Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
model merge extention for stable diffusion web ui
| Date | Stars |
|---|---|
| 2026-08-03 | 835 |
| 2026-08-06 | 835 |
| 2026-08-12 | 837 |
| 2026-08-21 | 837 |
| 2026-08-30 | 837 |
| 2026-09-04 | 835 |
| 2026-09-05 | 835 |
| 2026-09-13 | 835 |
| 2026-09-18 | 835 |
| 2026-09-23 | 835 |
| 2026-09-30 | 835 |
| 2026-10-02 | 834 |
| 2026-10-03 | 835 |
| 2026-10-08 | 836 |
| 2026-10-20 | 837 |
| 2026-10-22 | 838 |
| 2026-10-25 | 837 |
| 2026-11-04 | 837 |
Today
— stars today
This week
— stars this week
This month
+2 stars this month
Momentum
0.0
growth rate 0.00%/day
# SuperMerger
- Model merge extention for [AUTOMATIC1111's stable-diffusion-webui](https://github.com/AUTOMATIC1111/stable-diffusion-webui)
- Merge models can be loaded directly for generation without saving
[<img src="https://img.shields.io/badge/lang-Egnlish-red.svg?style=plastic" height="25" />](#overview)
[<img src="https://img.shields.io/badge/言語-日本語-green.svg?style=plastic" height="25" />](README_ja.md)
[<img src="https://img.shields.io/badge/Support-%E2%99%A5-magenta.svg?logo=github&style=plastic" height="25" />](https://github.com/sponsors/hako-mikan)
# Overview
This extension allows merged models to be loaded as models for image generation without saving them.
This extension can prevent the use of HDD and SSD.
This extension significantly improves the efficiency of model merging by eliminating the need for saving and loading the model and by generating continuously while changing ratios.
# Index
- [Merge Models](#merge-models)
- [Merge Block Weight](#merge-block-weight)
- [XYZ Plot](#xyz-plot)
- [Adjust](#adjust)
- [Let the Dice Roll](#let-the-dice-roll)
- [Elemental Merge](#elemental-merge)
- [Generation Parameters](#generation-parameters)
- [LoRA](#lora)
- [Merge LoRAs](#merge-loras)
- [Merge to Checipoint](#merge-to-checkpoint)
- [Extract from Checkpoints](#extract-from-checkpoints)
- [Extract from two LoRAs](#extract-from-two-loras)
- [Other Tabs](#other-tabs)
- [Calcomode](calcmode_en.md)
- [Elemental Merge](elemental_en.md)
# Recent Update
2026.10.25.0100(JST)
- Support merging and block merging MOSS-Image on Forge Neo
- Fix merging fp16, bf16 and fp32 models with each other (Triple Sum and others) failing on mismatched dtypes (#480)
- Fix the Clear Cache button doing nothing (#474)
- Fix the strength being ignored when a LoRA is merged with both a strength and block weights (#463)
- LoKr files are now recognised as LyCORIS: merging several of them reports that this is not supported instead of stopping with a KeyError (#459)
- Forge NeoのMOSS-Imageのマージと階層マージに対応しました
- fp16/bf16/fp32の混在したモデル同士のマージ(Triple Sumなど)がdtypeの不一致で落ちる問題を修正しました(#480)
- Clear Cacheボタンが動作しない問題を修正しました(#474)
- LoRAを強度とブロック重みを両方指定してマージした場合に強度が無視されていた問題を修正しました(#463)
- LoKrがLyCORISとして認識されるようになりました。複数のLoKrのマージは非対応である旨を表示し、KeyErrorで止まらなくなりました(#459)
2024.02.02 0000(JST)
- bug fix
- MakeLoRaの標識が変わりました。Dreanboothに習ってモデル名をTunedとOriginalに変更しています。
- The model name for MakeLoRA has changed. Following to the Dreambooth, the model names have been changed to Tuned and Original.
2023.11.18.0000(JST)
- add new option:use cuda
If you have enough GPU memory, merge speed will be 50x.
- 新しいオプション(use cuda)が追加されました
GPUメモリに余裕がある場合マージスピードを50倍以上にできます。
2023.11.12
- change LoRA extract script: support SDXL,2.X
- support LoRA merge to checkpoint in SD2.X
- add new feature: Include/Exclude
- Bugfix
- Add new calcmode `extract(checkpoint/LoRA)`, Thanks [subaqua](https://github.com/sbq0)!
## Change in Calculation Functions
In several modes, the functions used for calculations have been changed, resulting in faster merge computation speeds.
The calculation results should be the same, but if you cannot get the same result, please check the "use old calc method" in the options. The methods affected are as follows:
Weight Sum: normal, cosineA, cosineB
Sum Twice:normal
I am grateful to [wkpark](https://github.com/wkpark) for proposing this method.
**Attention!**
A minimum of 64GB of CPU memory is required for the XL model merge. Even with 64GB of memory, depending on the software you are using in conjunction, there is a risk of system instability. Therefore, please work in a state where it is acceptable for the system to crash. I encountered a blue screen for the first time in a while.
**注意！**
XLモデルのマージには最低64GBのCPUメモリが必要です。64Gのメモリであっても併用しているソフトによってはシステムが不安定になる恐れがあるのでシステムが落ちてもいい状態で作業して下さい。私は久しぶりにブルースクリーンに遭遇しました。
## Knowon issues/既知の問題
If you have other eExcerpt of 25,915 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:b69a8325b2df511b, desc:stable diffusion
