English
English
Most LoRA models are trained by users themselves. Here are some commonly used LoRA models. If you don’t like the following recommended models, you can also download them at a third-party model site by yourself, such as civitai: https://civitai.com/models , or Hugging Face: https://huggingface.co/models
MoXin - MoXin 1.0 | Stable Diffusion LoRA | Civitai
"Mo Xin" - In the past, Juanzi's "Quin Xin" and Wangsun's "Questive Heart", the heart is so beautiful, so it is used.
This product is made up of freehand works by masters such as Anji Wu Cangshi, Xinghua Banqiao, Bada Shanren, Shanyin Bonian, etc., supplemented by modern characters training. Supplemented with appropriate prompt words, recite the titles of the ancestors, and use them in both ancient and modern times, and clarify the beauty of the common people.
Let the whole world appreciate the unique charm of Chinese ink painting!
The wonder of the Mo Xin model is its wide applicability and can adapt to various types of models, whether it is real characters or anime characters, it can present amazing results.
checkpoint: ChilloutMix, Guofeng 3.2, etc.
Weight: below 0.85
CFG: Change Style
1~3: Low and low-quality
3~7: Gradually mechanized
Linear/comic style LoRA model, use this LoRA to generate excellent black and white liner drawings. The lines are smooth and controllable in thickness, and whether they are comic-style images or comic production, they can help us create amazing works.
V3.0 version has been fine-tuned to improve the effect of generating characters' limbs and complex backgrounds. Overall, it tends to be more comic-style than pure line-text.
checkpoint: Anything V4.5
Weight: 1
CFG: 7, low value can make lines thinner
3D rendering style - 3DMM_V11 | Stable Diffusion LoRA | Civitai
This universal 3D cartoon Lora model is used to describe the style of 3D rendering effects. I usually use it to convert 2D icons to 3D icons, suitable for line drawings with colors or hand-drawn drawings, to generate images on the right with stereo effects from flat icons on the left. It is recommended to use RevAnimated large models, combined with 3D-style Lora models such as blindbox and this model, while using Controlnet Lineart, Depth and Tile to control the edges, depth and color of the render.
checkpoint: revAnimated
ControlNet:
Lineart
Depth
Tile
After the release of the famous anime mockup Pony Diffusion V6 XL, many LoRA models and retrained big models quickly emerged.
Based on training of about 2.6M images, the LoRA model training data covers a variety of images such as anime, cartoon, furry animals and Pony Pony. This LoRA, paired with the Pony Diffusion XL model, creates breathtaking visual effects of anthropomorphic, wild or humanoid species and their interactions with just simple prompt words.
checkpoint: Pony Diffusion XL
Weight: 1
CFG: 5
SamDoesArts (Sam Yang) Style LoRA - offset | Stable Diffusion LoRA | Civitai
This LoRA model has modest flexibility to generate images with similar visual effects to Sam Yang on AnyPastel. Sam Yang is a post-95 illustrator who is popular on the Internet. His works have beautiful colors and a unique European and American thick paint style, giving people a full sense of atmosphere.
checkpoint: AnyPastel, AnyLoRA and AbyssOrangeMix2
Weight: Offset version 1, original version 0.65
CFG: 7 ~ 11
KoreanDollLikeness (v2.0) - v2.0 | Stable Diffusion LoRA | Civitai
This LoRA model is popular and powerful and can be used to draw cute female images with Korean idol style. The generated images usually have delicate facial features, gorgeous clothing and stylish hairstyles, showing a strong sense of reality that they were temporarily removed from Station C for being used for crimes.
checkpoint: chilloutmix, majic
Weight: 0.2 ~ 0.7
CFG: 5 ~ 9
JapaneseDollLikeness (v1.5) - v1.5 | Stable Diffusion LoRA | Civitai
The same author as the KoreanDollLikeness above, and the author also recommends several other excellent models: ChinaDollLikeness and TaiwanDollLikeness and ThaiDollLikeness.
Refer to KoreanDollLikeness.
