Sequence primitive for neural network layers.

This JIRA task is to construct a sequence primitive for the python API that is able to combine multiple neural network layers, and perform forward and backwards passes.

The result should be similar to Keras and PyTorch (not forcefully similar, but a nice design like theirs.)

from tensorflow.keras import layers, models
p=’same’
model = models.Sequential([
layers.Conv2D(32, (3, 3), activation=’relu’, padding=p, input_shape=(H, W, C)),
layers.Conv2D(64, (3, 3), activation=’relu’, padding=p),
layers.Conv2D(128, (3, 3), activation=’relu’, padding=p),
layers.Conv2D(1, (1, 1), padding=p)])
Or pytorch:
import torch.nn as nn
model = nn.Sequential(
nn.Conv2d(C, 32, kernel_size=3, padding=1), nn.ReLU(),
nn.Conv2d(32, 64, kernel_size=3, padding=1), nn.ReLU(),
nn.Conv2d(64, 128, kernel_size=3, padding=1), nn.ReLU(),
nn.Conv2d(128, 1, kernel_size=1, padding=0))