aditya-khant / neural-assembly-compiler

A neural assembly compiler for pyTorch based on adaptive-neural-compilation

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Neural Assembly Compiler

This repository contains an assembly to PyTorch compiler to be used with ANC. Please follow the instructions given below for the same.

Install

Haskell

The simplest solution to get the dependency would be to install Haskell Platform. Alternatively, if you want to get directly a binary of the compiler, you can find it on the project page

Compilation

Running cabal install should create the executable in ./dist/build/neulang-compiler/neulang-compiler

Usage

The input program should be provided in stdin to the executable. For example to compile the dijkstra's implementation, run cat tests/dijkstra.nl | ./dist/build/neulang-compiler/neulang-compiler. This will print the compiled program to stdout and create a configuration file ./dram.py containing this program that can be used directly by the ANC code.

References

@article{anc, title={Adaptive Neural Compilation}, author={Bunel, Rudy and Desmaison, Alban and Kohli, Pushmeet and Torr, Philip H.S and Kumar, M. Pawan}, journal={arXiv preprint arXiv:1605.07969}, year={2016} }

About

A neural assembly compiler for pyTorch based on adaptive-neural-compilation

License:GNU General Public License v3.0


Languages

Language:Haskell 45.0%Language:NewLisp 29.1%Language:Yacc 12.4%Language:Logos 5.9%Language:Lua 4.7%Language:Python 2.9%