The Problem of Creating Moving MNIST

[cyoungx]

Hello， Your work is really good. However, I wonder how to create the moving MNIST correctly.

After running your python file, MGP-VAE/create_data/create_moving_mnist_data.py, then it feedback the information as shown below.

Traceback (most recent call last):
  File "./create_data/create_moving_mnist_data.py", line 115, in <module>
    main(**{k:v for (k,v) in vars(args).items() if v is not None})
  File "./create_data/create_moving_mnist_data.py", line 82, in main
    dat = generate_moving_mnist(shape=(frame_size,frame_size), seq_len=seq_len, seqs=seqs, \
  File "./create_data/create_moving_mnist_data.py", line 67, in generate_moving_mnist
    canv.paste(mnist_images[i], tuple(map(lambda p: int(round(p)), positions[i])))
  File "/root/miniconda3/envs/MGP/lib/python3.8/site-packages/PIL/Image.py", line 1734, in paste
    self.im.paste(im, box)
TypeError: argument 2 must be sequence of length 4, not 0

If you can slove it, I'll be grateful.