Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/Akegarasu/lora-scripts
LoRA & Dreambooth training scripts & GUI use kohya-ss's trainer, for diffusion model.
https://github.com/Akegarasu/lora-scripts
dreambooth finetune lora
Last synced: 3 months ago
JSON representation
LoRA & Dreambooth training scripts & GUI use kohya-ss's trainer, for diffusion model.
- Host: GitHub
- URL: https://github.com/Akegarasu/lora-scripts
- Owner: Akegarasu
- License: agpl-3.0
- Created: 2023-01-22T13:50:45.000Z (almost 2 years ago)
- Default Branch: main
- Last Pushed: 2024-07-25T10:15:41.000Z (4 months ago)
- Last Synced: 2024-07-25T11:43:16.133Z (4 months ago)
- Topics: dreambooth, finetune, lora
- Language: Python
- Homepage:
- Size: 445 KB
- Stars: 4,147
- Watchers: 26
- Forks: 525
- Open Issues: 43
-
Metadata Files:
- Readme: README-zh.md
- License: LICENSE
Awesome Lists containing this project
- AiTreasureBox - Akegarasu/lora-scripts - 11-02_4562_0](https://img.shields.io/github/stars/Akegarasu/lora-scripts.svg) |LoRA training scripts use kohya-ss's trainer, for diffusion model.| (Repos)
- awesome - Akegarasu/lora-scripts - LoRA & Dreambooth training scripts & GUI use kohya-ss's trainer, for diffusion model. (Python)
README
# SD-Trainer
_✨ 享受 Stable Diffusion 训练! ✨_
LoRA-scripts(又名 SD-Trainer)
LoRA & Dreambooth 训练图形界面 & 脚本预设 & 一键训练环境,用于 [kohya-ss/sd-scripts](https://github.com/kohya-ss/sd-scripts.git)
## ✨新特性: 训练 WebUI
Stable Diffusion 训练工作台。一切集成于一个 WebUI 中。
按照下面的安装指南安装 GUI,然后运行 `run_gui.ps1`(Windows) 或 `run_gui.sh`(Linux) 来启动 GUI。
![image](https://github.com/Akegarasu/lora-scripts/assets/36563862/d3fcf5ad-fb8f-4e1d-81f9-c903376c19c6)
| Tensorboard | WD 1.4 标签器 | 标签编辑器 |
| ------------ | ------------ | ------------ |
| ![image](https://github.com/Akegarasu/lora-scripts/assets/36563862/b2ac5c36-3edf-43a6-9719-cb00b757fc76) | ![image](https://github.com/Akegarasu/lora-scripts/assets/36563862/9504fad1-7d77-46a7-a68f-91fbbdbc7407) | ![image](https://github.com/Akegarasu/lora-scripts/assets/36563862/4597917b-caa8-4e90-b950-8b01738996f2) |# 使用方法
### 必要依赖
Python 3.10 和 Git
### 克隆带子模块的仓库
```sh
git clone --recurse-submodules https://github.com/Akegarasu/lora-scripts
```## ✨ SD-Trainer GUI
### Windows
#### 安装
运行 `install-cn.ps1` 将自动为您创建虚拟环境并安装必要的依赖。
#### 训练
运行 `run_gui.ps1`,程序将自动打开 [http://127.0.0.1:28000](http://127.0.0.1:28000)
### Linux
#### 安装
运行 `install.bash` 将创建虚拟环境并安装必要的依赖。
#### 训练
运行 `bash run_gui.bash`,程序将自动打开 [http://127.0.0.1:28000](http://127.0.0.1:28000)
## 通过手动运行脚本的传统训练方式
### Windows
#### 安装
运行 `install.ps1` 将自动为您创建虚拟环境并安装必要的依赖。
#### 训练
编辑 `train.ps1`,然后运行它。
### Linux
#### 安装
运行 `install.bash` 将创建虚拟环境并安装必要的依赖。
#### 训练
训练
脚本 `train.sh` **不会** 为您激活虚拟环境。您应该先激活虚拟环境。
```sh
source venv/bin/activate
```编辑 `train.sh`,然后运行它。
#### TensorBoard
运行 `tensorboard.ps1` 将在 http://localhost:6006/ 启动 TensorBoard
## 程序参数
| 参数名称 | 类型 | 默认值 | 描述 |
|------------------------------|-------|--------------|-------------------------------------------------|
| `--host` | str | "127.0.0.1" | 服务器的主机名 |
| `--port` | int | 28000 | 运行服务器的端口 |
| `--listen` | bool | false | 启用服务器的监听模式 |
| `--skip-prepare-environment` | bool | false | 跳过环境准备步骤 |
| `--disable-tensorboard` | bool | false | 禁用 TensorBoard |
| `--disable-tageditor` | bool | false | 禁用标签编辑器 |
| `--tensorboard-host` | str | "127.0.0.1" | 运行 TensorBoard 的主机 |
| `--tensorboard-port` | int | 6006 | 运行 TensorBoard 的端口 |
| `--localization` | str | | 界面的本地化设置 |
| `--dev` | bool | false | 开发者模式,用于禁用某些检查 |