🎉 New Year Sale: Get 20% OFF on all plans — Use code NEWYEAR2026.

Upgrade now
XSMerge-RealisticVisionV3-ForArchitectural thumbnail

ModelsLab/XSMerge-RealisticVisionV3-ForArchitectural

xsmerge-realisticvisionv3
Text to Image Community ModelFree for Premium UsersLLMs.txt

注意:本模型并非原创模型,只在其基础上对目前的行业使用习惯进行了简单的微调,原模型作者链接:Realistic Vision V3.0 - v3.0 (VAE) | Stable Diffusion Checkpoint | Civitai

使用建议:

自RV3.0作者发布模型以来,很多建筑、室内向的AI研究者反应模型在这些方向的表现有些单一,以至于部分prompt如同陷入了过拟合等类似的情况(列如出图时玻璃的框架几乎都是黑色金属,需要特别标注prompt才能改善等等),并不能很好的进行多方位的图面表现。我将RV3.0模型从底层参数上进行了微调,让其模型更偏向于建筑等我所在研究的行业中,以达到让更多使用者有更好使用体验的目的。我在微调后进行了merge,是想要达到很好的泛化性,但似乎丧失了一些真实感。在出图时看似图面的质量会略微低于RV3.0一些,但是可以通过我所制作的轻量模型(Lora等)进行改善,甚至使用后出图会高于原本只用RV3.0模型的出图。

 

Since the release of the model by the RV3.0 authors, many architectural and interior AI researchers have responded that the performance of the model in these directions is somewhat simple, so that some prompt seems to fall into similar situations such as overfitting (for example, the glass frame is almost all black metal when drawing, requiring special marking prompt to improve, etc.). It is not very good to carry out multi-directional graphic performance. I fine-tuned the RV3.0 model from the underlying parameters to make the model more oriented to the industry I studied, such as architecture, so as to provide more users with better experience. I did merge after fine-tuning to achieve good generalization, but it seems to have lost some of its realism. In the drawing, it seems that the quality of the drawing surface will be slightly lower than RV3.0, but it can be improved by the lightweight model (Lora, etc.) I made, and even the drawing will be higher than the original drawing using only RV3.0 model.

 

RV3.0の作者がモデルを発表して以来、大勢の建築・室内向けのAI研究者は、モデルの表現がやや画一的なため、一部のpromptが過適合に陥っていると反応している(図の場合、ガラスのフレームはほとんどが黒金属で、promptを明記する必要があるなど)。多面的なグラフィック表現がうまくできていない。RV3.0モデルを基礎となるパラメーターから微調整し、建築など私の研究する業界に近づけることで、より多くのユーザーに体験を楽しんでもらうことを目指した。一般化を意識して微調整してmergeをしたのが、実感がなくなってしまったようなってみたい。見た目の品質はRV3.0よりも若干落ちてけど、私が作った軽量モデル(Loraなど)で改善できので、この問題を解決した。使用後はRV3.0のみのものよりも高くなることもある。

 

配合一些TI, VAE等使用文件

如果您有任何问题,请通过[email protected]与我联系

Suggestions for use form XS:

TI, VAE and other documents for use

If you have any questions, please contact me at [email protected]

XSからの使用アドバイス:

TI、VAEなどのファイルを一緒で使ってください

何かご質問がございましたら、[email protected]にメールしてください。

API PlaygroundAPI Documentation

API Endpoint URL

Base URL for all API requests to this endpoint.

https://modelslab.com/api/v6/images/text2img

API Authentication

Authentication requires a valid API key included in the request. Generate and manage your API keys from your developer dashboard. Include the key in the key parameter for all API requests.

Integration Examples

Production-ready code samples for API integration

{
"prompt": "R3alisticF, hauntingly beautiful oriental necromancer, long flowing brown hair, bangs, darkly tanned skin, earrings, bone necklaces, dark eyeshadow, red lips, vibrant, front-laced transparent, filmy silk blouse, cleavage, holding skull, in a sandstone room lit by candles, High Detail, Perfect Composition, high contrast, silhouetted, chiascuro",
"model_id": "xsmerge-realisticvisionv3",
"lora_model": [],
"width": "1024",
"height": "1024",
"negative_prompt": "(worst quality:2), (low quality:2), (normal quality:2), (jpeg artifacts), (blurry), (duplicate), (morbid), (mutilated), (out of frame), (extra limbs), (bad anatomy), (disfigured), (deformed), (cross-eye), (glitch), (oversaturated), (overexposed), (underexposed), (bad proportions), (bad hands), (bad feet), (cloned face), (long neck), (missing arms), (missing legs), (extra fingers), (fused fingers), (poorly drawn hands), (poorly drawn face), (mutation), (deformed eyes), watermark, text, logo, signature, grainy, tiling, censored, nsfw, ugly, blurry eyes, noisy image, bad lighting, unnatural skin, asymmetry",
"num_inference_steps": "31",
"scheduler": "DPMSolverMultistepScheduler",
"guidance_scale": "7.5",
"enhance_prompt": false,
"key": "YOUR_API_KEY"
}

SDKs

Official SDKs

Production-ready SDKs and client libraries for all major programming languages

API Parameters

Technical specifications for API request parameters.

Field NameParameterTechnical Description
promptpromptprompt help in image generation
Modelmodel_idEnter model_id that can help in image generation
lora Modellora_modelNo description available
Widthwidthwidth of the image
Heightheightheight of the image
Negative Promptnegative_promptNegative prompt help in avoid things that you do not want in image
Stepsnum_inference_stepsNumber of inference steps
SchedulerschedulerSampling scheduler
Guidance Scaleguidance_scaleHow closely to follow the prompt (1-10)
enhance_promptenhance_promptAutomatically enhance the prompt