Z-Image Asian Utop 亚洲美女

Z-Image Asian Utopian 亚洲美女是一款基于ZImageTurbo、ZImageBase架构深度训练的AI绘画大模型。它精准捕捉东方女性的古典韵味与现代美感，无论是灵动的眉眼线条、温婉的神态气质，还是适配古风、潮流等多元风格的造型，都能高效生成细节饱满、氛围感拉满的优质画面，满足创作者对亚洲美女形象的多样化创作需求。

关于v2.0 ZImageBase版本

与v2.0 ZImageTurbo一致，均使用2.4MP高解析度数据集训练fp32，dim 64 Lora，并重新调整v1.5 - ZSVD融合配重，重新制作为fp32全精度版本，转换为bf16，GGUF。

fp8 的损耗太大；GGUF Q8_0 会更好

推荐VAE：diffusion_pytorch_model.safetensors（包含下载）

下载包含bf16、fp8和gguf三个版本，以及工作流

作品参照

A young woman with brown hair in twin buns, adorned with a blue and pink flower hair clip, rests her arms on a red pool edge. She wears a white bikini top, gazing thoughtfully away. Water droplets cling to her skin. Soft, diffused daylight illuminates the tranquil, turquoise pool water.
Negative prompt: low query, blur, bad anatomy, bad hands, missing fingers, extra fingers, bad arms, missing legs, missing arms, poorly drawn face, bad face, fused face, cloned face, three crus, fused feet, fused thigh, extra crus, ugly fingers, horn, cartoon, cg, 3d, unreal, animate, amputation, disconnected limbs
Steps: 30, CFG scale: 4, Sampler: er_sde, Seed: 342796270213760, width: 768, height: 1344

图片提示词

A highly detailed photorealistic portrait of a stunning Japanese cosplayer dressed as a Holstein cow girl under bright sunny weather with a perfectly clear deep blue sky, solo female, natural long black hair styled in twintails, wearing a black non-woven face mask covering mouth and nose, cute puffy tear bags under her eyes, direct eye contact looking straight at viewer, realistic skindentation from clothing, slim thin waist and slender legs, overall slim and elegant body, thick leather collar with large cowbell around neck, white cow horn and cow ear headband, frilly black-and-white cow print bikini top, oversized white knit long-sleeve bolero slightly slipping off shoulders, high-waisted red and black plaid ruffle mini skirt, black nail polish, exposed navel, one hand making a cute peace sign V gesture, the other hand holding a classic japanese glass milk bottle, standing in a lush green ranch field with grazing Holstein cows softly in the background, brilliant clear blue sky with no clouds, golden hour natural sunlight pouring down, realistic skin texture with subtle pores and details, live-action cosplay photography style, sharp focus, cinematic depth of field, 85mm lens, f/1.8 aperture, ultra-detailed, photorealism, masterpiece quality
Negative prompt: low query, jpeg artifacts, blur, bad anatomy, bad hands, missing fingers, extra fingers, bad arms, missing legs, missing arms, poorly drawn face, bad face, fused face, cloned face, three crus, fused feet, fused thigh, extra crus, ugly fingers, horn, cartoon, cg, 3d, unreal, animate, amputation, disconnected limbs
Steps: 30, CFG scale: 2.6, Sampler: res_2m, Seed: 45, width: [object Object], height: [object Object]

Z-Image Asian Utopian 亚洲美女是一款基于ZImageTurbo架构深度训练的AI绘画大模型。它精准捕捉东方女性的古典韵味与现代美感，无论是灵动的眉眼线条、温婉的神态气质，还是适配古风、潮流等多元风格的造型，都能高效生成细节饱满、氛围感拉满的优质画面，满足创作者对亚洲美女形象的多样化创作需求。

关于v2.0版本

FF = 全片幅，全精度

使用2.4MP高解析度数据集训练fp32，dim 64 Lora，并重新调整v1.5 - ZSVD融合配重，重新制作为fp32全精度版本，转换为bf16，GGUF。

