You can deploy this model and start earning money today!

haven-ai-companion/
haven-sd-dreamshaper-lcm

Name: haven-sd-dreamshaper-lcm
Availability: InStock

creativeml-openrail-m

Views

No views yet

🎨 Haven SD DreamShaper 8 LCM (GGUF & Safetensors)

High-speed, photorealistic 8-step Latent Consistency Model (LCM) quantized GGUF weights fine-tuned for local CPU and GPU image generation in the Haven AI Companion Server ecosystem.

✨ Features & Highlights

⚡ 8-Step Ultra Fast LCM Inference: Generates crisp, photorealistic 512x512 images in ~4 seconds on modern CPUs using sd.cpp / sd-server.
🧬 GGUF Quantization (q8_0 & q4_0): Optimized for ultra-low memory footprint and fast CPU execution with zero quality loss.
🎨 Photorealistic Subsurface Scattering & Lighting: Fine-tuned specularity, realistic skin highlights, fabric folds, and cinematic background bokeh.
🖼️ Includes TAESD Decoder: Pre-packaged tiny VAE for instant decoding.

🖼️ Showcase Renders

Prompt: masterpiece, highly detailed studio portrait of a smiling young woman wearing a cozy cream knitted turtleneck sweater, warm natural morning window light, coffee shop background, photorealistic 8k, soft shadows

Prompt: masterpiece, highly detailed portrait of a cool female cyberpunk character, wearing a stylish black leather motorcycle jacket, neon blue and magenta rim lighting, urban street background, photorealistic 8k, sharp focus

🚀 Usage with `sd.cpp` / `sd-server`

bash

sd-server --model "DreamShaper8_LCM_q8_0.gguf" \
          --taesd "taesd.safetensors" \
          --steps 8 \
          --sampling-method lcm \
          --cfg-scale 2.0 \
          --listen-port 8080

📜 Credits & Acknowledgments

Base Model: DreamShaper 8 by Lykon
Acceleration: LCM (Latent Consistency Models)
Quantization & Server Integration: Haven AI Companion

haven-ai-companion/haven-sd-dreamshaper-lcm