monster/Qwen-Image-Layered-MRP-MLX
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Small README update (#244)

Browse Source
This commit is contained in:
Filip Strand 2025-08-03 11:27:30 +02:00 committed by GitHub
parent beee9b9509
commit 4f9af80956
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 17 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
CHANGELOG.md
View File

@ -12,6 +12,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
### 🎨 Model Improvements
- **FLUX.1 Krea [dev] Support!**
- **FLUX.1-Krea-dev-mflux-4bit Model**: Added [filipstrand/FLUX.1-Krea-dev-mflux-4bit](https://huggingface.co/filipstrand/FLUX.1-Krea-dev-mflux-4bit) quantized model to HF
- **FLUX.1-Kontext-dev-mflux-4bit Model**: Added [akx/FLUX.1-Kontext-dev-mflux-4bit](https://huggingface.co/akx/FLUX.1-Kontext-dev-mflux-4bit) quantized model to HF, contributed by @akx
### ✨ New Features
@ -31,6 +33,12 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
### 👩‍💻 Developer Experience
- **Streamlined Release Process**: Removed TestPyPi publishing step from release workflow for simplified deployment
### 🙏 Contributors
- **[@filipstrand](https://github.com/filipstrand)** - FLUX.1 Krea [dev] model support, 5-bit quantization, enhanced defaults, and various improvements
- **[@akx](https://github.com/akx)** - Added 4-bit quantized Kontext model to HF
---
## [0.9.6] - 2025-07-20

18
README.md
View File

@ -750,6 +750,7 @@ The model sizes for both `schnell` and `dev` at various quantization levels are
|--------|--------|---------|---------|---------|-------------------|
| 7.52GB | 9.61GB | 11.71GB | 13.81GB | 18.01GB | 33.73GB |
*4bit quantized model files are [available for direct download on Hugging Face](#-loading-and-running-a-quantized-version-from-disk), eliminating the need to download full models and quantize them yourself.*
#### 💾 Saving a quantized version to disk
@ -808,21 +809,20 @@ In other words, you can reclaim the 34GB diskspace (per model) by deleting the f
- [dhairyashil/FLUX.1-schnell-mflux-v0.6.2-4bit](https://huggingface.co/dhairyashil/FLUX.1-schnell-mflux-v0.6.2-4bit)
- [dhairyashil/FLUX.1-dev-mflux-4bit](https://huggingface.co/dhairyashil/FLUX.1-dev-mflux-4bit)
- [akx/FLUX.1-Kontext-dev-mflux-4bit](https://huggingface.co/akx/FLUX.1-Kontext-dev-mflux-4bit)
- [filipstrand/FLUX.1-Krea-dev-mflux-4bit](https://huggingface.co/filipstrand/FLUX.1-Krea-dev-mflux-4bit)
<details>
<summary>Using the community model support, the quantized weights can be also be automatically downloaded</summary>
Using the [community model support](#-third-party-huggingface-model-support), the quantized weights can be also be automatically downloaded when running the generate command:
```sh
mflux-generate \
--model "dhairyashil/FLUX.1-schnell-mflux-v0.6.2-4bit" \
--base-model schnell \
--steps 2 \
--seed 2 \
--prompt "Luxury food photograph"
--model filipstrand/FLUX.1-Krea-dev-mflux-4bit \
--base-model krea-dev \
--prompt "A photo of a dog" \
--steps 25 \
--seed 2674888
```
</details>
---
### 💽 Running a non-quantized model directly from disk