Can't convert cuda:0 device type tensor to numpy. Use Tensor.cpu()

Question

My code in colab worked in January and now it doesn't. It's a Style Transfer program using pytorch (https://github.com/leongatys/PytorchNeuralStyleTransfer). This is the relevant snippet:

max_iter = 500
show_iter = 50
optimizer = optim.LBFGS([opt_img]);
n_iter=[0]

while n_iter[0] <= max_iter:

def closure():
    optimizer.zero_grad()
    out = vgg(opt_img, loss_layers)
    layer_losses = [weights[a] * loss_fns[a](A, targets[a]) for a,A in enumerate(out)]
    loss = sum(layer_losses)
    loss.backward()
    n_iter[0]+=1
    if n_iter[0]%show_iter == (show_iter-1):
        print('Iteration: %d, loss: %f'%(n_iter[0]+1, loss.item()))
enumerate(layer_losses)])
    return loss

optimizer.step(closure)

out_img = postp(opt_img.data[0].cpu().squeeze())
imshow(out_img)
gcf().set_size_inches(10,10)

I get an error as follows:

---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
<ipython-input-11-79d2e162c3f3> in <module>()
     20         return loss
     21 
---> 22     optimizer.step(closure)
     23 
     24 #display result

7 frames
/usr/local/lib/python3.7/dist-packages/torch/optim/optimizer.py in wrapper(*args, **kwargs)
     86                 profile_name = "Optimizer.step#{}.step".format(obj.__class__.__name__)
     87                 with torch.autograd.profiler.record_function(profile_name):
---> 88                     return func(*args, **kwargs)
     89             return wrapper
     90 

/usr/local/lib/python3.7/dist-packages/torch/autograd/grad_mode.py in decorate_context(*args, **kwargs)
     25         def decorate_context(*args, **kwargs):
     26             with self.clone():
---> 27                 return func(*args, **kwargs)
     28         return cast(F, decorate_context)
     29 

/usr/local/lib/python3.7/dist-packages/torch/optim/lbfgs.py in step(self, closure)
    309 
    310         # evaluate initial f(x) and df/dx
--> 311         orig_loss = closure()
    312         loss = float(orig_loss)
    313         current_evals = 1

/usr/local/lib/python3.7/dist-packages/torch/autograd/grad_mode.py in decorate_context(*args, **kwargs)
     25         def decorate_context(*args, **kwargs):
     26             with self.clone():
---> 27                 return func(*args, **kwargs)
     28         return cast(F, decorate_context)
     29 

<ipython-input-11-79d2e162c3f3> in closure()
     11         out = vgg(opt_img, loss_layers)
     12         layer_losses = [weights[a] * loss_fns[a](A, targets[a]) for a,A in enumerate(out)]
---> 13         loss = sum(layer_losses)
     14         loss.backward()
     15         n_iter[0]+=1

<__array_function__ internals> in sum(*args, **kwargs)

/usr/local/lib/python3.7/dist-packages/numpy/core/fromnumeric.py in sum(a, axis, dtype, out, keepdims, initial, where)
   2258 
   2259     return _wrapreduction(a, np.add, 'sum', axis, dtype, out, keepdims=keepdims,
-> 2260                           initial=initial, where=where)
   2261 
   2262 

/usr/local/lib/python3.7/dist-packages/numpy/core/fromnumeric.py in _wrapreduction(obj, ufunc, method, axis, dtype, out, **kwargs)
     84                 return reduction(axis=axis, out=out, **passkwargs)
     85 
---> 86     return ufunc.reduce(obj, axis, dtype, out, **passkwargs)
     87 
     88 

/usr/local/lib/python3.7/dist-packages/torch/_tensor.py in __array__(self, dtype)
    730             return handle_torch_function(Tensor.__array__, (self,), self, dtype=dtype)
    731         if dtype is None:
--> 732             return self.numpy()
    733         else:
    734             return self.numpy().astype(dtype, copy=False)

RuntimeError: Can't call numpy() on Tensor that requires grad. Use tensor.detach().numpy() instead.
    
    TypeError: can't convert cuda:0 device type tensor to numpy. 

I also previously got:

Use Tensor.cpu() to copy the tensor to host memory first.

I understand that the tensor cannot be converted to numpy. However, I've tried adding .cpu() to various positions in the code, but where is the tensor?

I've also tried modifying _tensor.py in dist-packages/torch/ with no luck. And changing the sum to torch.sum

Is it possible that this is related to this issue:

https://github.com/pytorch/pytorch/issues/44023

I do not have the GPU power to run this on a local machine. Thanks!

UPDATE: I've created a workaround by rolling back the versions of torch, torchvision and numpy in Colab. It's the interaction of the latest versions of numpy and pytorch that is causing the error.