EarthnDusk/SD-Convert-1-5

---
license: creativeml-openrail-m
language:
- en
tags:
- code
pretty_name: Convert SD 1.5 to Diffusers
size_categories:
- n<1K
---

# **SD 1.5 Model Converter**

<a target="_blank" href="https://colab.research.google.com/github/kieranxsomer/convert-scripts/blob/main/Converter_SD1_5_V2_Duct_TapeVersion.ipynb">
<img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/>
</a>

**A Colab Notebook To Convert SD 1.5 Checkpoint to Diffusers format**

But a horribly duct taped edition. THIS IS IN ALPHA STAGES, WILL BE PATCHING THE CODE AS I GO ALONG.

♻ - USE ONLY FOR NOW: Converter_SD1_5_V2_Duct_TapeVersion.ipynb

♻ - THIS IN THEORY SHOULD WORK ON VAST/RUNPOD - BUT IT IS UNTESTED, JUST CHANGE YOUR DIRECTORIES AS NEEDED!

RIGHT NOW THE INSTRUCTIONS ARE AS FOLLOWS:

♻ - Install/Clone etc

♻ - Download model - Direct port from Linaqruf.

♻ - Open code panel, replace model details. - don't move after you hit play, it does it really quickly.

♻ - Check file browser, if the model/yourmodelhere looks like a diffusers format you're good to go!

♻ - Write Token + Set up your Repo!

♻ - Upload Diffusers!

---

***Patched from*** : https://colab.research.google.com/github/Linaqruf/sdxl-model-converter/blob/main/sdxl_model_converter.ipynb

***Linaqruf @ Github***: https://github.com/Linaqruf

 [](https://lookup.guru/850007095775723532) [](https://ko-fi.com/linaqruf) <a href="https://saweria.co/linaqruf"><img alt="Saweria" src="https://img.shields.io/badge/Saweria-7B3F00?style=flat&logo=ko-fi&logoColor=white"/></a>

**Please use their main scripts for SDXL HERE:**

| Notebook Name | Description | Link |
| --- | --- | --- |
| [Kohya LoRA Trainer XL](https://github.com/Linaqruf/kohya-trainer/blob/main/kohya-LoRA-trainer-XL.ipynb) | LoRA Training | [](https://colab.research.google.com/github/Linaqruf/kohya-trainer/blob/main/kohya-LoRA-trainer-XL.ipynb) |
| [Kohya Trainer XL](https://github.com/Linaqruf/kohya-trainer/blob/main/kohya-trainer-XL.ipynb) | Native Training | [](https://colab.research.google.com/github/Linaqruf/kohya-trainer/blob/main/kohya-trainer-XL.ipynb) |

SD 1.5 Scripts:

| Notebook Name | Description | Link | V14 |
| --- | --- | --- | --- |
| [Kohya LoRA Dreambooth](https://github.com/Linaqruf/kohya-trainer/blob/main/kohya-LoRA-dreambooth.ipynb) | LoRA Training (Dreambooth method) | [](https://colab.research.google.com/github/Linaqruf/kohya-trainer/blob/main/kohya-LoRA-dreambooth.ipynb) | [](https://colab.research.google.com/github/Linaqruf/kohya-trainer/blob/ff701379c65380c967cd956e4e9e8f6349563878/kohya-LoRA-dreambooth.ipynb) |
| [Kohya LoRA Fine-Tuning](https://github.com/Linaqruf/kohya-trainer/blob/main/kohya-LoRA-finetuner.ipynb) | LoRA Training (Fine-tune method) | [](https://colab.research.google.com/github/Linaqruf/kohya-trainer/blob/main/kohya-LoRA-finetuner.ipynb) | [](https://colab.research.google.com/github/Linaqruf/kohya-trainer/blob/ff701379c65380c967cd956e4e9e8f6349563878/kohya-LoRA-finetuner.ipynb) |
| [Kohya Trainer](https://github.com/Linaqruf/kohya-trainer/blob/main/kohya-trainer.ipynb) | Native Training | [](https://colab.research.google.com/github/Linaqruf/kohya-trainer/blob/main/kohya-trainer.ipynb) | [](https://colab.research.google.com/github/Linaqruf/kohya-trainer/blob/ff701379c65380c967cd956e4e9e8f6349563878/kohya-trainer.ipynb) |
| [Kohya Dreambooth](https://github.com/Linaqruf/kohya-trainer/blob/main/kohya-dreambooth.ipynb) | Dreambooth Training | [](https://colab.research.google.com/github/Linaqruf/kohya-trainer/blob/main/kohya-dreambooth.ipynb) | [](https://colab.research.google.com/github/Linaqruf/kohya-trainer/blob/ff701379c65380c967cd956e4e9e8f6349563878/kohya-dreambooth.ipynb) |

Ahoy! you're looking for our Huggingface backup that is again patched from Linaqruf and others?

| Notebook Name | Description | Link |
| --- | --- | --- |
| [Huggingface Backup](https://colab.research.google.com/github/kieranxsomer/HuggingFace_Backup/blob/main/HuggingFace_Backup.ipynb) | backup checkpoints! | [](https://colab.research.google.com/github/kieranxsomer/HuggingFace_Backup/blob/main/HuggingFace_Backup.ipynb)
| [1.5 Conversions](https://github.com/kieranxsomer/convert-scripts/blob/main/Converter_SD1_5_V2_Duct_TapeVersion.ipynb) | Convert to Diffusers! | [](https://github.com/kieranxsomer/convert-scripts/blob/main/Converter_SD1_5_V2_Duct_TapeVersion.ipynb)
## Duskfall/ Earth & Dusk Socials

| Social Network | Link |
| --- | --- |
|Discord|[Invite](https://discord.gg/5t2kYxt7An)
|CivitAi|[Duskfallcrew](https://civitai.com/user/duskfallcrew/)
|Huggingface|[Earth & Dusk](https://huggingface.co/EarthnDusk)
|Ko-Fi| [Dusk's Kofi](https://ko-fi.com/duskfallcrew/)

原始信息汇总

SD 1.5 Model Converter

概述

• 名称: SD 1.5 Model Converter
• 描述: 一个Colab笔记本，用于将SD 1.5检查点转换为Diffusers格式。当前版本为Alpha阶段，代码将持续更新。
• 标签: code
• 语言: en
• 大小: n<1K
• 许可证: creativeml-openrail-m

使用说明

1. 安装/克隆等
2. 下载模型 - 直接从Linaqruf移植。
3. 打开代码面板，替换模型细节 - 不要在点击播放后移动，它会非常快速地完成。
4. 检查文件浏览器，如果模型/yourmodelhere看起来像Diffusers格式，则可以继续。
5. 写入Token并设置仓库
6. 上传Diffusers

注意事项

• 当前仅建议使用Converter_SD1_5_V2_Duct_TapeVersion.ipynb。
• 理论上可以在VAST/RUNPOD上运行，但尚未测试，只需根据需要更改目录。

来源

• 从Linaqruf的SDXL模型转换器笔记本中修补而来。
• Linaqruf的GitHub主页: https://github.com/Linaqruf