Doll Likeness - by EDG - Bond Doll Likeness | Stable Diffusion LoRA | Civitai
This LoRA series has similar features to KoreanDollLikeness, focusing on realistic rendering of faces from different countries, including facial features, unique makeup and hairstyles, and body proportions. At present, it has covered many loras such as France, Colombia, Argentina, Poland, Egypt, the Philippines, Greece, Canada, Blonde, and Blue Eyes.
checkpoint: Various models (from 2D to real style)
Weight: 0.7 ~ 0.9
CFG: 5 ~ 8
Detail Tweaker LoRA - v1.0 | Stable Diffusion LoRA | Civitai
This LoRA model is used to enhance or weaken details while maintaining overall style and features. For example, hair details can be enhanced to make the image look more realistic, or facial details can be reduced, rough and frizzy. It can be used with various basic models (whether it is anime style or realistic style), style and feature LoRA, and can maintain overall consistency well.
checkpoint: Various models (from 2D to real style)
Weight: -2 ~ 2 (positive numbers enhance details, negative numbers weaken details)
CFG: Match the model
xl_more_art-full / xl_real / Enhancer - xl_more_art-full-v1 | Stable Diffusion LoRA | Civitai
This LoRA model is SDXL version and can enhance and improve images to make them more aesthetically pleasing, artistic and creative. In addition, the model trains aerial photography, so it can be used to create interesting aerial images. Functionally similar to our commonly used LoRA model (add-detail-xl) that adds details, the difference is that xl_more_art-full is better at improving the artistic effect of the picture.
checkpoint: sd_xl_base_1.0_0.9vae
Weight: 0.8 ~ 1, no more than 1
CFG: 4 ~ 8
400GB-LoRA-XL-Repository | Civitai
Models are carefully extracted from a variety of models, and can be mixed and matched with various styles to create unique and artistic works. Rather than simply replicas, these selected LoRAs capture the essence of the original model, adding creative influence to “in style” or “influenced by…”. The ability to generate images with rich details and high-quality diversity provides endless options and creative space.
checkpoint: FFXL-400-v3
Weight: 0.2 (subtle details) ~ 2 (strongly changing)
CFG: 7
epi_noiseoffset - v2 | Stable Diffusion LoRA | Civitai
This model uses noise offset (offset noise) technology to process image quality, and by enhancing contrast and saturation, it generates high-quality images with sharp contrast and light and darkness. Compared with the cumbersome attempts to add and modify the prompt words, the visual effect of the image can be more effectively improved and more vivid, rich and attractive.
checkpoint: Various models (from 2D to real style)
Weight: 0.5 ~ 3 (note that it may change the composition if it is too large)
CFG: Checkpoint and Sampler with the matching
DetailedEyes_XL - v3.0 | Stable Diffusion LoRA | Civitai
This is a LoRA model dedicated to finely beautifying the eyes, suitable for SDXL models, with the main purpose of generating more detailed and realistic eyes. This model not only enhances the beauty of the eyes, but also adds subtle details to the image, which significantly optimizes hair, clothing, skin, etc. Judging from the actual usage, the improvement effect on the cartoon picture is particularly obvious. Negative values are also effective, which can reduce eye details and detailed performance of elements such as hair and clothing.
checkpoint: Various models (from 2D to real style)
Weight: -1 ~ 1 (positive number enhances details, negative number weakens details)
CFG: Checkpoint and Sampler with the matching
Multiple XL Sliders - Age, weight, Hands, etc - age | Stable Diffusion LoRA | Civitai
Generally speaking, it is difficult to accurately control visual attributes such as age, hair, and eye size through text prompts alone. However, this powerful LoRA model uses the concept slider to adjust the character's visual attributes more effectively while minimizing the impact on other attributes as much as possible. It can shape the character image more accurately and meet the requirements for visual details.
checkpoint: Various models (from 2D to real style)
Weight: -2 ~ 2 (negative number reduces age)
CFG: Checkpoint and Sampler with the matching
Skin Tone Slider - LoRA - v1.0 | Stable Diffusion LoRA | Civitai
Similar to Multiple XL Sliders, it is also a concept slider type lora. It is a very practical skin tone adjustment tool that can be adjusted by operating the concept slider. Bidirectional slider adjustments can be made in the range of -5 to +5 without causing significant distortion. When the weight value is close to a negative number, the skin color will become whiter; when the weight value is close to a positive number, the skin color will become darker. According to the generated image, when the skin color is whiter, the face shape of the character will be more European-oriented, and the hair color will change accordingly.
