AriannaV4
Production all-in-one local assistant for Surge's home AI fleet.
AriannaV4 = google/gemma-4-12B + the Aria v4 weights merged — one standalone, multimodal core. The published gguf/ quant ladder is provided for llama.cpp / Ollama / LM Studio.
Architecture
The weights carry the core: text, reasoning, identity, memory, math, code, grammar,
storytelling, tool-selection, safety, and vision-/audio-/video-understanding. Every other
modality is a sidecar the core orchestrates — realtime voice (Whisper STT / Piper TTS),
video generation, image generation, retrieval embeddings — never baked into the weights.
Files
- Repo root — merged standalone weights (AutoModelForImageTextToText.from_pretrained).
- gguf/ — GGUF quant ladder (F16 + Q2_K … Q8_0) for llama.cpp / Ollama / LM Studio.
License
Built by Sergio Williams / Surge.
