pytorch-tensors
Introduction to Pytorch tensor basic operations.
Here there are some jupyter notebook (.ipynb) files that guides you through the basics of pytorch-tensors.
Anyone intrested can clone this repository and run in their local machine in the continous cycle of learning and exploring. I have written these codes with the reference of official pytorch documentation and some tutorials.
You can find the pytorch documentation at : https://pytorch.org/docs/stable/index.html
TABLE OF CONTENTS:
I will provide you the overview of each notebook files so you can choose the best for you.
1. getting_started.ipynb
-> This notebook file contains :
a.Chossing devives (CPU or GPU)
b.Working with tensors
c.Common initialization methods
d.Initializing and converting tensors into different types
e.Numpy array to tensor conversion and vice versa
2.math_operation.ipynb
-> This notebook files contains :
a.Tensor Math and Comparision Operation
-Addition
-Subtraction
-Division ( element wise division)
-Exponentiation
-Inplace operations
-Single Comparision
-Matrix Multiplication
-Matrix Exponentiation
-Element Wise matrix multiplication
-Dot Product
-Batch matrix multiplicaiton
b.Broadcasting Example
3.tensor_indexing.ipynb
-> This notebook file contains:
a.Tensor Indexing
b.Finding the features of first batch
c.Gettin features from every batches
d.Fancy Indexing
e.Advanced Indexing
f.Useful operation and finding dimension
4.tensor_reshaping.ipynb
-> This notebook file contains:
a.Reshaping tensors
b.Transpose
INSTALLATION:
To run these files you should install pytorch in your local environment. I suggest creating virtual environment using Anconda Navigator. Install anaconda navigator from https://docs.anaconda.com/anaconda/install/.
Create virutal enviroment by following the given article : https://medium.com/@chinmay.s077/how-to-create-environment-in-anaconda-982ffc9f3a07
After sucessfully creating the virtual environment run the following command to install pytorch :
For linux OS using CPU in conda package : conda install pytorch torchvision torchaudio cpuonly -c pytorch
If you have different systems then you can choose as per your requirement by visiting https://pytorch.org/.
Thank you. Asim Mahat ( April 14,2021)