checkpoint: Various models (from 2D to real style)
Weight: -5 ~ 5 (negative number turns white)
CFG: Checkpoint and Sampler with the matching
Muscle Slider - LoRA - v1.0 | Stable Diffusion LoRA | Civitai
In addition, there is a concept slider type LoRA model for muscle adjustment. Bidirectional slider adjustments can be achieved in the range of -3 to +5 to achieve more natural muscle performance without causing excessive distortion. Positive values will increase muscle mass, while negative values will decrease muscle mass. However, a problem was found during use: as muscle mass increases, the character's skin tone becomes darker. This may be because bodybuilders use oil to enhance the visual effects of their muscles, and people with a well-melted muscle may be more likely to tan. To solve this problem, the aforementioned skin tone slider LoRA can be used to eliminate this effect, thereby better controlling the character’s skin tone while adjusting the muscle mass.
checkpoint: Various models (from 2D to real style)
Weight: -3 ~ 5 (positive muscle gain)
CFG: Checkpoint and Sampler with the matching
Hair Length Slider - LoRA - v1.0 | Stable Diffusion LoRA | Civitai
This LoRA model is also a concept slider type tool that helps adjust the hair length of a character. The slider can be adjusted in both directions from -8 to +8 to minimize distortion. Positive values will make hair longer, while negative values will make hair shorter. However, it should be noted that excessive value may lead to problems such as male characters having female characteristics or female characters having beards. In this case, it is recommended to use the Gender Concept slider to make adjustments to ensure that the character's appearance matches the gender settings.
checkpoint: Various models (from 2D to real style)
Weight: -8 ~ 8 (positive long hair)
CFG: Checkpoint and Sampler with the matching
People Count Slider - LoRA - v1.0 | Stable Diffusion LoRA | Civitai
This is a simple LoRA model that helps adjust the number of people in the picture, which can be adjusted left and right between -8 and +8 without causing too much distortion. The larger the value, the more people there will be. When adjusted to a larger value, a large number of people can be generated. Negative values are used to reduce the number of people, but the effect may not be ideal when used. By adjusting the parameters, the number of people can be easily controlled and a visual effect that meets the needs is presented.
checkpoint: Various models (from 2D to real style)
Weight: -8 ~ 8 (there are many positive people)
CFG: Checkpoint and Sampler with the matching
hanfu Hanfu - v3.0 | Stable Diffusion LoRA | Civitai
High-quality Chinese costume LoRA model supports a variety of styles, including Han style, Tang style, Song style, Ming style and Jin style. The clothing style is highly realistic and can more accurately present the characteristics of clothing in various dynasties. It is almost like the feeling of dreaming back to ancient dynasties. It is very suitable for the generation of ancient style photos and animation pictures.
checkpoint: chilloutmix, national style, ghostmix
Weight: 0.5 ~ 0.7
CFG: 6 ~ 8
This Lolita style LoRA model has a modern Victorian design concept and can present a wide range of images of Lolita dresses including styles, fabrics, colors and decorative details. Trained with realistic-style images, it may not perform well when using 2D large-scale graphics. The training focus of this model is on clothing, excluding other elements such as face and background. Therefore, it can be used with other LoRA models to create a richer and complete visual effect.
checkpoint: Real style (Mi Orange)
Weight: 0.65 ~ 0.85
CFG: 7 ~ 10
Royal Gowns - an EDG collection - Flower Gowns | Stable Diffusion LoRA | Civitai
Lora models related to the clothing series related to exquisite royal gowns that generate gorgeous dress decorations and scenes. In addition to the floral dresses in the picture below, they also cover Victoria's angel dresses, afternoon tea dresses, college-style dresses, fire dragon robes and mechanical gowns. Through the model, you can generate exquisite royal dress images in various styles, adding gorgeous and unique charm to the picture.
