Hugging Face's logo

Sennke
/
RealCosplayAnima_All-IN-ONE_FullFinetuned

Text-to-Image
English
realistic
3D
anime
anima
CG
danbooru
cosplay
Model card Files
xet
 Community
1
RealCosplayAnima_All-IN-ONE_FullFinetuned
File size: 2,242 Bytes
9497978
 
bd2e655
9497978
bd2e655
 
 
 
 
 
 
 
 
 
 
 
 
8994c20
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a99af8b
8994c20
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
---
license: other
license_name: derivative.model-circlestone-labs-non-commercial-license
license_link: LICENSE
language:
- en
base_model:
- circlestone-labs/Anima
pipeline_tag: text-to-image
tags:
- realistic
- 3D
- anime
- anima
- CG
- danbooru
- cosplay
---

# RealCosplay&Anime&3D

*A full fine-tuned Anima model designed to bridge Anime, Cosplay, Realistic Photography, and 3DCG styles in a single checkpoint.*

While preserving Anima's character knowledge and artist-style tag understanding, this model introduces a large amount of realistic visual knowledge to improve generalization

<Gallery>
  
![Section 1](https://cdn-uploads.huggingface.co/production/uploads/64972aa55551ac0d48be42dc/Irwts_6u8cTSdsQed4REY.jpeg)
  
![Frame 1](https://cdn-uploads.huggingface.co/production/uploads/64972aa55551ac0d48be42dc/cp7bwvVtpUpIvhhxFTS1g.jpeg)
</Gallery>

### More Examples

Civitai:
https://civitai.red/models/2671085/realcosplayandanimeand3d-all-in-one-animabase

## Recommended Prompts

Positive:

```text
masterpiece, best quality, newest, high resolution
```

For stronger realism:

```text
photo (medium)
```

Recommended weight:

```text
photo (medium): 0.5 ~ 3.0
```

Negative:

```text
worst quality, low quality, lowres, signature, text, logo
```

## Suggested Parameters

| Setting | Recommendation |
|----------|----------|
| Sampler | ER SDE / Euler SDE / Heun / DPM++ 2M SDE |
| Scheduler | Beta / Simple |
| CFG | 3 ~ 5 |
| Realistic CFG | 4.5 |
| Shift | 3 |

## Artist Styles

Supports most Danbooru artist tags inherited from Anima.

```text
artist:@artist_name
```

## Notes

- `solo` is not recommended and may lead to more 2D-style outputs. Use `solo focus` instead.
- Redundant tags such as `shirt, red shirt` are unnecessary, as duplicate and synonymous tags were filtered during training.

## Training

Full Fine-Tuning using:

https://github.com/wochenlong/lora-scripts-next

Special thanks to Chenkin (且听风吟) for the training framework and technical support.

## License

This model is a derivative work based on Anima.

It follows the original Anima license from Circle Stone Labs. Please review the included LICENSE file before use.

Commercial use is subject to the terms of the original Anima license.