Create PR to torchvision for MovingMNIST

[tsugumi-sys]

pytorch/vision#3562

[tsugumi-sys]

1. Transform for each frames: Done
2. There two ways (1: Read form file and split. 2: Generate dataset with this class and split.) I prefer 1 to 2 and tensorflow also do 1. The moving sequence dataset implemented by tensorflow is an other dataset.
3. Simply return a flames, Not split into input and label. : remove input_seq_length. Moving MNIST is used both supervised and unsupervised training, so we should just return frames here.

[tsugumi-sys]

Add tests

• https://github.com/pytorch/vision/blob/main/test/test_datasets_download.py
• https://github.com/pytorch/vision/blob/main/test/test_datasets.py

Change docs

• https://github.com/pytorch/vision/blob/main/docs/source/datasets.rst