🪹 Bird's Nest — Non-Transformer AI Hub for Mac. Local AI chat, image generation, and tools. No cloud, no API keys.
A world model that replaces attention and recurrence with reaction-diffusion PDEs. Latent states evolve through Laplacian diffusion and learned reaction terms on a 2D spatial grid, giving the model a built-in inductive bias for spatial coherence, adaptive computation depth, and O(N) scaling in the number of spatial tokens.
FluidLM is a Transformer-free language model replacing O(N²) self-attention with reaction-diffusion PDEs - achieving O(N) scaling, adaptive computation, and no KV-cache.
Aletheion State Models: an attention-free causal state-model research family derived from DRM, evaluated through ablations and scaling laws.
Hierarchical planning via multi-scale reaction-diffusion PDEs.
Backend-neutral kernel of predictive primitives. O(n) attention is deception.
FluidVLA is Transformer-free Vision-Language-Action model for real-time edge robotics. - Replacing O(N²) self-attention with Reaction-Diffusion PDEs.
To associate your repository with the non-transformer topic, visit your repo's landing page and select "manage topics."