Infrastructure

AirLLM lowers peak accelerator memory by loading model layers sequentially instead of keeping the full checkpoint resident during inference.

Cache-to-Cache learns projections that fuse a source model's KV cache into a receiver, letting paired language models exchange latent state without generating intermediate text.

FastAPI-MCP exposes selected FastAPI operations as MCP tools while reusing request schemas, endpoint documentation, ASGI integration, and existing dependency-based authentication.

KV Cache Transform Coding applies learned decorrelation, adaptive quantization, and entropy coding to reduce storage and transfer costs for reusable language-model KV caches.

OmniExtractBench provides a common structured-document extraction dataset, runner, and deterministic scorer whose verdicts distinguish matches, omissions, misreads, and fabricated values.