How to Create Tensor from List in PyTorch: Simple Guide

To create a tensor from a list in PyTorch, use torch.tensor() and pass your list as the argument. This converts the list into a PyTorch tensor that you can use for machine learning tasks.

Syntax

The basic syntax to create a tensor from a list in PyTorch is:

• torch.tensor(data, dtype=None, device=None)

Here, data is your Python list or nested lists. dtype is optional and sets the data type (like float or int). device is optional and sets where the tensor lives (CPU or GPU).

python

import torch

# Create a tensor from a simple list
my_list = [1, 2, 3]
tensor = torch.tensor(my_list)
print(tensor)

Output

tensor([1, 2, 3])

Example

This example shows how to create a 2D tensor from a nested list and specify the data type as float.

python

import torch

nested_list = [[1, 2], [3, 4]]
tensor_2d = torch.tensor(nested_list, dtype=torch.float32)
print(tensor_2d)
print('Shape:', tensor_2d.shape)
print('Data type:', tensor_2d.dtype)

Output

tensor([[1., 2.],
        [3., 4.]])
Shape: torch.Size([2, 2])
Data type: torch.float32

Common Pitfalls

Common mistakes when creating tensors from lists include:

• Passing lists with mixed data types, which can cause unexpected tensor types.
• Not specifying dtype when you need a specific type, leading to default types that may not fit your needs.
• Trying to create tensors from irregular nested lists (lists of different lengths), which will raise errors.

Example of wrong and right ways:

python

import torch

# Wrong: irregular nested list
try:
    bad_list = [[1, 2], [3]]
    bad_tensor = torch.tensor(bad_list)
except Exception as e:
    print('Error:', e)

# Right: regular nested list
good_list = [[1, 2], [3, 4]]
good_tensor = torch.tensor(good_list)
print(good_tensor)

Output

Error: expected sequence of length 2 at dim 1 (got 1)
tensor([[1, 2],
        [3, 4]])

Quick Reference

Tips for creating tensors from lists in PyTorch:

• Use torch.tensor() to convert lists to tensors.
• Specify dtype to control the tensor's data type.
• Ensure nested lists are regular (all inner lists have the same length).
• Use device to place tensors on CPU or GPU.

Key Takeaways

Use torch.tensor() to convert Python lists into PyTorch tensors easily.

Specify dtype to control the data type of the tensor for your needs.

Nested lists must be regular; irregular shapes cause errors.

You can place tensors on CPU or GPU by setting the device parameter.

Always check tensor shape and type after creation to avoid bugs.