ppw_v8_Illuv1_128.safetensors

972 MB · SafeTensor

Vyžaduje členství

<peoples-works-illus-xl>

v10.5 - 本次更新没有添加新功能。 This update does not introduce any new features. - 添加了一个使用较大参数DiT模型生成的数据集，略微加强cowboy shot 和 full body构图下的解剖稳定性。人物头部过大的几率降低了。 Added a dataset generated using larger DiT models, slightly improving anatomical stability in cowboy shot and full-body compositions. The probability of overly large heads has also been reduced. - 使用nano banana和seedream修复了一部分“因为瑕疵面积过大而难以手工修复，在过去几个版本被移除数据集”的图片。 Used Nano Banana and Seedream to repair some images that had previously been removed from the dataset because their defects were too extensive to fix manually. - 已适配Anima。 Now compatible with Anima. 快速上手 | Quick Start （v10.5 已更新） 这是什么？ | What is this? - People's Works 是一个免费的入门级插画模型系列，专注于中、近景风格化人像生成。这个模型的数据集由数千张AI社区用户发布的图片和作者使用AI合成的图片构成，经过人工编辑和标注后用于训练。除此之外还有一个由真实照片、游戏CG构成的辅助数据集。 - People's Works is a free, entry-level illustration model series specializing in stylized portrait generation, primarily focused on medium and close-up compositions. The dataset consists of several thousand images published by AI community users, along with images synthesized by the author using AI. These images were manually edited and annotated before being used for training. In addition, there is an auxiliary dataset composed of real photographs and game CGs. 模型功能 | Model features - 风格化：使用 ppw 触发一种介于写实和漫画之间、采集自 的特定风格，并且对 flat color, no lineart, realistic 等风格/绘画肌理tag进行优化。 - 质量与审美提示词：对 masterpiece, very aesthetic, best quality 三个modifier进行了优化。 - 人物精细化控制：在训练集中增加了对人物的族裔、年龄的训练。增加生成人像的多样性。 - 使用高清训练素材，支持1024 - 1536分辨率生成。 - Stylization: Use ppw to trigger a distinctive style between realism and anime illustration originate from . The model is also optimized for specific style and texture tags, including flat color, no lineart, and realistic. - Quality & Aesthetic Modifier: Optimized for following modifier tags: masterpiece, very aesthetic, best quality . - Detailed Character Control: The training dataset includes increased coverage of ethnicity and age, improving the diversity of generated portraits. - Trained with high-resolution images and supports generation at resolutions between 1024 - 1536. 使用方法 | Usage v10+ 快速开始：直接将这行tag放在prompt开头： Quick Start: Place these tags at the beginning of your prompt: ppw, masterpiece, best quality, very aesthetic (v10.5 updated) 以下是所有已训练概念的列表 | Below is a list of all trained concepts: 触发风格 | trigger stylization: ppw, 肌理 | texture tags: flat color, no lineart, realistic, 质量与审美提示词 | Quality & Aesthetic Modifier: masterpiece, best quality, very aesthetic, 年龄控制 | Age Control: child, teenage, adult, mature, 族裔 | Ethnicity: Anime, Asian, Caucasian, African, Indian, v7-v9 positive: masterpiece, best quality, very aesthetic negative: low quality, displeasing 更新记录 | Change log v10 注意：这个版本尝试额外添加了一个触发词“ppw”。把它放在提示词的开始处来触发lora。 Note: This version experimentally adds an extra trigger word: “ppw”. Place it at the beginning of your prompt to trigger the LoRA. - 降低了写实照片的训练轮数，尽管对于族裔标签的响应效果弱于v9，但是提升了生成质量。 In this version, I reduced the number of training epochs for the photo dataset. Although the response to ethnicity tags is weaker than in v9, the overall generation quality has improved. - 新增一个族裔标签Anime。使用它可以让人物获得更接近传统二次元的脸型。 Added a new ethnicity tag Anime. Using this tag can give characters a face shape closer to traditional anime. - 移除了质量提示词“low quality”的训练数据集。 