Heterogeneous programming in Julia

Light-weight expression manipulation tools

-

Primitives for construction of objects

Julia package for various special functions based on `log` and `exp`.

CUDA programming in Julia.

The implementation of NameResolution.jl for Julia language.

A package for your code to understand scoping rules(including mutability of free variables).

AD-backend agnostic system defining custom forward and reverse mode rules. This is the light weight core to allow you to define rules for your functions in your packages, without depending on any particular AD system.

-

-

Canonical default initial values and identity elements for Julia

One More Einsum for Julia! With runtime order-specification and high-level adjoints for AD

A data-focused namespace for packages to share functions

Pretty print that makes sense

Optimisers.jl defines many standard optimisers and utilities for learning loops.

Types and operations for basis represented by bits in linear algebra.

Julia implementation of Random123.

C-compatible enum for Julia

Print data in formatted tables.

Batched routines (BLAS, LAPACK, etc.) for multi-dimensional arrays

Forward and reverse mode automatic differentiation primitives for Julia Base + StdLibs

Julia implementation of zlib decompression

Julia implementation for the BFloat16 number type

Mike's Little Intermediate Representation

Julia functional programming infrastructures and metaprogramming facilities

21st century AD

An interface for tables in Julia

Julia implementation of associative containers that preserve insertion order

Fast parsing machinery for basic types in Julia

Efficient implementation of struct arrays in Julia

The Arnoldi Method with Krylov-Schur restart, natively in Julia.

-

Update deeply nested immutable structs.

Julia package for lazily representing matrices filled with a single entry

Traits for julia iterators

Type stable methods for small tuples

A package which provides an API for querying differentiation results at multiple orders simultaneously

A simple shared suite of common derivative definitions

Neural Network primitives with multiple backends