---
license: creativeml-openrail-m
tags:
- stable-diffusion
- stable-diffusion-cpp
- gguf
- lcm
- image-generation
---

# DreamShaper 8 LCM GGUF Files

Quantized GGUF weights of DreamShaper 8 LCM, optimized for fast local CPU and low-memory GPU image generation using `stable-diffusion.cpp`.

These models are designed specifically to run inside the **Haven AI Companion** local portrait generation engine.

## Files Available

*   **DreamShaper8_LCM_q8_0.gguf** (1.8 GB) - High-quality 8-bit quantization. Recommended for most systems.
*   **DreamShaper8_LCM_q4_0.gguf** (1.6 GB) - Fast 4-bit quantization for low-memory environments.

## Quick Start with stable-diffusion.cpp

Run the local server:

```bash
sd-server.exe --model ./DreamShaper8_LCM_q8_0.gguf --taesd ./taesd.safetensors --steps 8 --sampling-method lcm --cfg-scale 2.0
```