The training datasets for the quality prompt tags "low quality" have been removed. - 对过去的数据集进行了维护。降低了学习率，增加训练步数，缓解了一些人物解剖错误、肢体扭曲的问题。 The version includes a maintenance of the previous dataset. The learning rate was reduced and the number of training steps was increased, which helps alleviate some anatomy errors and limb distortion issues. v9 - 更改了系列名称。自这个版本起，训练集中来自Pony v6 XL的图片已经在所有AIGC内容中占比不足1/3。随着越来越多的新模型出现，我有计划在明年将这个系列拓展到其他模型上。为了避免未来用户使用上的混乱和误解，这个系列从这个版本起更改命名。 The series name has been changed. Starting from this version, images sourced from Pony v6 XL make up less than one third of the training data across all AIGC content. As increasing number of new models are emerging, I plan to expand this series to other models next year. To avoid potential confusion and misunderstanding for users in the future, the series name has been changed starting from this version. - 这个版本的训练方式是直接训练LoCon，而非训练Checkpoint后再抽取LoRA。模型相较前一个版本效果更强。 This version is trained directly as a LoCon, rather than training a checkpoint first and then extracting a LoRA. Compared to the previous versions, the model delivers stronger effects. - v9全系列使用1536分辨率的训练集。现在使用这个Lora生成图片时支持单边768-1536的分辨率，使用高清修复时也可以尝试更高的denoise参数了。 All v9 models use a 1536-resolution training set. When generating images with this LoRA, single-side resolutions from 768 to 1536 are now supported. When using high-res fix, you can also try higher denoise values. - 对训练图片调色。现在模型在没有指定色彩时更倾向于生成暖色调的图片，并且色彩的饱和度略微提高。较暗的场景明暗对比更大了。 Color adjustments were applied to images. When no specific color is specified, the model now tends to produce warmer tones, with slightly increased saturation. Darker scenes also have stronger contrast between light and shadow. - 之前版本的数据集中，人物鼻子的画法不统一。出于作者本人的兴趣，手工修改了其中约300幅图片，并暂时排除了约200幅来不及修改的图片。现在人物的鼻子有鼻翼了。 In earlier versions of dataset, nose depiction was inconsistent. Out of personal interest, the author manually modified around 300 images and temporarily excluded about 200 images that could not be edited in time. Characters’ noses now have nose wings. - 删除了旧版本中数百张过时的低质量训练数据。 Hundreds of outdated, low-quality images from older versions of dataset have been removed. - 添加了一个新的实验性数据集： 使用真人相片作为引导，现在你可以使用以下年龄和族裔标签了： A new experimental dataset has been added. Using real photographs as guidance, you can now use the following age and ethnicity tags: child, teenage, adult, mature Caucasian, Asian, Indian, African - 我的标签设计优先选择Danbooru数据集中已经存在的标签，尽管其中很多只有很少量的数据，在原版模型中几乎无法触发。重新启用了已经被删除的danbooru词条Caucasian和teenage，增设了adult和African两个标签。此外，loli和shota因为其文化背景中强烈的性暗示倾向，这两个词条被完全替换，根据具体情况分流入child和teenage。 My tag design prioritizes labels that already exist in

SHA256

1B5B763D83961BB5745F3AF8271BA83F1D4FD69C16278DAE6D5B4E194BDDE97A

ID modelu ModelTrends.ai

#15796

Sampler

Euler a Karras

Kroky

26

Guidance

6

Seed

169189539948130

Velikost

832x1216

Sampler

Euler a

Kroky

19

Guidance

5

Seed

1765271748

Velikost

832x1216

Clip skip

2

Použitá váha

1

Sampler

Euler a

Kroky

19

Guidance

5

Seed

144026216

Velikost

832x1216

Clip skip

2

Použitá váha

1

Sampler

DPM++ 2M Karras

Kroky

25

Guidance

7

Seed

772081301561909

Velikost

1664x2432

Sampler

Euler a

Kroky

27

Guidance

4.5

Seed

2007059945

Velikost

672x1176

Clip skip

2

Sampler

Euler a

Kroky

19

Guidance

6

Seed

2088205673

Velikost

832x1216

Clip skip

2

Použitá váha

0.5