checkpoint: Various models (from 2D to real style)
Weight: 0.8 ~ 1
CFG: 5.5 ~ 8
A-Mecha Musume A body machine girl - SSS | Stable Diffusion LoRA | Civitai
This is a LoRA model specially used to generate the second-dimensional beauty of mecha. It is used in conjunction with real-life models to create characters similar to those in Japanese special films, or characters with cosplay style. The generated characters feature a streamlined body, robotic arms and unique head design that demonstrates the beauty of a perfect combination of technology and cuteness. The image of a mecha girl with different colors, hairstyles, clothing and accessories can be generated by using different prompt words.
checkpoint: Various models (from 2D to real style)
Weight: 0.7 ~ 0.8
CFG: 6 ~ 7
concept Head-mounted display - v1.0 | Stable Diffusion LoRA | Civitai
This LoRA model can generate images of various head-mounted displays (HMD)/virtual reality (VR) glasses. It is recommended to try adding some descriptive words to target specific image elements such as X marks, cross marks, heart marks, etc. Combined with neon colors, science fiction, cyberpunk and other styles, you can create pictures with unique science fiction cyberpunk characteristics. The disadvantage is that color and styling control needs to be improved.
checkpoint: Various models (from 2D to real style)
Weight: 0.8 ~ 1
CFG: 7 ~ 8
Urban Samurai | v0.14 | Clothing LoRA - urbansamurai_v0.3 | Stable Diffusion LoRA | Civitai
This LoRA can wear a character in a Techwear-style outfit and is suitable for most 2D, 2.5D and live models, especially for full-body characters. Easily add elements such as functional jackets, masks, tactical vests and handheld samurai to the image, creating modern and retro clothing, weapon accessories, and unique hairstyles and makeup, providing rich creative possibilities to generate images with urban samurai style. It should be noted that the current version may not be ideal for handheld samurai swords, but subsequent versions will be improved to improve quality.
checkpoint: Various models (from 2D to real style)
Weight: 0.5 ~ 0.9
CFG: 7 ~ 12
Gacha splash LORA - v4.0 | Stable Diffusion LoRA | Civitai
This model is specifically used to generate images similar to the Japanese gacha style, i.e., drawings with backgrounds. Its characteristics are bright colors and cute images, which are especially suitable for 2D character design and game portraits. The generated images conform to the core elements of the vertical painting style, including color matching, line fluency, character expressions and postures.
checkpoint: 2D style
Weight: 0.8 ~ 1
CFG: 7 ~ 11
Resolution: 640 x 640
paper cut ,background,pastel,papercut style,background,kk - v1.0 | Stable Diffusion LoRA | Civitai
This is a LoRA model with a paper-cut art style and blends soft pastel tones that can be used for stylized translation. It is especially suitable for presenting natural pictures such as flowers, rivers, lakes and seas in paper-cut style. It can also be used for character description. If face collapse occurs, it can be handled through the facial repair function. In addition, the model also has good results in animal performance. If you want to create a dazzling picture, you can combine the description of enhanced visual effects or match it with other Lora models, such as: (fractal art: 1.3).
checkpoint: realistic or 2.5D (revAnimated, dreamshaper)
Weight: 0.6 ~ 1
CFG: 7 ~ 10
Anime Tarot Card Art Style LoRA - v3.1 | Stable Diffusion LoRA | Civitai
This LoRA model focuses on the anime tarot card art style and has been trained in Phantom Star illustrations. It covers card design, color matching and symbolic symbols. You can choose according to your preferences and use the characters' LoRA model to generate pictures with unique styles. For negative embeddings, EasyNegative and badhandv4 are recommended.
checkpoint: 2D (Anything, orangeminx), 2.5D has unique effects
Weight: 0.8 ~ 1
CFG: 7 ~ 9
GHIBLI_Background - END | Stable Diffusion LoRA | Civitai
The model is able to draw a unique style with Japanese animation watercolor illustrations, clean and full of Ghibli. It has good performance in depicting characters and scenery. The overall background effect is stronger than the character effect. Whether it is a city, natural scenery or fantasy scene, the model has excellent performance.
