# Construct new tensor on correct device based on input **URL:** \ **Category:** Uncategorized\ **Created:** [February 13, 2017, 2:54am UTC](https://discuss.pytorch.org/t/construct-new-tensor-on-correct-device-based-on-input/469 "2017-02-13T02:54:51Z")\ **Posts on this page:** 9\ **Page:** 1
**Author:** ![ypxie](https://discuss.pytorch.org/letter_avatar_proxy/v4/letter/y/43a26b/32.png) [@ypxie](https://discuss.pytorch.org/u/ypxie)\ **Post date:** [February 13, 2017, 2:54am UTC](https://discuss.pytorch.org/t/construct-new-tensor-on-correct-device-based-on-input/469/1 "2017-02-13T02:54:51Z")
I know `input.new()` will create a new tensor of the same dtype to `input`. But what I want to do is to create a new tensor on the same device to input, but has different data types. for example, input is `torch.cuda.FloatTensor` but I wanna something like `torch.cuda.LongTensor` Is there any way to do it without `if_else` wrt `use_cuda` ? ---
**Author:** ![apaszke](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/apaszke/32/21_2.png) [@apaszke](https://discuss.pytorch.org/u/apaszke)\ **Post date:** [February 13, 2017, 12:40pm UTC](https://discuss.pytorch.org/t/construct-new-tensor-on-correct-device-based-on-input/469/2 "2017-02-13T12:40:33Z")
`input.new().long()` should do it. Type casts retain the device, and if you give no arguments to `new` the tensor isn’t going to have any memory allocated, so the cast is nearly free. ---
**Author:** ![wouter](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/wouter/32/30625_2.png) [@wouter](https://discuss.pytorch.org/u/wouter)\ **Post date:** [January 8, 2018, 1:01pm UTC](https://discuss.pytorch.org/t/construct-new-tensor-on-correct-device-based-on-input/469/3 "2018-01-08T13:01:38Z")
Hi! I’m sorry to reply to an old post but I thought replying here would be better than starting a new topic. What if you do give arguments to `new`, for instance to create 0’s or from a numpy list of data? I found some ways but I’m not sure what’s the best way to do it. So let’s say we have given `X`: ``` device = 1 X = torch.rand(10).cuda(device) # or X = torch.rand(10) ``` How to make code below agnostic to whether X is on CPU/GPU and which device? ``` # Fill with numpy data data = np.array([1, 2, 3]) # For example indices calculated by custom Python procedure X.new().long().new(data) # Works on CPU, error on GPU? X.new().long().new(*data.shape).copy_(torch.from_numpy(data)) # Works but verbose X.new(data.astype(float)).long() # Inefficient, lossy? # X.new(data) # Error # Or fill with zeros X.new(10).long().zero_() # Works, but inefficient? X.new().long().new(10).zero_() # Works, but verbose? ``` ---
**Author:** ![smth](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/smth/32/13_2.png) [@smth](https://discuss.pytorch.org/u/smth)\ **Post date:** [January 8, 2018, 2:17pm UTC](https://discuss.pytorch.org/t/construct-new-tensor-on-correct-device-based-on-input/469/4 "2018-01-08T14:17:59Z")
@wouter your post is unrelated to the parent topic (parent topic is about GPU tensors) ---
**Author:** ![wouter](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/wouter/32/30625_2.png) [@wouter](https://discuss.pytorch.org/u/wouter)\ **Post date:** [January 10, 2018, 8:58am UTC](https://discuss.pytorch.org/t/construct-new-tensor-on-correct-device-based-on-input/469/5 "2018-01-10T08:58:57Z")
@smth I’m sorry if it was not clear. Actually I meant that `X` can be on any device, then how to construct a Tensor with numpy data or 0’s on that same device. I edited my post, please let me know if I should open a new topic for this. ---
**Author:** ![bgobbi](https://discuss.pytorch.org/letter_avatar_proxy/v4/letter/b/fbc32d/32.png) [@bgobbi](https://discuss.pytorch.org/u/bgobbi)\ **Post date:** [August 16, 2018, 1:38pm UTC](https://discuss.pytorch.org/t/construct-new-tensor-on-correct-device-based-on-input/469/6 "2018-08-16T13:38:26Z")
The new() method is deprecated. > What is the new alternative for this use case? ---
**Author:** ![shlomia](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/shlomia/32/25585_2.png) [@shlomia](https://discuss.pytorch.org/u/shlomia)\ **Post date:** [June 23, 2020, 6:49am UTC](https://discuss.pytorch.org/t/construct-new-tensor-on-correct-device-based-on-input/469/7 "2020-06-23T06:49:39Z")
new\_tensor = new\_tensor.to(input.device) will change new tensor to be cuda if needed. ---
**Author:** ![Noam\_Salomonski](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/noam_salomonski/32/9359_2.png) [@Noam\_Salomonski](https://discuss.pytorch.org/u/Noam_Salomonski)\ **Post date:** [January 19, 2021, 3:33pm UTC](https://discuss.pytorch.org/t/construct-new-tensor-on-correct-device-based-on-input/469/8 "2021-01-19T15:33:35Z")
> new\_tensor = new\_tensor.to(input.device) > will change new tensor to be cuda if needed. This creates very ugly (and slow) code such as ``` if std.is_cuda: eps = torch.FloatTensor(std.size()).cuda().normal_() else: eps = torch.FloatTensor(std.size()).normal_() ``` instead of the much better `eps = std.new().normal_()` Isn’t there a better way? ---
**Author:** ![Noam\_Salomonski](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/noam_salomonski/32/9359_2.png) [@Noam\_Salomonski](https://discuss.pytorch.org/u/Noam_Salomonski)\ **Post date:** [October 16, 2022, 9:22am UTC](https://discuss.pytorch.org/t/construct-new-tensor-on-correct-device-based-on-input/469/9 "2022-10-16T09:22:34Z")
[answer to my question](https://stackoverflow.com/a/65794995/913098)