torch.onnx.export failes due to missing __round__ function

🐛 Describe the bug

When exporting a trained CNN to onnx file format using torch.onnx.export(), an exception is thrown from torchvision/transforms/functional.py and the model is not saved.

To Reproduce

Steps to reproduce the behavior:

1. Define a CNN-torch.Model instance similar to a VGG-Network, train it
2. Try to export the trained model to an *.onnx - file using the save_model() function below

from os import makedirs, join
import torch
import torch.onnx as onnx

def save_model(model, batch_size, image_size, save_dir, input_channels, tensor_type):
    makedirs(save_dir, exist_ok=True)

    torch.save(model.state_dict(), join(save_dir, f"model.pt"))

    x = torch.randn(batch_size, len(input_channels), *image_size, requires_grad=True)  # a bogus input tensor; needed for tracing onnx export
    x = x.type(tensor_type) # move to device

    onnx.export(model,                                          # model being run
                x,                                              # model input (or a tuple for multiple inputs)
                join(save_dir, f"model.onnx"),                  # where to save the model (can be a file or file-like object)
                export_params=True,                             # store the trained parameter weights inside the model file
                opset_version=10,                               # the ONNX version to export the model to
                do_constant_folding=True,                       # whether to execute constant folding for optimization
                input_names = ['input'],                        # the model's input names
                output_names = ['output'],                      # the model's output names
                dynamic_axes={'input' : {0 : 'batch_size'},     # variable length axes
                            'output' : {0 : 'batch_size'}})
    print("onnx model exported")  

~/miniconda3/envs/playground/lib/python3.8/site-packages/torchvision/transforms/functional.py in center_crop(img, output_size)
    514             return img
    515 
--> 516     crop_top = int(round((image_height - crop_height) / 2.))
    517     crop_left = int(round((image_width - crop_width) / 2.))
    518     return crop(img, crop_top, crop_left, crop_height, crop_width)

TypeError: type Tensor doesn't define __round__ method

Expected behavior

Model is saved to file and can be used with onnxruntime library later on.

Workaround (or possible solution)

in torchvision/transforms/functional.py function center_crop change lines 516-517 to use the numpy.round() instead of built-in round():

# old code for reference:    
# crop_top = int(round((image_height - crop_height) / 2.))
# crop_left = int(round((image_width - crop_width) / 2.))
# working by using numpy.round():
crop_top = int(np.round((image_height - crop_height) / 2.))
crop_left = int(np.round((image_width - crop_width) / 2.))

Versions

Collecting environment information...
PyTorch version: 1.9.1+cu102
Is debug build: False
CUDA used to build PyTorch: 10.2
ROCM used to build PyTorch: N/A

OS: Ubuntu 20.04.3 LTS (x86_64)
GCC version: (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0
Clang version: 10.0.0-4ubuntu1
CMake version: version 3.16.3
Libc version: glibc-2.31

Python version: 3.8.10 (default, Jun 4 2021, 15:09:15) [GCC 7.5.0] (64-bit runtime)
Python platform: Linux-5.4.0-81-generic-x86_64-with-glibc2.17
Is CUDA available: True
CUDA runtime version: Could not collect
GPU models and configuration: GPU 0: GeForce RTX 2080 Ti
Nvidia driver version: 460.91.03
cuDNN version: Probably one of the following:
/usr/lib/x86_64-linux-gnu/libcudnn.so.7.6.5
/usr/lib/x86_64-linux-gnu/libcudnn.so.8.2.4
/usr/lib/x86_64-linux-gnu/libcudnn_adv_infer.so.8.2.4
/usr/lib/x86_64-linux-gnu/libcudnn_adv_train.so.8.2.4
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_infer.so.8.2.4
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_train.so.8.2.4
/usr/lib/x86_64-linux-gnu/libcudnn_ops_infer.so.8.2.4
/usr/lib/x86_64-linux-gnu/libcudnn_ops_train.so.8.2.4
/usr/local/cuda-10.1/targets/x86_64-linux/lib/libcudnn.so.7
HIP runtime version: N/A
MIOpen runtime version: N/A

Versions of relevant libraries:
[pip3] numpy==1.21.2
[pip3] torch==1.9.1
[pip3] torchvision==0.10.1
[conda] numpy 1.21.2 pypi_0 pypi
[conda] torch 1.9.1 pypi_0 pypi
[conda] torchvision 0.10.1 pypi_0 pypi

cc @neginraoof

Start in torchvision/transforms/functional.py at center_crop, then reproduce the reported torch.onnx.export failure with the provided CNN-style example and environment details. Check the existing transform behavior and relevant test coverage; done means the export completes without the round TypeError and the ONNX file is saved for later use.