A PyTorch architecture mapping LLM Attention and KV-Caches into non-Euclidean hyperbolic space to drastically reduce memory bottlenecks.
machine-learning ai pytorch hyperbolic-geometry kv-cache large-language-models htf hyperbolic-tensor-folding attentian-mechanism
-
Updated
Jul 24, 2026 - Python