推荐VAE：diffusion_pytorch_model.safetensors（包含下载）

下载包含bf16、fp8和gguf三个版本，以及工作流

作品参照

图片提示词

图片提示词

图片提示词

Z-Image Asian Utopian 亚洲美女是一款基于ZImageTurbo架构深度训练的AI绘画大模型。它精准捕捉东方女性的古典韵味与现代美感，无论是灵动的眉眼线条、温婉的神态气质，还是适配古风、潮流等多元风格的造型，都能高效生成细节饱满、氛围感拉满的优质画面，满足创作者对亚洲美女形象的多样化创作需求。

关于v1.5版本

额外使用 1000 张高画质训练 dim 128 Lora 并 SVD 融合[Z Image Turbo] Asian Mix Lora v3.78与原版 Z Image Turbo bf16。

推荐VAE：diffusion_pytorch_model.safetensors（包含下载）

下载包含工作流

作品参照

the photo of woman pose, realistic, 1girl, solo, glasses, bed, long hair, legs, brown eyes, feet, necklace, black hair, brown hair, skirt, cleavage, no shoes, bare shoulders, lipstick, black pantyhose, underwear, on side, in the image, a woman is lying on a bed, wearing black stockings and a white outfit. she is positioned in a way that suggests she is posing for a photo. the image appears to be a close-up of the woman, capturing her outfit and the bed she is lying on. it could be a promotional or artistic image, showcasing the woman's attire or the bed's design. the image does not depict any specific event or action, but rather focuses on the woman's appearance and the setting.
Steps: 9, CFG scale: 1, Sampler: er_sde, Seed: 564631875756868, width: 768, height: 1344

(Smiling, mischievous,serious,normal.smirk,composed face-displaying any expressions.),(Ultra-realistic photos in 4K resolution,cinematic lighting,highly detailed ultra realistic photo of A stunningly beautiful/cute 28 years old Japanese actress without a single flaw,detailed ski,detailed/Accentuated  droopy eyes,Emphasized small/oval face,Emphasized Snow-white skin,Emphasized narrow, sharp-angled jaw,Emphasized thin, sharp eyebrows, slender, pointed eyelashes,a small, prominent, high-bridged nose,Long eyelashes),(Height 173cm, Body accentuated by beautiful curve,emphasized medium round breasts and emphasized slim waist, large hips, long slender arms and legs), ((posing like a fashion model's seductive pose),looking away,full body,View the whole body,Light from the left illuminates the body, casting a shadow behind it.:1.3), (low braided ponytail over one shoulder with blunt cut),(wearing light makeup,orange eyeshadow),(light grey white trim with written "Z-Image" sports bra under white unbuttoned blouse shoulder off,light grey white trim with written "Z-Image" sports thong under lowrise destroyed denim pants zipper is undoned ,pull down pants with both hands,transparent heel sandals)
Steps: 9, CFG scale: 1, Sampler: er_sde, Seed: 824095699356026, width: 768, height: 1344

Create a photorealistic image of a young Japanese woman inside a crowded public bus, capturing a vibrant urban atmosphere. The subject is seated on a metallic bench, and the shot is composed using a low-angle perspective that looks up at her, creating slight foreshortening. Focus the composition on her legs and torso, with her knees pulled up to her chest. She holds an iPhone in her left hand as if taking a selfie, while her gaze is directed away to the right, accompanied by a devious, confident smile.

The woman has long, flowing black hair and is dressed in a luxurious pink satin robe that flows elegantly, left open to reveal a black lace bra and garter belt, paired with sophisticated black thigh-high stockings. The contrast between the soft, glossy satin and intricate lace textures, as well as the bold pink and black color palette, should be emphasized.

The mood of the image should exude confidence, intimacy, and a subtle sense of rebellion, blending the private nature of her pose with the gritty, bustling public setting. Use cinematic lighting to create dynamic highlights and shadows, ensuring sharp focus and high detail throughout. Incorporate a depth of field effect to enhance the three-dimensional quality of the image, drawing attention to the subject while subtly blurring the crowded background to maintain context without distraction.
Steps: 9, CFG scale: 1, Sampler: er_sde, Seed: 1004684992480169, width: 768, height: 1344

Z-Image Asian Utopian 亚洲美女是一款基于ZImageTurbo架构深度训练的AI绘画大模型。它精准捕捉东方女性的古典韵味与现代美感，无论是灵动的眉眼线条、温婉的神态气质，还是适配古风、潮流等多元风格的造型，都能高效生成细节饱满、氛围感拉满的优质画面，满足创作者对亚洲美女形象的多样化创作需求。