Checkpoint: aniflat and other two-dimensional styles
Weight: 0.8 ~ 1
CFG: 5 ~ 8
Ink scenery | Ink landscape - v1.0 | Stable Diffusion LoRA | Civitai
The model focuses on creating ink-wash landscape-style images, which can easily generate landscape paintings with Chinese style. Its uniqueness is that it can generate delicate ink strokes and a three-dimensional sense of depth, thereby creating a surreal visual experience between 2D and 3D. Using it with different large models and other LoRA models can create a richer and more diverse artwork with ink style.
checkpoint: dreamshaper, darkSushiMix and RealCartoon3D, etc.
Weight: 0.8 ~ 1
CFG: 5 ~ 11
[LoCon/LoRA] Octans/Eight-Number Style - pynoiseLoHa-v2.0 | Stable Diffusion LoRA | Civitai
This is a LoCon model trained through 150 images of galaxies and ruins backgrounds. The Milky Way reminds the author of a long voyage, thus becoming the source of the model’s name. Using this model, you can design amazing visual effects such as night sky, reflections in the water, and flowers.
checkpoint: Various models (from 2D to real style)
Weight: 0.7 ~ 1
CFG: 7 ~ 8
XSArchi_127 New Science Fiction Neo Sci-Fi - v1.0 | Stable Diffusion LoRA | Civitai
The model can create a stunning science fiction-style architecture, cleverly utilize elements such as futurism, space exploration, high-tech elements, and cyberpunk to present visual effects beyond reality and full of endless imagination. The resulting building lines are smooth, and the materials and textures reflect the sense of futuristic technology.
checkpoint: revAnimated, dreamshaper, etc.
Weight: 0.75 ~ 1
CFG: 7 ~ 20
Cyberpunk Interior Design - v1.0 | Stable Diffusion LoRA | Civitai
This LoRA model focuses on cyberpunk-style interior design and can generate images with futuristic and technologically driven aesthetics. It combines cyberpunk elements such as high-tech materials and equipment, smooth styling and bold color matching, so as to make the generated interior design unique in sci-fi and futuristic style. Models are suitable for creating cyberpunk-style interior design images, as well as visual backgrounds in anime or cartoon style.
checkpoint: hellopure, revAnimated, theAllysMix, etc.
Weight: 0.8 ~ 1
CFG: 7
M_Scene Cartoon Scene - v2.0 | Stable Diffusion LoRA | Civitai
This model can depict a variety of colorful cartoon scenes, including individual buildings and sandbox environments, with cartoon, 3D and Pixar style features.
Small resolution produces single buildings, and large resolution produces sandbox.
checkpoint: revAnimated
Weight: 0.5 ~ 0.8
CFG: 7
Stylized Setting (Isometric) SDXL & SD1.5 - SDXL | Stable Diffusion LoRA | Civitai
The model is designed to generate game assets or scene concept maps with isometric perspectives, and is often used to present 3D scenes with a sense of scale and depth in two-dimensional images. The combination of prompts has infinite possibilities that can be created based on imagination, such as the combination of elements such as caves, crystals, magic and light. Supports SD1.5 and SDXL versions.
checkpoint: revAnimated, gameIconInstitute, etc.
Weight: 0.6 ~ 0.7
CFG: 7 ~ 10
size: The effect of drawing is different in different proportions. It is recommended to be 1:1. Here is a 512 x 768 image output example
Japan Architecture - Japan Interior v1.0 | Stable Diffusion LoRA | Civitai
A LoRA model specializing in the creation of classic Japanese architecture, including three models: shrine, hot spring and traditional and indoor. Japanese interior design, also known as the “Wa” (Haginal Style) or “Nihon” style, is a unique aesthetic style that reflects the cultural and traditional elements of Japan. The model can capture and reproduce the subtleties of Japanese architecture, such as its structural structure, building materials (such as tatami, wood, etc.), and unique decorative elements.
checkpoint: Various models (from 2D to real style)
Weight: 0.5 ~ 0.9
CFG: 7 ~ 9.5
concept Loong(china dragoneastern dragon) China Dragon- v1.0 | Stable Diffusion LoRA | Civitai
There are various dragon-drawing models on Station C, but there are significant differences between Chinese dragons and Western dragons. This model is different from the dragon in Western mythology. It focuses on presenting the Chinese dragon/oriental dragon image with traditional Chinese style characteristics, from the dragon's body shape, scale texture to posture expression. The model strives to accurately capture the unique charm of Chinese dragons.
