RuntimeError when converting image to tensor

[ariangilesgarcia]

Hi @amdegroot , I was trying to get the demo running and I'm having a problem when calling transform(img) of the BaseTransform class.

When doing python test.py the output is the following.

Finished loading model!
Testing image 1/4952....
Traceback (most recent call last):
  File "test.py", line 84, in <module>
    thresh=args.visual_threshold)
  File "test.py", line 39, in test_net
    x = Variable(transform(img).unsqueeze(0))
  File "/home/arian/Documents/proyecto-integrador/models/ssd/ssd-pytorch/data/data_augment.py", line 119, in __call__
    return torch.Tensor(img)
RuntimeError: tried to construct a tensor from a nested float sequence, but found an item of type numpy.float32 at index (0, 0, 0)

This happens in the demo notebook and in the test.py file.
Do you have any idea why this could be happening?.

Thanks,
Arian.

[ariangilesgarcia]

Ok sorry for leaving an issue before trying everything I could to fix it.
I don't use torch so I had no idea what to do, but here's how I fixed it just in case.

On data_augment.py I had to chage return torch.Tensor(img) to return torch.FloatTensor(img).

Then an error appear saying it was expecting a CPU tensor and got a CUDA tensor instead, so I had to add .cuda() to the input variables (I don't know if there is another way to do it, I don't really know how to use pytorch yet). For example, changing x = Variable(transform(img).unsqueeze(0)) to x = Variable(transform(img).unsqueeze(0).cuda())

Again, sorry for the issue!

[amdegroot]

Ah, thanks for pointing that out, it must only be the case on GPU. For some reason I was using torch.Tensor() instead of torch.from_numpy() in data_augment.py. It should all be fixed now. Feel free to send a Pull Request if you find any other errors!