这是使用了[Z Image Turbo] Asian Mix Lora v3.78与原版Z Image Turbo bf16，另外加入了10k数据集（我拥有所有使用权，其实，都是我自己拍摄的风景照），使用
gradient_accumulation: 4额外训练的Lora来调配权重，并根据AI-Toolkit的结果中，将DiT根据30个不同的Blocks各别调整成的模型。

作品参照

One girl, portrayed in a charming, festive style. She has long, straight dark hair cascading smoothly down her back, intricately adorned with numerous flower-themed accessories woven throughout, adding elegance and softness to her appearance. She wears delicate long lace elbow-length gloves with fine, intricate patterns, enhancing a refined and graceful aesthetic. On her feet are cute Mary Jane shoes, polished and neatly fitted, each decorated with dainty bows that add a sweet, youthful touch.

She holds an ornate, decorative sign with both hands, designed in a festive, elegant style, clearly displaying the words “Merry Christmas” in beautifully styled lettering. The sign appears fancy and well-crafted, possibly with subtle embellishments such as gold trim, ribbons, or delicate patterns. Her posture is gentle and composed, with a warm, inviting expression that conveys holiday cheer.

The overall mood is festive and cozy, evoking a classic Christmas atmosphere. Soft, pleasing lighting highlights her outfit details and accessories, while maintaining a clean, harmonious color palette. The composition feels balanced and visually appealing, focusing on her festive attire, floral hair decorations, and the holiday message she proudly presents.
Steps: 8, CFG scale: 1, Sampler: er_sde, Seed: 200004093593613, width: 768, height: 1344

Create a photorealistic image of a young Japanese woman inside a crowded public bus, capturing a vibrant urban atmosphere. The subject is seated on a metallic bench, and the shot is composed using a low-angle perspective that looks up at her, creating slight foreshortening. Focus the composition on her legs and torso, with her knees pulled up to her chest. She holds an iPhone in her left hand as if taking a selfie, while her gaze is directed away to the right, accompanied by a devious, confident smile.

The woman has long, flowing black hair and is dressed in a luxurious pink satin robe that flows elegantly, left open to reveal a black lace bra and garter belt, paired with sophisticated black thigh-high stockings. The contrast between the soft, glossy satin and intricate lace textures, as well as the bold pink and black color palette, should be emphasized.

The mood of the image should exude confidence, intimacy, and a subtle sense of rebellion, blending the private nature of her pose with the gritty, bustling public setting. Use cinematic lighting to create dynamic highlights and shadows, ensuring sharp focus and high detail throughout. Incorporate a depth of field effect to enhance the three-dimensional quality of the image, drawing attention to the subject while subtly blurring the crowded background to maintain context without distraction.
Steps: 8, CFG scale: 1, Sampler: er_sde, Seed: 44095192612708, width: 768, height: 1344

(Smiling, mischievous,serious,normal.smirk,composed face-displaying any expressions.),(Ultra-realistic photos in 4K resolution,cinematic lighting,highly detailed ultra realistic photo of A stunningly beautiful/cute 28 years old Japanese actress without a single flaw,detailed ski,detailed/Accentuated  droopy eyes,Emphasized small/oval face,Emphasized Snow-white skin,Emphasized narrow, sharp-angled jaw,Emphasized thin, sharp eyebrows, slender, pointed eyelashes,a small, prominent, high-bridged nose,Long eyelashes),(Height 173cm, Body accentuated by beautiful curve,emphasized medium round breasts and emphasized slim waist, large hips, long slender arms and legs), ((posing like a fashion model's seductive pose),looking away,full body,View the whole body,Light from the left illuminates the body, casting a shadow behind it.:1.3), (low braided ponytail over one shoulder with blunt cut),(wearing light makeup,orange eyeshadow),(light grey white trim with written "Z-Image" sports bra under white unbuttoned blouse shoulder off,light grey white trim with written "Z-Image" sports thong under lowrise destroyed denim pants zipper is undoned ,pull down pants with both hands,transparent heel sandals)
Steps: 8, CFG scale: 1, Sampler: er_sde, Seed: 162315931213032, width: 768, height: 1344