D3g3n41 – Visual Thinking Model (Decoder-only)
Experimental decoder-only image generation model.
- 4-layer attention transformer
- Quantized int8
- Designed for low-spec GPU inference
- Native image output via latent diffusion (DALL·E-style)
This model is part of the Degencode initiative: a local-first AI architecture combining symbolic, visual, and text reasoning.
No base model. No dependencies. Fully original.
