/145643355.mp4" width="200" controls muted autoplay loop></video></td>
27
+ <td><video src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/4be5210e-ef3f-413d-bd4f-e6461de1f14a/original=true/145645731.mp4" width="200" controls muted autoplay loop></video></td>
28
+ <td><video src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/0d8d6d33-66a8-424c-8b84-49ce04fede80/original=true/145643531.mp4" width="200" controls muted autoplay loop></video></td>
29
+ </tr>
30
+ <tr>
31
+ <td><video src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/8a6578b0-b841-4b11-84a3-912f3d81644a/original=true/145643436.mp4" width="200" controls muted autoplay loop></video></td>
32
+ <td><video src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/53b3231f-3a55-476f-bbe2-1e96fe0565ac/original=true/145643854.mp4" width="200" controls muted autoplay loop></video></td>
33
+ <td><video src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/c809bc0c-867b-4b29-9df1-22a5d2bb1e4a/original=true/145643749.mp4" width="200" controls muted autoplay loop></video></td>
34
+ </tr>
35
+ <tr>
36
+ <td><video src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/546233e7-3829-46be-9b71-8d4571c06009/original=true/145643730.mp4" width="200" controls muted autoplay loop></video></td>
37
+ </tr>
38
+ </table>
39
+
40
+ ## Description
41
+
42
+ [Please follow me on X/Twitter if you're interested in my work. I will be posting there more often with early looks at what I'm working on.](https://x.com/tazmannner)
43
+
44
+ V2:
45
+ ---
46
+
47
+ Trained using 489 clips + 356 images from the show at the native 768x576 resolution. This is my first time doing images + videos together with the Ostris adapter v3 on musubi fork and the result is great! The motion is more alive and fluid than the base model comparisons. I think it makes things more interesting while bringing in the style from the original show. Same triggers as V1.
48
+
49
+ Images were microbatching with 8; videos up to 124 frames max length thus the low # of steps.
50
+
51
+ V1:
52
+ ---
53
+
54
+ Trained on 369 screen captures from Golden Boy at native res 768x576.
55
+
56
+ H3 is wonderful, it knows anime quite well, I will potentially try training it on video clips from the show but it does a pretty good job with just images.
57
+
58
+ It trains mega fast on images at this res, around 1.1s/it and this lora trained in around 3 hours. I'll do some extra training tomorrow but I am happy enough with the 12k checkpoint.
59
+
60
+ ### Trigger:
61
+
62
+ retro 90s 2d anime style, golden boy style
63
+
64
+ Not tested with any speed up loras like lightx. All examples are t2v 1mp 20 steps, euler beta57.
65
+
66
+ ### Training:
67
+
68
+ Trained using musubi tuner fork, lr 5e-5, --h3\_spatial\_density\_jitter 0.2 --h3\_guidance\_distillation\_scale 3.5 --h3\_base\_preservation\_loss\_weight 0.02 \
69
+
70
+ Rank 16
71
+
72
+ ---
73
+
74
+ Author: [tazmannner379](https://civitai.com/user/tazmannner379)
75
+
76
+ Model: [CivitAI Model Page](https://civitai.com/models/2909069?modelVersionId=3404835)
77
+
78
+ Archive: [CivArchive Page](https://civarchive.com/models/2909069?modelVersionId=3404835)
79
+
80
+ <!-- Version: 20261014_upload -->