Exploring PyTorch Tensor

Short exploration and explanation of PyTorch Tensor function

An short introduction about PyTorch and about the chosen functions.

torch.logspace
torch.argmin
torch.cat
torch.split
torch.mean

# Import torch and other required modules
import torch
import numpy as np
import matplotlib.pyplot as plt

Function 1 - torch.logspace

The torch.logspace function returns a 1-D tensor of steps point logarithmically spaced with base base between base^start and base^end.

Note: torch.linspace is a linear spacing and torch.logspace is logarithmic spacing

# Example 1 - working
torch.logspace(start=-1, end=1, steps=5, dtype=torch.int32)

tensor([ 0,  0,  1,  3, 10], dtype=torch.int32)

This example returns a tensor with 5 integers from 10^-1 to 10¹ with logarithmically spaced. Here, the values are rounded since the desired data type is set to an Integer (torch.int32)