🌐 Supra-A2A-Nano-Exp

An experimental ~30M-parameter any-to-any multimodal model from SupraLabs. A single autoregressive GPT that reads and writes both text and images through one unified token vocabulary.

⚠️ This is an educational prototype, not a polished product. Expect low-resolution 64×64 images and short text.

Model: SupraLabs/Supra-A2A-Nano-Exp · ~30M params · 384-token context · 64×64 image output · Apache-2.0