MovingMNIST in PyTorch

Buy Me a Coffee☕

*Memos:

• My post explains Moving MNIST.
• My post explains KMNIST.
• My post explains MNIST().
• My post explains EMNIST().
• My post explains QMNIST().
• My post explains KMNIST().
• My post explains FashionMNIST().

MovingMNIST() can use Moving MNIST dataset as shown below:

*Memos:

• The 1st argument is root(Required-Type:str or pathlib.Path). *An absolute or relative path is possible.
• The 2nd argument is split(Optional-Default:None-Type:str): *Memos:
  • None, "train" or "test" can be set to it.
  • If it's None, all 20 frames(images) of each video are returned, ignoring split_ratio.
• The 3rd argument is split_ratio(Optional-Default:10-Type:int): *Memos:
  • If split is "train", data[:, :split_ratio] is returned.
  • If split is "test", data[:, split_ratio:] is returned.
  • If split is None, it's ignored. ignoring split_ratio.
• The 4th argument is transform(Optional-Default:None-Type:callable).
• The 5th argument is download(Optional-Default:False-Type:bool): *Memos:
  • If it's True, the dataset is downloaded from the internet to root.
  • If it's True and the dataset is already downloaded, it's extracted.
  • If it's True and the dataset is already downloaded, nothing happens.
  • It should be False if the dataset is already downloaded because it's faster.
  • You can manually download and extract the dataset(mnist_test_seq.npy) from here to data/MovingMNIST/.

from torchvision.datasets import MovingMNIST

all_data = MovingMNIST(
    root="data"
)

all_data = MovingMNIST(
    root="data",
    split=None,
    split_ratio=10,
    download=False,
    transform=None
)

train_data = MovingMNIST(
    root="data",
    split="train"
)

test_data = MovingMNIST(
    root="data",
    split="test"
)

len(all_data), len(train_data), len(test_data)
# (10000, 10000, 10000)

len(all_data[0]), len(train_data[0]), len(test_data[0])
# (20, 10, 10)

all_data
# Dataset MovingMNIST
#     Number of datapoints: 10000
#     Root location: data

all_data.root
# 'data'

print(all_data.split)
# None

all_data.split_ratio
# 10

all_data.download
# <bound method MovingMNIST.download of Dataset MovingMNIST
#     Number of datapoints: 10000
#     Root location: data>

print(all_data.transform)
# None

all_data[0]
# tensor([[[[0, 0, 0,  ..., 0, 0, 0],
#           ...,
#           [0, 0, 0,  ..., 0, 0, 0]]],
#         ...
#         [[[0, 0, 0,  ..., 0, 0, 0],
#           ...,
#           [0, 0, 0,  ..., 0, 0, 0]]]], dtype=torch.uint8)

all_data[1]
# tensor([[[[0, 0, 0,  ..., 0, 0, 0],
#           ...,
#           [0, 0, 0,  ..., 0, 0, 0]]],
#         ...
#         [[[0, 0, 0,  ..., 0, 0, 0],
#           ...,
#           [0, 0, 0,  ..., 0, 0, 0]]]], dtype=torch.uint8)

all_data[2]
# tensor([[[[0, 0, 0,  ..., 0, 0, 0],
#           ...,
#           [0, 0, 0,  ..., 0, 0, 0]]],
#         ...
#         [[[0, 0, 0,  ..., 0, 0, 0],
#           ...,
#           [0, 0, 0,  ..., 0, 0, 0]]]], dtype=torch.uint8)

import matplotlib.pyplot as plt

def show_images(data, labs):
    plt.figure(figsize=(8, 4))
    for i, (vid, lab) in enumerate(iterable=zip(data, labs), start=1):
        plt.subplot(1, 3, i)
        plt.imshow(X=vid.squeeze()[0])
        plt.title(label=lab)
    plt.tight_layout()
    plt.show()

videos = (all_data[0], train_data[0], test_data[0])
titles = ("all_data", "train_data", "test_data")

show_images(data=videos, labs=titles)

from torchvision.datasets import MovingMNIST

all_data = MovingMNIST(
    root="data",
    split=None
)

train_data = MovingMNIST(
    root="data",
    split="train"
)

test_data = MovingMNIST(
    root="data",
    split="test"
)

import matplotlib.pyplot as plt

def show_images(data, main_title=None):
    plt.figure(figsize=(12, 10))
    plt.suptitle(t=main_title, y=1.0, fontsize=14)
    for i, im in enumerate(iterable=data.squeeze(), start=1):
        plt.subplot(4, 5, i)
        plt.title(label=i)
        plt.imshow(X=im)
    plt.tight_layout()
    plt.show()

show_images(data=all_data[0], main_title="all_data")
show_images(data=train_data[0], main_title="train_data")
show_images(data=test_data[0], main_title="test_data")

from torchvision.datasets import MovingMNIST

all_data = MovingMNIST(
    root="data",
    split=None
)

train_data = MovingMNIST(
    root="data",
    split="train"
)

test_data = MovingMNIST(
    root="data",
    split="test"
)

import matplotlib.pyplot as plt

def show_images(data, main_title=None):
    plt.figure(figsize=(10, 8))
    plt.suptitle(t=main_title, y=1.0, fontsize=14)
    for i, vid in zip(range(1, 6), data):
        plt.subplot(4, 5, i)
        plt.title(label=i)
        plt.imshow(X=vid.squeeze()[0])
    plt.tight_layout()
    plt.show()

show_images(data=all_data, main_title="all_data")
show_images(data=train_data, main_title="train_data")
show_images(data=test_data, main_title="test_data")

from torchvision.datasets import MovingMNIST
import matplotlib.animation as animation

all_data = MovingMNIST(
    root="data"
)

import matplotlib.pyplot as plt
from IPython.display import HTML

figure, axis = plt.subplots()

# ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ `ArtistAnimation()` ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓
ims = []
for im in all_data[0].squeeze():
    ims.append([axis.imshow(X=im)])
ani = animation.ArtistAnimation(fig=figure, artists=ims,
                                interval=100)
# ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ `ArtistAnimation()` ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑

# ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ `FuncAnimation()` ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓
# def animate(i):
#     axis.imshow(X=all_data[0].squeeze()[i])
#
# ani = animation.FuncAnimation(fig=figure, func=animate,
#                               frames=20, interval=100)
# ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ `FuncAnimation()` ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑

# ani.save('result.gif') # Save the animation as a `.gif` file

plt.ioff() # Hide a useless image

# ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ Show animation ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓
HTML(ani.to_jshtml()) # Animation operator
# HTML(ani.to_html5_video()) # Animation video
# ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ Show animation ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑

# ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ Show animation ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓
# plt.rcParams["animation.html"] = "jshtml" # Animation operator
# plt.rcParams["animation.html"] = "html5" # Animation video
# ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ Show animation ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑

from torchvision.datasets import MovingMNIST
from ipywidgets import interact, IntSlider

all_data = MovingMNIST(
    root="data"
)

import matplotlib.pyplot as plt
from IPython.display import HTML

def func(i):
    plt.imshow(X=all_data[0].squeeze()[i])

interact(func, i=(0, 19, 1))
# interact(func, i=IntSlider(min=0, max=19, step=1, value=0))
# ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ Set the start value ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑
plt.show()