Torcyx

Torcyx is a extension of Cytnx Tensor network library with ML capability base on pytorch. 

API documentation:

https://kaihsin.github.io/Torcyx/docs/html/index.html

Install:

• Without CUDA, CPU only. python 3.6/3.7/3.8

    $conda install -c kaihsinwu torcyx

• With CUDA. cuda 10.2 only. python 3.6/3.7/3.8

   ** currently unavailable **

Simple demo:

• C++ API:

    #include "torcyx.hpp"
    #include <torch/torch.h>
    #include <iostream>

    using namespace std;
    int main(int argc, char* argv[]){
       auto A = torch::ones({3,4});
       auto cA = torcyx::CyTensor(A,1);

       cA.print_diagram();
       cout << cA << endl;

    } 

• Python API:

    import torcyx
    import torch
    
    A = torch.ones([3,4])
    cA = torcyx.CyTensor(A,1)
    cA.print_diagram()
    print(cA)

• Output:

    --------CyTensor-------
    tensor Name : 
    tensor Rank : 2
    block_form  : false
    is_diag     : False
    TensorOptions(dtype=float, device=cpu, layout=Strided, requires_grad=false (default), pinned_memory=false (default), memory_format=(nullopt))
                -------------      
               /             \     
         0 ____| 3         4 |____ 1  
               \             /     
                -------------      
    Tensor name: 
    braket_form : False
    is_diag    : False
     1  1  1  1
     1  1  1  1
     1  1  1  1
    [ CPUFloatType{3,4} ]

For Contributors:

To build Torcyx, there are two dependency:
    
* 1. pytorch
* 2. cytnx 


Step:

1) install pytorch and cytnx
2) change the -DCYTNX_PATH to the path of installed cytnx directory. 
3) create a build folder:
    $mkdir build
    $cd build
    
4) copy compile.sh into build folder
5) execute compile.sh, and make
    $sh compile.sh
    $make 

Developer:

Kai-Hsin Wu: kaihsinwu@gmail.com