checkpoint: Various models (from 2D to real style)
Weight: 0.7 ~ 1
CFG: 7 ~ 10
Cat/Cute cat/midjourney style cat Lora - v1.0 | Stable Diffusion LoRA | Civitai
A model with a cute cat picture in Midjourney style can be generated. The author has selected a large number of cat pictures generated by Midjourney 2D models for training. The effect is very good, the painting style is restored very well, and it shows a very cute cat image. Just trigger the word cat to get the effect, and its compatibility with other types of models is also very good.
checkpoint: ghost (stable), majic (realistic), meina (second dimension), irise (dream)
Weight: 0.5 ~ 0.9
CFG: 7
SDXL_Sacred_beast - v1.0 | Stable Diffusion LoRA | Civitai
This SDXL-lora model is themed on the theme of the divine beasts and uses various behemoths for training, presenting a strong sci-fi and mythological atmosphere. In addition, the model can also add decorative elements to characters and scenes, such as clouds, flames, sky and water surfaces, etc. It has a significant effect on enhancing the expressiveness and atmosphere of the picture, and can give the work a richer sense of layering. He is good at capturing unique features of divine beasts, such as complex textures, colors and postures. Use this model to shape the characters, monsters or environment elements in the game, injecting fantasy colors and attraction into the game.
checkpoint: sd_xl_base, dreamshaperXL, copaxTimelessxlSDXL
Weight: 0.5 ~ 0.8
CFG: 7
Dressed animals - SD XL | Stable Diffusion LoRA | Civitai
This LoRA model focuses on generating images of animals wearing costumes, providing a wealth of creative tools and possibilities. It is easy to add a variety of gorgeous costumes and decorative elements to the animals, creating unique and interesting pieces. The boundary between reality and imagination is blurred, thereby fully demonstrating creativity and imagination through animal images.
checkpoint: SDXL Base 1.0, dreamshaperXL
Weight: 0.5 (recommended with xl_more_art-full_v1) ~ 1
CFG: 3 ~ 10
Animal Mecha - v1.0 | Stable Diffusion LoRA | Civitai
A LoRA model focusing on creating creative images that combine animals with machinery, suitable for non-human only, generates semi-mechanical animals, robotic animals and mecha animals. Prompt words can include specific animal species (such as lions, tigers, eagles, etc.), mecha components (such as mechanical wings, armor, weapons, etc.), and the desired scene and background. It can also be used with LoRA models for other animals or animal characters, but pay attention to adjusting the weight accordingly.
checkpoint: real or 2.5D
Weight: 0.6 ~ 1
CFG: 7
For models specially developed for solo games, the generated pixel map is a real pixel map, rather than a pseudo-pixel map that is not a lattice. With clear edges, no blur or contamination, it is also quite compatible with other LoRA models, and can be used simultaneously with other characters LoRAs and with Pixel mash to achieve true alignment of pixel points. In addition, it can also be used to cooperate with ControlNet to control the bones using openpose to generate sequence frame animations.
checkpoint: anything4.5, anylora, ReV Animated
Weight: 0.6 ~ 1
CFG: 7
Pecha Swords Generator - v1.2 | Stable Diffusion LoRA | Civitai
This LoRA model can draw weapons such as swords, axes, bows, etc., and the author plans to expand more types in the future. In practice, not limited by the type of weapon that can be generated, prompt words can be provided and creative as needed, such as a hammer or spear. But the current version cannot make completely symmetrical weapons. In some cases, problems may occur, such as when trying to generate an axe, it generates a sword or a mixture of multiple weapons. Increasing the weight of the prompt word or including undesired types in the negative prompt word can help optimize the generation of results.
checkpoint: Abyss Orange, ReV Animated
Weight: 0.6 ~ 1
CFG: 6 ~ 8
CGgameguniconcsw - v1.0 | Stable Diffusion LoRA | Civitai
This is a LoRA model about weapons in the game, which can generate various different types of game weapons according to needs after calling the trigger word. Adding some extra tip words, such as the appearance, size, material, etc. of the weapon, can become different effects, thereby adding more visual effects and creative elements.
