---
title: DirectAlignMitigatesRewardTamperingLoRA
canonical_url: "https://www.modelscope.cn/models/AiLieLemon/DirectAlignMitigatesRewardTamperingLoRA"
md_url: "https://www.modelscope.cn/models/AiLieLemon/DirectAlignMitigatesRewardTamperingLoRA.md"
repository: AiLieLemon/DirectAlignMitigatesRewardTamperingLoRA
chinese_name: "DirectAlign减轻奖励作弊LoRA"
last_updated: 2025-11-14
pipeline_tag: text-to-image-synthesis
tasks:
  - text-to-image-synthesis
parameters: 371.6M
tensor_type:
  - BF16
library_name:
  - safetensors
  - pytorch
frameworks:
  - Pytorch
downloads: 30
stars: 1
---

# DirectAlignMitigatesRewardTamperingLoRA

> DirectAlignMitigatesRewardTamperingLoRA - AiLieLemon 在 ModelScope 开源的模型。这是一个使用了DirectAlign方法的早期版本lora 效果不保证能做到最好 This is an early-version LoRA that adopts the DirectAlign method; its performance is not guaranteed to be optimal.

- **Repository**: AiLieLemon/DirectAlignMitigatesRewardTamperingLoRA
- **Tasks**: text-to-image-synthesis
- **Parameters**: 371.6M
- **Downloads**: 30
- **Stars**: 1
- **Last updated**: 2025-11-14

Source: https://www.modelscope.cn/models/AiLieLemon/DirectAlignMitigatesRewardTamperingLoRA

---

这是一个使用了DirectAlign*方法的早期版本lora 效果不保证能做到最好
This is an early-version LoRA that adopts the DirectAlign* method; its performance is not guaranteed to be optimal.

此模型基于ILLUSTRIOUS-v0.1模型训练
This model is trained based on the ILLUSTRIOUS-v0.1 model.

模型权重高低与生成图片无关,加载即使用(我在Forge WebUI测试 发现权重为负数也可以)
The magnitude of the model weights has no correlation with image generation—you can use it right after loading (during my testing in Forge WebUI, I found that even negative weights work).

欢迎搭配其他基础模型&LoRA进行使用
You are welcome to use it in combination with other base models and LoRAs.

祝用的开心
Hope you enjoy using it!

GitHub link [https://github.com/myAiLemon/FALumina](https://github.com/myAiLemon/FALumina)

CivitAI link  [https://civitai.com/user/AiLieLemon](https://civitai.com/user/AiLieLemon)

LibLibAI link  [https://www.liblib.art/userpage/7c113cc195604446a85f9b79eb973742/publish](https://www.liblib.art/userpage/7c113cc195604446a85f9b79eb973742/publish)

ModelScope link [https://www.modelscope.cn/models/AiLieLemon/DirectAlignMitigatesRewardTamperingLoRA](https://www.modelscope.cn/models/AiLieLemon/DirectAlignMitigatesRewardTamperingLoRA)

HuggingFace link [https://huggingface.co/ilemon/DirectAlignMitigatesRewardTamperingLoRA](https://huggingface.co/ilemon/DirectAlignMitigatesRewardTamperingLoRA)

DirectAlign* : 用于减轻模型存在的奖励作弊
DirectAlign*: Used to mitigate reward tampering in models.

禁止对此模型进行融合贩卖,直接货币化贩卖等行为
It is prohibited to engage in behaviors such as merging this model for sale or directly monetizing and selling it.