hipoly — STABLE DIFFUSION 1.5 AI Model

(日本語は後半にあります) High-poly 3D Model LoRA This is a LoRA trained on high-polygon 3D model images. It can provide clean, high-resolution skin and hair materials, as well as detailed clothing elements. For those interested, I have compiled the technical insights gained during the training of ver.2 in the following article. Changes in 2.0 - Increased the number of training images - Increased the training resolution (ver.1: 768, ver.2: 896) - Revised the tagging - Removed the trigger word as well - Reviewed the training parameters Improvements in 2.0 - It can reproduce a more 3D-like texture and stereoscopi effect than ver.1 - LoRA can be applied without a trigger word - The style can be controlled using 3d and realistic tags - When added to Positive Prompt, it enhances the 3D feel - When added to Negative Prompt, it adds details such as clothing while maintaining the model's art style - The issue of eroding the model's art style with tags such as intricate, detailed has been alleviated Additional information The effect of LoRA was confirmed using the following my custom merged models: - - - High-poly 3D Model LoRA ハイポリゴン 3D モデル画像を学習した LoRA です。 3D らしいクリーンで高精細な肌・髪の質感や、衣服のディテールを与えることができます。 ver.2 の学習時に得られた技術的知見をまとめましたので、興味がある方は次の記事をご覧ください。 2.0 での変更点 - 教師画像を増やしました - 学習解像度を向上しました (ver.1: 768, ver.2: 896) - タグ付けを見直しました - トリガーワードも削除しました - 学習パラメータを見直しました 2.0 での改善点 - ver.1 よりも 3D らしいテクスチャ感、奥行き感が再現できます - トリガーワードなしで LoRA を適用することが可能です - 3d, realistic タグを使用することで絵柄のコントロールが可能です - Positive Prompt に入れた場合、より 3D 感を強化します - Negative Prompt に入れた場合、モデルの絵柄を保ちつつ、衣装などのディテールを追加することができます - intricate, detailed などのタグによってモデルの絵柄を侵食してしまう問題が緩和されましたその他 LoRA の動作確認は以下の自作マージモデルを使用して行いました。 - - -

98,352 downloads · 7,557 likes on Civitai

SHA256

3BE24D2E6D581BAEC264ECAA843B0CA45524C95984C576DA641E8A8406F0EC5F

ModelTrends.ai Model ID

#30

Flag content

Example renders

Prompts and settings shared by the people who made these renders on Civitai. Pick a render to see how it was made.

Get this prompt

Render by realitylvn on Civitai

Flag

Settings

Sampler

DPM++ 2S a Karras

Steps

Guidance

7.5

Seed

3363075371

Size

2048x3072

Clip skip

Prompt

(full body, (dynamic pose), action pose), <the subject is a female/male android humanoid, in a flowing torn thick cloth cloak, dark orange metallic armor, wielding an obsidian blade, featuring hard surface designs in their gear>,<the scene is an advanced civilization, with towering structures in the background, immersed in the beautiful lights, the cityscape is a fully detailed environment>, (rim lighting, studio lighting, distant moon light, night, bloom), (cinematic, best quality, masterpiece, ultra HD textures, highly detailed, hyper realistic, intricate detail, 8k, photorealistic, concept art, matte painting, autodesk maya, vray render, ray tracing, hdr), (dslr, full frame, 16mm focal length, f/8 aperture, dynamic perspective, dynamic angle, golden ratio, wide photography, wide field of view, deep depth of field, zoom out) <lora:epiNoiseoffset:0.3> <lora:urbanSamuraiClothing:0.25> techwear, urbansamurai <lora:3dCGStyleRealistic:0.2> <lora:hipoly3DModelLora:0.2>3d, realistic <lora:cyberhelmetWearable:0.45> cyborg in a cyberhelmet head <hypernet:dr0ne:0.15> <hypernet:sxzBloom_sxzBloom:0.2> <hypernet:LuisapSciFiHard:0.1>

Negative prompt

(worst quality:2), (low quality:2), (normal quality:2), (lowres:2), bad anatomy, fat, ugly, cartoon, comic, sketch, anime, monochrome, logo, signature, watermark, deformed, mutation, text, ng_deepnegative_v1_75t, render by bad-artist, bad-hands-5, bad_prompt_version2, EasyNegative, advntr, cheerful, happy, snow, close up, portrait, zoom in, crop in, shallow depth of field, medium depth of field, subject in center, subject close to camera, imperfect circle, imperfect bolts, imperfect screws, open helmet, half helmet, large moon, 3d, realistic, mask, image by bad-artist-anime, bad design, low detail, low texture quality, repetition, old, rusty, 1k, 2k

