Tensor Array Python

This machine learning library using Tensor-Array library

This project is still in alpha version, we are trying to make this look like the main framework but it is easier to code.

How to install Tensor-Array python version.

Before install this library please install NVIDIA CUDA toolkit first.

It can not work without NVIDIA CUDA toolkit.

If you did not install Python then install Python 3:

apt-get update
apt-get install python3

After that go to command and install:

pip install TensorArray

Testing with the Tensor object.

The Tensor class is a storage that store value and calculate the tensor.

The Tensor.calc_grad() method can do automatic differentiation.

The Tensor.get_grad() method can get the gradient after call Tensor.calc_grad().

import tensor_array.core as ta
import numpy as np

def test_add():
    example_tensor_array = ta.Tensor(np.array([[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12], [13, 14, 15, 16]], dtype=np.int32))
    example_tensor_array_scalar = ta.Tensor(100)
    example_tensor_sum = example_tensor_array + example_tensor_array_scalar
    print(example_tensor_sum)
    example_tensor_sum.calc_grad()
    print(example_tensor_array.get_grad())
    print(example_tensor_array_scalar.get_grad())

test_add()