README.md

---
license: other
license_name: bespoke-lora-trained-license
license_link: https://multimodal.art/civitai-licenses?allowNoCredit=True&allowCommercialUse=Rent&allowDerivatives=True&allowDifferentLicense=True
tags:
- text-to-image
- stable-diffusion
- lora
- diffusers
- template:sd-lora
- concept
- portrait
- parallelview
- stereoscopic
- crossview

base_model: runwayml/stable-diffusion-v1-5
instance_prompt: 
widget:
- text: 'monkey king,
masterpiece, high detail, 8k, high detailed skin, 8k uhd, high quality
'
  output:
    url: >-
      2690164.jpeg
- text: 'a young lady holding a cat, full body
masterpiece, high detail, 8k, high detailed skin, 8k uhd, high quality
'
  output:
    url: >-
      2690163.jpeg
- text: 'a young woman with glasses in campus, half body,
masterpiece, high detail, 8k, high detailed skin, 8k uhd, high quality
'
  output:
    url: >-
      2690160.jpeg
- text: 'a young man in a park,
masterpiece, high detail, 8k, high detailed skin, 8k uhd, high quality
'
  output:
    url: >-
      2690167.jpeg
- text: 'a princess in front of a castle, half body
masterpiece, high detail, 8k, high detailed skin, 8k uhd, high quality
'
  output:
    url: >-
      2690166.jpeg

---

# Stereoscopic Portrait - 3D 立体肖像 

<Gallery />



([CivitAI](https://civitai.com/models/152741))

## Model description

<p>This is an experimental project designed to produce <span style="color:rgb(253, 126, 20)">Stereoscopic Portraits</span>. It supports both <span style="color:rgb(253, 126, 20)">Parallel-view</span> and <span style="color:rgb(253, 126, 20)">Cross-view</span> modes, corresponding to two downloadable versions.</p><p><strong>Performance:</strong></p><p>When used as recommended, the subject of the portrait achieves a relatively high success rate. The background, in contrast, might require multiple attempts and might achieve a lower success rate.</p><p><strong>User Guide:</strong> For optimal results and image quality, strictly follow the procedure outlined below:</p><ul><li><p><strong>Checkpoint:</strong> Realistic Vision V5.1</p></li><li><p><strong>Textual Inversion:</strong> ng_deepnegative_v1_75t</p></li><li><p><strong>Positive prompts:</strong> {prompt}, masterpiece, high detail, 8k, high detailed skin, 8k uhd, high quality</p></li><li><p><strong>Negative prompts:</strong> ugly, disfigured, deformed, worst quality, low quality, ng_deepnegative_v1_75t</p></li><li><p><strong>LoRA Weight:</strong> 0.8-1.0</p></li><li><p><strong>Image Dimensions:</strong> Width: 832, Height: 560</p></li><li><p><strong>Optional:</strong> Use Hires. fix to enhance image quality with a Denoising strength of 0.3.</p></li></ul><p><strong>Note:</strong></p><ul><li><p>Using the Realistic Vision V5.1 in tandem with ng_deepnegative_v1_75t can significantly improve the success rate and generated image quality. See the "Suggested Resources" section below for more details.</p></li><li><p>This LoRA is specifically designed for generating Stereoscopic Portraits with a human subject. Without a human subject, a 3D effect cannot be achieved.</p></li><li><p>For detailed information on Stereoscopy, refer to the Wikipedia entry titled "<a target="_blank" rel="ugc" href="https://en.wikipedia.org/wiki/Stereoscopy">Stereoscopy</a>".</p></li></ul><hr /><p>这是一个实验性项目，用来生成 <span style="color:rgb(253, 126, 20)">立体视觉肖像（Stereoscopic Portrait）</span>，支持 <span style="color:rgb(253, 126, 20)">Parallel-view</span> 与 <span style="color:rgb(253, 126, 20)">Cross-view</span> 两种模式（对应两个下载版本）。</p><p>在按照推荐方法使用的前提下，人物主体可以获得比较高的成功率，背景的成功率相较人物主体要低一些，有时需要多次尝试。</p><p><strong>使用说明：</strong>（严格按照以下流程生成图片，可以获得比较高的成功率与图像质量。）</p><ul><li><p><strong>模型（checkpoint）：</strong>Realistic Vision V5.1</p></li><li><p><strong>嵌入式（Textual Inversion）：</strong>ng_deepnegative_v1_75t</p></li><li><p><strong>正面提示词：</strong>{prompt}, masterpiece, high detail, 8k, high detailed skin, 8k uhd, high quality</p></li><li><p><strong>负面提示词：</strong>ugly, disfigured, deformed, worst quality, low quality, ng_deepnegative_v1_75t</p></li><li><p><strong>LoRA 权重：</strong>0.8-1.0</p></li><li><p><strong>图像尺寸：</strong>宽度：832，高度：560</p></li><li><p><strong>可选：</strong>使用高清修复（Hires. fix）提升图像质量，重绘幅度（Denoising strength）0.3</p></li></ul><p><strong><em>注意：</em></strong></p><ul><li><p>配合使用 Realistic Vision V5.1 与 ng_deepnegative_v1_75t 可以极大提高成功率与生成图像质量。具体见下方“Suggested Resources”。</p></li><li><p>此 LoRA 只用于生成带有人物主体的立体视觉肖像图，如果不带有人物主体则无法产生立体效果。</p></li><li><p>关于立体视觉（Stereoscopy）的具体说明，参见维基百科“<a target="_blank" rel="ugc" href="https://zh.wikipedia.org/zh-cn/%E7%AB%8B%E4%BD%93%E5%9B%BE">立体图</a>”词条。</p></li></ul>



## Download model

Weights for this model are available in Safetensors format.

[Download](/lordjia/stereoscopic-portrait-3d-li-ti-xiao-xiang/tree/main) them in the Files & versions tab.

## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)

```py
from diffusers import AutoPipelineForText2Image
import torch

pipeline = AutoPipelineForText2Image.from_pretrained('runwayml/stable-diffusion-v1-5', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('lordjia/stereoscopic-portrait-3d-li-ti-xiao-xiang', weight_name='stereoscopic_parallel_v1.0.safetensors')
image = pipeline('a princess in front of a castle, half body
masterpiece, high detail, 8k, high detailed skin, 8k uhd, high quality
').images[0]
```

For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)