Core tensor-based key-value storage with HNSW indexing and tiered caching
An AI-native embedded database
Versatile Storage Format
Provides functions to read and write safetensors which aim to be safer than their PyTorch counterpart. The format is 8 bytes which is an unsized int, being the size of a JSON header, the JSON header refers the `dtype` the `shape` and `data_offsets` which are the offsets for the values in the rest of the file.
Rust DSL for writing Apple Metal GPU kernels — write once, run fast on Apple Silicon
Quantum Database with vector store, MPS compression, Grover search, and fault-tolerant storage
Forward-mode automatic differentiation with const-generic, stack-allocated first-, second-, and third-order jets (Jet1/Jet2/Jet3) for gradients, Hessians, and third-order tensors
Rust wrappers for the PyTorch C++ api (libtorch).
Rust wrappers for the PyTorch C++ api (libtorch).
AXON — the formal cognitive language: a deterministic, proof-carrying AI runtime. Native Rust lexer/parser/type-checker/IR generator (re-exported from axon-frontend) plus the runtime: typed channels (π-calculus mobility, capability extrusion), algebraic effects via Free Monad CPS handlers, lease kernel + reconcile loop, the Epistemic Security Kernel, Trust Types, Proof-Carrying Code (independently verifiable proof objects), and the closed-catalog extension mechanism. Crate publishes as `axon-lang`; library import is `use axon::*` so existing call sites keep working unchanged.
Data types for the RelayRL framework.
Ergonomic auto differentiation in Rust, with pytorch like apis.