Render by NeoClassicalRibbon on Civitai

Flag

Settings

Sampler

UniPC

Steps

Guidance

Seed

1041864763

Size

512x896

Clip skip

Prompt

<lora:hipoly3DModelLora_v20:0.6>, 1girl, looking at viewer, upper body, 3D, realistic, large breasts, excessively frilled princess dress, draped clothes, jewelry, ornament, flower, lace trim, masterpiece, best quality, 8k, detailed skin texture, detailed cloth texture, beautiful detailed face, intricate details, ultra detailed, rim lighting, side lighting, cinematic light, ultra high res, 8k uhd, film grain,best shadow, delicate, RAW

Negative prompt

[:EasyNegative:0.5], extra fingers,fewer fingers, cowboy, mole, freckles, facepaint, (simple background:1.3),

Render by realitylvn on Civitai

Flag

Settings

Sampler

Euler a

Steps

Guidance

Seed

2958810892

Size

512x768

Clip skip

Prompt

((best quality)), ((masterpiece:1.2)), (8k, high quality, cinematic, hyper realistic, illustration), (autodesk maya, vray render, ray tracing, hdr), (DSLR, full frame, 16mm focal length, f/4 aperture, dynamic perspective, deep depth of field), (A black and gold, humanoid bounty hunter, clad in a reinforced exoskeleton and adaptive techwear, stalking their prey through the dark, neon-lit alleyways of a cyberpunk city), <lora:mshn:0.35> (mshn robot:0.55, mshn:0.25), <hypernet:dr0ne:0.55> (subdivision, cyber muscle, angular design, mechanical, carbon-fiber, high-tech, pistons, robotic frame, metal joints, rubber cables), <hypernet:LuisapSciFiHard:0.25> (scifi, hard surface, armor, metal bolts, robotic, industrial, structures, sub-d), <lora:urbanSamuraiClothing:0.25> (urbansamurai, techwear), <lora:hipoly3DModelLora:0.25> (CGI, 3D, realistic), <lora:epiNoiseoffset:0.5> (bloom, rim lighting, studio lighting, soft lighting, low key), <lora:cyberhelmetWearable:0.5> (cyberhelmet:0.25, glowing robotic eyes, full of battlegear)

Negative prompt

(low quality, worst quality:1.4), [bad-picture-chill-75v:bad-picture-chill-32v:0.5], (bad_prompt_version2-neg:0.75), ((render by bad-artist:0.5)), 3D, CGI, realistic, ((mask)), (EasyNegative:0.5), ((badhandsv5-neg:0.8)), imperfect bolts, imperfect circle, bad anatomy:0.8, detailed

Render by NeoClassicalRibbon on Civitai

Flag

Settings

Sampler

UniPC

Steps

Guidance

Seed

126391336

Size

512x896

Clip skip

Prompt

<lora:hipoly3DModelLora_v20:0.7>, 1girl, looking at viewer, upper body, 3D, realistic, large breasts, excessively frilled princess dress, draped clothes, jewelry, ornament, flower, lace trim, masterpiece, best quality, 8k, detailed skin texture, detailed cloth texture, beautiful detailed face, intricate details, ultra detailed, rim lighting, side lighting, cinematic light, ultra high res, 8k uhd, film grain,best shadow, delicate, RAW

Negative prompt

[:EasyNegative:0.5], extra fingers,fewer fingers, cowboy, mole, freckles, facepaint, (simple background:1.3),

Render by DDDLSTE on Civitai

Flag

Settings

Size

480x854

Prompt

Scene setup: - A vast, dreamy pink landscape made entirely of infinite cherry-blossom “clouds.” Branches are paper-thin and nearly invisible, hidden behind blossoms so the background reads as a fluffy pink sky. - Gentle pink petals continuously drift down from above, swirling softly in the air. Main character: - A super-cute, fluffy baby kitten with a round face, small body, and oversized sparkling blue eyes. - Its fur is stylized with soft, rounded contours and subtle highlights to emphasize fluffiness. - Petals settle naturally on its head, ears, and back like delicate accessories (no twigs or branches). Animation: - Camera fixed, front view, slightly low angle so the kitten fills the frame as it approaches. - The kitten walks smoothly toward the camera at a relaxed pace, occasionally blinking, tilting its head, or pausing to sniff a falling petal. - No sudden turns or jerky motions—movement should feel fluid and charming. Lighting & Sound: - Soft, diffused lighting with gentle gradients and highlights; pastel color palette dominated by pinks and soft whites. - Ambient soundtrack: light, whimsical music with the faint sound of petals rustling.

