Converting python list to pytorch tensor

Question

I have a problem converting a python list of numbers to pytorch Tensor : this is my code :

caption_feat = [int(x)  if x < 11660  else 3 for x in caption_feat]

printing caption_feat gives : [1, 9903, 7876, 9971, 2770, 2435, 10441, 9370, 2]
I do the converting like this : tmp2 = torch.Tensor(caption_feat) now printing tmp2 gives : tensor([1.0000e+00, 9.9030e+03, 7.8760e+03, 9.9710e+03, 2.7700e+03, 2.4350e+03, 1.0441e+04, 9.3700e+03, 2.0000e+00])
However I expected to get : tensor([1. , 9903, , 9971. ......]) Any Idea?

It's just the scientific notation, numbers are ok. However, you should probably use torch.tensor() for construction of tensors. — dedObed
– dedObed, Commented Feb 6, 2020 at 8:05

Devanshi · Accepted Answer · 2021-01-10 02:45:19Z

20

You can directly convert python list to a pytorch Tensor by defining the dtype. For example,

import torch

a_list = [3,23,53,32,53] 
a_tensor = torch.Tensor(a_list)
print(a_tensor.int())

>>> tensor([3,23,53,32,53])

answered Jan 10, 2021 at 2:45

Devanshi

2112 silver badges5 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Brian Cannard Over a year ago

It says "only one element tensors can be converted to Python scalars"

Dishin Goyani · Accepted Answer · 2020-02-06 08:34:16Z

6

If all elements are integer you can make integer torch tensor by defining dtype

>>> a_list = [1, 9903, 7876, 9971, 2770, 2435, 10441, 9370, 2]
>>> tmp2 = torch.tensor(a_list, dtype=torch.int)
>>> tmp2
tensor([    1,  9903,  7876,  9971,  2770,  2435, 10441,  9370,     2],
       dtype=torch.int32)

While torch.Tensor returns torch.float32 which made it to print number in scientific notation

>>> tmp2 = torch.Tensor(a_list)
>>> tmp2
tensor([1.0000e+00, 9.9030e+03, 7.8760e+03, 9.9710e+03, 2.7700e+03, 2.4350e+03,
        1.0441e+04, 9.3700e+03, 2.0000e+00])
>>> tmp2.dtype
torch.float32

edited Feb 6, 2020 at 8:34

answered Feb 6, 2020 at 8:18

Dishin Goyani

7,7533 gold badges33 silver badges42 bronze badges

Comments

GarAust89 · Accepted Answer · 2020-04-29 19:18:03Z

1

A simple option is to convert your list to a numpy array, specify the dtype you want and call torch.from_numpy on your new array.

Toy example:

some_list = [1, 10, 100, 9999, 99999]
tensor = torch.from_numpy(np.array(some_list, dtype=np.int))

Another option as others have suggested is to specify the type when you create the tensor:

torch.tensor(some_list, dtype=torch.int)

Both should work fine.

answered Apr 29, 2020 at 19:18

GarAust89

4572 silver badges10 bronze badges

Comments

ychnh · Accepted Answer · 2020-02-06 08:52:54Z

0

Try

torch.IntTensor(caption_feat)

You can see the other types here https://pytorch.org/docs/stable/tensors.html

answered Feb 6, 2020 at 8:52

ychnh

1971 silver badge12 bronze badges

Collectives™ on Stack Overflow

Converting python list to pytorch tensor

4 Answers 4

1 Comment

Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

1 Comment

Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related