Instructions to use PrismLive/Expressive-LoRA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use PrismLive/Expressive-LoRA with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir Expressive-LoRA PrismLive/Expressive-LoRA
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Expressive LoRA
The Expressive style for Prism Image Studio (macOS / iOS): the Master Drawing LoRA, a subtle hand-drawn look layered on top of PrismImage.
| File | Contents |
|---|---|
master-drawing.safetensors |
Master Drawing LoRA, rank 32, 8-bit row-wise (int8 weights + f16 per-row scales) |
Prism Image Studio applies it at strength 0.5 whenever the style switch is set to Expressive; Standard runs PrismImage on its own. The 8-bit row-wise format is read by the app's own loader, not by generic LoRA tooling.
The app downloads this file on first launch together with PrismImage; nothing here needs to be fetched by hand.
Hardware compatibility
Log In to add your hardware
Quantized