Downloads · 30 days
4
14% of all-time downloads
AriannaV1
SurgeFF/AriannaV1
AriannaV1 is a image-text-to-text model from SurgeFF. Use it for the image-text-to-text task on the model card, and read the license before you ship it in a product. It is set up for transformers. The card lists the license as gemma.
All-in-one local assistant model for Surge's home AI fleet. AriannaV1 is gemma-4-12b-it (multimodal, encoder-free) with the Aria adapter-v17 LoRA merged into the base weights — a single standalone model, not a bare ad…
All-time downloads
28
Public
Parameters
12B
28.6 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors28.5 GB · 100%
At a glance
- Task
- Image-Text-to-Text
- Library
- transformers
- License
- gemma
- Model type
- gemma4_unified
- Access
- Public
- Created
- Aug 9, 2026
- Updated
- Aug 9, 2026
- SHA
- 155ec116
Try a prompt
Base models
Model card
From the Hugging Face model README
AriannaV1
All-in-one local assistant model for Surge's home AI fleet. AriannaV1 is
gemma-4-12b-it (multimodal, encoder-free) with the Aria adapter-v17 LoRA
merged into the base weights — a single standalone model, not a bare adapter.
Architecture
Arianna's weights carry the core: text, reasoning, identity, memory, math, code, grammar, storytelling, tool-selection, safety, and vision-/audio-understanding. Every other modality is a sidecar the core orchestrates, never baked into the weights: realtime voice (STT/TTS), video generation, image generation, and retrieval embeddings. The AgentOS runtime — self-improvement, self-healing, toolsmith, curiosity, semantic memory, append-only event log, reflection — wraps this core as the self-modifying brain.
Gated eval scores (v17, suite fingerprint 2be660e12d96bc71)
| capability | score |
|---|---|
| math (100-item GSM8K held-out) | 0.90 |
| tools | 1.00 |
| identity | 0.90 |
| identity_bare (monitor-only) | 0.625 |
| memory | 0.90 |
Disjoint 150-problem math confirmation: 140/150 = 93.3%.
Files
- Repo root — merged standalone weights (
AutoModelForImageTextToText.from_pretrainedjust works). adapter/— the original Aria v17 LoRA, to stack on the base yourself.
License
Gemma license, inherited from the base model. Built by Sergio Williams / Surge.
- Task
- Image-Text-to-Text
- Library
- transformers
- Type
- gemma4_unified
- License
- gemma
- Created
- Aug 9, 2026
- Updated
- Aug 9, 2026