Dog X Bloods 狗血

---

The Misadventures of a Misfortunate Dog, developed by WetDogs

丧家犬与落水狗的冒险剧，由落水狗（WETDOGS) 制作

Following the perspective of the protagonist, the player will decide who, what, where, why, and how the story unfolds. This LoRa is capable of generating the four main characters of the story: The Protagonist, Qianbei, The Prince, and The Cop.

跟随主角的视角，玩家将决定故事展开的人物、事件、地点、原因及方式。该LoRa能够生成故事的四个核心角色：主人公、前辈、王子殿下与 Cop。

---

The Protagonist 主人公

Character: dxb_protagonist, tugou

Outfit: white shirt, yellow jacket, black shorts, brown belt, elbow pads, fingerless gloves

Qianbei 前辈

Character: dxb_qianbei, fox

Reinforcement: red eyes, earrings

Outfit: green jacket, black vest, brown pants

The Prince 王子殿下

Character: dxb_prince, husky

Outfit: boxer briefs, transparent raincoat

The Cop

Character: dxb_cop, rottweiler

Outfit: police hat, blue underwear

Two Characters 双主角

• Protagonist with Prince 主人公与王子殿下: protaxprince (recommend to use v0.5)

• Protagonist with Cop 主人公与Cop : protaxcop (recommend to use v1.0)

• Protagonist with Qianbei 主人公与前辈: protaxqian (recommend to use v1.0)

v0.5 - Original dataset, capable of producing The Protagonist and The Prince well. Used as base model for v1.0 dataset. The images in the earlier post were generated using this LoRa.

v0.5 - 原始数据集能够稳定生成主人公与王子殿下角色，已作为v1.0版本的基础模型使用。先前发布的图像正是基于该LoRa生成。

v1.0 - Improved dataset with some images from v0.5 LoRa. Great in producing images of two characters together, as well as characters Qianbei and The Cop compared to v0.5

v1.0 - 改进版数据集融合了v0.5 LoRa的部分图像素材。相较于v0.5版本，本版本在双角色同框生成方面表现优异，且前辈与Cop 角色的生成效果显著提升。

---

Feel free to review, comment, and share your images. If you enjoy this model, consider leaving a tip — it helps me continue creating and sharing more!

And don't forget to support the game developers who brought this amazing works to life!
https://www.dogbloods.com/en

Thank you!

---

Remark:

Sorry for dragging this LoRa out for so longggg... I've been busy with real-life commitments lately. On top of that, with such a complicated LoRa (at least for me) and the limitations of the Civitai trainer, I haven't been able to deliver the quality I originally hoped for. I apologize if this has disappointed anyone who had high expectations for it. Hopefully in the future, I’ll be able to retrain this LoRa once I’ve gained enough knowledge about creating multi-character LoRas and have my own powerful PC to work with.

Anyway, for a single-character LoRa, I recommend checking out the one created by talented trainer 🫶

备注:

最近一直忙于生活琐事，导致这款 LoRa 的发布时间一再拖延…… 相较于我以往制作的 LoRa（无论是单人角色还是双人组合），这次的难度（对我而言）实在太高了。无论是图片数量的要求、Civitai 本身训练功能的限制，还是我第一次尝试制作包含多角色互动的 LoRa，都让我感到力不从心…… 如果这款 LoRa 生成的图像质量未能达到您的预期，在此向大家诚挚道歉！希望将来我拥有自己的训练设备，并对多人 LoRa 有更深入的掌握时，能再接再厉，将这款 LoRa 打磨得更加出色！