Render by realitylvn on Civitai

Flag

Settings

Sampler

Euler a

Steps

Guidance

7.5

Seed

2971199474

Size

2048x3072

Clip skip

Prompt

NOUSR ROBOT:1.1, FULL BODY:1.6 (A black and gold, humanoid bounty hunter, clad in a reinforced exoskeleton and adaptive techwear, stalking their prey through the dark, neon-lit alleyways of a cyberpunk city), <lora:mshn:0.45> (mshn robot:0.55, mshn:0.25), <hypernet:dr0ne:0.85> (subdivision, cyber muscle, angular design, mechanical, carbon-fiber, high-tech, pistons, robotic frame, metal joints, rubber cables), <hypernet:LuisapSciFiHard:0.25> (scifi, hard surface, armor, metal bolts, robotic, industrial, structures, sub-d), <lora:urbanSamuraiClothing:0.55> (urbansamurai, techwear), <lora:hipoly3DModelLora:0.35> (CGI, 3D, realistic), <lora:cyberhelmetWearable:0.35> a cyborg in a cyberhelmet head

Negative prompt

Render by realitylvn on Civitai

Flag

Settings

Sampler

DPM++ 2M alt Karras

Steps

Guidance

Seed

4260550867

Size

2144x2688

Clip skip

Prompt

((best quality)), ((masterpiece:1.2)), (extremely detailed:1.1), (8k, high quality, cinematic, hyper realistic, illustration), (autodesk maya, octane render, unreal engine, game character, ray tracing, hdr), (16mm focal length, f/4 aperture, dynamic perspective, deep depth of field), (1girl, dynamic angle, casting pose. Ice witch, long hair glistening like frost, crystal-woven gown, glacial gloves, frostbite boots. Snowy cloak, underlying icy blue dress, frozen tundra, ice sculptures sparkling around her), <lora:hipoly3DModelLora:0.5> 3d, realistic, <lora:3DcgStyle:0.75> CG, 3D model, <lora:EnvyElegance:0.85> beautiful, elegant, confident, <hypernet:sxzCinematic_800StrongEnough:0.55>, <lora:epiNoiseoffset:0.55> (hdri, bloom, rim lighting, soft lighting, low key), <lora:zhongfenghuaxiyou:0.4> zhongfenghua, <lora:LowRA:0.5>, <lora:delicateArmor:0.2> delicate\(armor\) wizard robe, wizard hat, <lora:blackMageFashion:0.15> mage_glam, caftan

Negative prompt

(low quality:1.4, worst quality:1.8), [bad-picture-chill-75v:bad-picture-chill-32v:0.5], (bad_prompt_version2-neg:0.75), ((render by bad-artist:1.1)), 3D, realistic, bad anatomy:1.1, ng_deepnegative_v1_75t, EasyNegativeV2, verybadimagenegative_v1.3, bad_pictures:0.8, badhandv4:1.2, ugly, missing fingers, missing hands, watercolor, paintings, sketches, anime, illustration, 2d, drawing, low detail, low textures, poorly detailed, distorted hands, mutation, deformation, ((monochrome)), ((grayscale)), normal quality, extra limbs, vile, gross, disgusting, extra arms, extra legs

Render by realitylvn on Civitai

Flag

Settings

Sampler

DPM++ 2S a Karras

Steps

Guidance

Seed

3168837767

Size

512x768

Clip skip

Prompt

((best quality)), ((masterpiece:1.2)), (extremely detailed:1.1), (8k, high quality, cinematic, hyper realistic, illustration) <lora:epiNoiseoffset:0.3> (rim lighting) <lora:urbanSamuraiClothing:0.3> (urbansamurai, techwear) <lora:hipoly3DModelLora:0.25> (CG, 3D, realistic), <lora:cyberhelmetWearable:0.25> (cyberhelmet), Cybernetic samurai, sleek metallic armor with intricate engravings, glowing energy lines coursing through the body, a plasma katana emitting a humming sound, facial features obscured by a high-tech visor. Rain-soaked streets, neon signs reflecting off puddles, bustling crowds wearing futuristic attire, towering holograms advertising the latest technology, an undercurrent of tension and danger <hypernet:dr0ne:0.75> (mechanical design, carbon-fiber, high-tech, exoskeletal) , <lora:mshn:0.1> mshn robot

Negative prompt

(low quality, worst quality:1.4), [bad-picture-chill-75v:bad-picture-chill-32v:0.5], ((render by bad-artist:0.35)), 3D, realistic, ((mask)), tiny details, bolts, screws

Similar Models

Badges for creators

Made this model? Put its standing on your Civitai page, your Hugging Face card, your README or your site. Each badge is a small picture that shows the standing on the day you copy it and never changes afterwards, however the rankings move later. The date is part of the badge.

AI Models for Image, Video & Text Generation

hipoly is a LoRA for the Stable Diffusion 1.5 family listed on ModelTrends.ai, a read-only catalog of open source AI image, video and text models. Compare it with other Stable Diffusion 1.5 models, check its heat score to see how it is trending, and open it on PirateDiffusion or BitVector to try it.

More Stable Diffusion 1.5 models More Stable Diffusion 1.5 LoRAs effects 3d clothing positives girls styles game characters fantasy

Browse by Model Family