Research language for array processing in the Haskell/ML family (by google-research)

Dex-lang Alternatives

Similar projects and alternatives to dex-lang

  • GitHub repo Pytorch

    Tensors and Dynamic neural networks in Python with strong GPU acceleration

  • GitHub repo jax

    Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more

  • GitHub repo julia

    The Julia Programming Language

  • GitHub repo tutorials

    PyTorch tutorials. (by pytorch)

  • GitHub repo futhark

    :boom::computer::boom: A data-parallel functional programming language

  • GitHub repo FlexFlow

    A distributed deep learning framework that supports flexible parallelization strategies.

  • GitHub repo CIPs

  • GitHub repo tensor_annotations

    Annotating tensor shapes using Python types

  • GitHub repo TablaM

    The practical relational programing language for data-oriented applications

NOTE: The number of mentions on this list indicates mentions on common posts. Hence, a higher number means a better dex-lang alternative or higher similarity.


Posts where dex-lang has been mentioned. We have used some of these posts to build our list of alternatives and similar projects - the last one was on 2021-06-10.
  • Questions for Charles Hoskinson - post from Lex Fridman | 2021-06-10
    Good luck parsing Dex to distinguish from confusion with Decentralized Exchanges (DEX). | 2021-06-10
    So now we are seeing ML research tools like Dex, a Haskell-based research language developed to have support provided by the features of a functional language.
  • Matrix Multiplication Inches Closer to Mythic Goal | 2021-03-23
    Not that it has bindings to other tools, but it sounds like Dex[1] would be relevant to your interests!


  • Dex: A research language for array processing in the Haskell/ML family | 2021-03-11
  • Looking to get into Parallel Computing and System Programming Research | 2021-03-11
  • Dex is a research language for typed, functional array processing | 2021-03-11
  • [N] PyTorch 1.8 Release, including Compiler and Distributed Training updates, New Mobile Tutorials
    When I've interacted with this kind of thing before -- for example the PyTorch JIT (which I believe also parses an AST to produce an IR? Is this the same parser/IR as fx or different?), the JAX JIT, or something like Zygote in Julia -- I've always hit these kinds of issues. I think the fundamental problem is choosing to parse an AST for something inherently more flexible, rather than building a graph-based DSL a la TensorFlow v1 (despite its flaws), or though I've not tried it, maybe building something like dex.
  • [News] Swift for TensorFlow Final Report
    Glad it's dead, it was a failure from the moment Swift was selected because it was Chris Lattner's pet project. Dex is an experimental project from Google that is a lot more promising:
  • A comparison of Futhark and Dex
    You may well be right. It's possible I was deceived by Dex declaring types using the same binding constructs as terms, but that might just be a semi-syntactic trick. It does look like a parameter of term-level type (n) is being promoted to the type level, though.


Basic dex-lang repo stats
1 day ago

google-research/dex-lang is an open source project licensed under BSD 3-clause "New" or "Revised" License which is an OSI approved license.