Embedding on cuda fails

[attardi]

Issue description

Calling an Embedding moved to cuda fails.
The same code on CPU works.

Code example

import torch
import torch.nn as nn
e = nn.Embedding(17,50)
dev = torch.device('cuda')
e.to(dev, torch.float)
idx=torch.tensor([1,2,3])
e(idx)
Traceback (most recent call last):
File "", line 1, in
File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/module.py", line 491, in call
result = self.forward(*input, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/sparse.py", line 108, in forward
self.norm_type, self.scale_grad_by_freq, self.sparse)
File "/usr/local/lib/python3.5/dist-packages/torch/nn/functional.py", line 107 6, in embedding
return torch.embedding(weight, input, padding_idx, scale_grad_by_freq, spars e)
RuntimeError: Expected object of type torch.cuda.LongTensor but found type torch .LongTensor for argument #3 'index'

System Info

PyTorch version: 0.4.0
Is debug build: No
CUDA used to build PyTorch: 9.1.85

OS: Ubuntu 16.04.4 LTS
GCC version: (Ubuntu 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609
CMake version: version 3.5.1

Python version: 3.5
Is CUDA available: Yes
CUDA runtime version: 8.0.61
GPU models and configuration:
GPU 0: Tesla P100-PCIE-16GB
GPU 1: Tesla P100-PCIE-16GB
GPU 2: Tesla P100-PCIE-16GB
GPU 3: Tesla P100-PCIE-16GB

Nvidia driver version: 390.30
cuDNN version: Probably one of the following:
/usr/lib/x86_64-linux-gnu/libcudnn.so.6.0.21
/usr/lib/x86_64-linux-gnu/libcudnn.so.7.0.5
/usr/lib/x86_64-linux-gnu/libcudnn_static_v7.a
/usr/local/MATLAB/R2017a/bin/glnxa64/libcudnn.so.5.1.5

Versions of relevant libraries:
[pip3] numpy (1.14.0)
[pip3] torch (0.4.0)
[pip3] torchvision (0.2.1)

[apaszke]

The error states that it expects a torch.cuda.LongTensor, while you gave it a torch.LongTensor (a CPU array). Move the inputs to the GPU using the .cuda() method and the error should go away.

[attardi]

I must be missing something:

import torch
import torch.nn as nn
e = nn.Embedding(7,5)
e.cuda()
Embedding(7, 5)
idx=torch.tensor([1,2,3])
idx.cuda()
tensor([ 1, 2, 3], device='cuda:0')
e(idx)
Traceback (most recent call last):
File "", line 1, in
File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/module.py", line 491, in call
result = self.forward(*input, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/sparse.py", line 108, in forward
self.norm_type, self.scale_grad_by_freq, self.sparse)
File "/usr/local/lib/python3.5/dist-packages/torch/nn/functional.py", line 1076, in embedding
return torch.embedding(weight, input, padding_idx, scale_grad_by_freq, sparse)
RuntimeError: Expected object of type torch.cuda.LongTensor but found type torch.LongTensor for argument #3 'index'

[apaszke]

e.cuda() modifies the actual embedding module (it's expensive to clone modules). idx.cuda() returns a new tensor, but idx stays on the CPU. Change it to idx = idx.cuda() and it will work fine.

[attardi]

Thanks.