checkpoint: AAM XL (Anime Mix)
Weight: 0.8 ~ 1
CFG: 7
Ui-Ux (LORA) SD 1.5 & SDXL - SDXL | Stable Diffusion LoRA | Civitai
This LoRA model focuses on the conceptual design of quickly generating user interfaces. It covers layout, controls, colors and fonts, and maintains consistent style and quality. Currently, there are two versions of SD15 and SDXL. Although they cannot be used directly for prototyping, they can still serve as a source of inspiration for web design creativity.
checkpoint: AAM XL (Anime Mix)
Weight: 0.5 ~ 0.7
CFG: 7 ~ 10
Badass Cars - Badass Cars | Stable Diffusion LoRA | Civitai
This car design model has rich creative potential and can create a wide range of car types, including retro, modern, realistic and fantasy. Whether it’s a powerful monster truck in off-road terrain or a streamlined cutting-edge car full of cyberpunk, the possibilities are endless. Compatibility with other models is one of its main advantages, and can be combined with realistic models to achieve precise automotive design, or injected with anime style to add creativity, both of which provide endless possibilities for creation.
checkpoint: Various models (from 2D to real style)
Weight: 0.45 ~ 0.65
CFG: 7
Futuristic Aircraft/Spacecraft Generator Concept XL - v1.0 | Stable Diffusion LoRA | Civitai
This model focuses on futuristic spacecraft (spacecraft) design, which can generate aircraft of various shapes. From commercial jets, supersonic vehicles, to spacecraft exploring Mars, this model can create unprecedented high-quality concept maps that inspire unlimited imagination space.
checkpoint: Realism from HaDeS XL
Weight: 1
CFG: 7 ~ 10
Motorbike EX | Transportation LoRA - YHmotorbike V0.9 | Stable Diffusion LoRA | Civitai
This is a model dedicated to generating cool motorcycle/locomotive images. Not only can the image of the motorcycle be generated individually, but it can also be combined with characters, such as a girl riding a motorcycle. Different backgrounds can be set according to your needs, such as city, cyberpunk city, neon effect, outdoor, etc.
checkpoint: Various models (from 2D to real style)
Weight: 0.7 ~ 0.9
CFG: 6 ~ 7
【Y5】Blue print (Disassembly diagram) - v1.0 | Stable Diffusion LoRA | Civitai
This blueprint/demolition diagram LoRA model is based on anything4.5 pruned training and is designed to generate images similar to object blueprints. This model can be paired with LoRA models with added details, such as add_detail and epi_noiseoffset2, to show the internal structure of the object after being disassembled from various angles and to support the generation of blueprints/decomposition diagrams of different styles or levels of detail.
checkpoint: ReVAnimated, chilloutmix
Weight: 0.8 ~ 1
CFG: 7 ~ 10
Food Photography - v1.0 | Stable Diffusion LoRA | Civitai
This LoRA model focuses on generating a variety of food materials. In fact, the shooting effect is very visually attractive and looks super appetizing. This model helps to conceive a culinary feast, including a variety of dishes, desserts and drinks, creating more professional and attractive food materials. Whether it is the 2D style or the real-life food, this LoRA model is versatile and easily creates mouth-watering food photos.
checkpoint: Various models (from 2D to real style)
Weight: 0.6 ~ 0.8
CFG: 5 ~ 7
Guns (Lots Of Guns) XL (Choose from list) - Carl Gustaf m45 v1.0 | Stable Diffusion LoRA | Civitai
A LoRA series model focused on gun theme, trained based on the basic model xl 1.0. It covers a number of gun models, including shotgun, Desert Eagle, AK-47, revolver, etc. This model works well with realistic stock photos, and it has also been tested to have good compatibility with large models with other styles, such as ponyDuffisonXL.
checkpoint: real-life models (such as cinevisionXL, JuggernautXL, dreamshaperXL, etc.), other styles (such as ponyDuffisionXL), basic xl models with different XL Styles
Weight: 0.8 ~ 1
CFG: 4 ~ 10