def __init__(self) -> None:
self.model: Sequential = Sequential()
self.model.add(Conv2D(filters=6, kernel_size=(5, 5),
activation="relu"))
self.model.add(MaxPool2D(pool_size=(2, 2)))
self.model.add(
Conv2D(filters=16, kernel_size=(5, 5), activation="relu"))
self.model.add(MaxPool2D(pool_size=(2, 2)))
self.model.add(Flatten())
self.model.add(FC(units=120, activation="relu"))
self.model.add(FC(units=84, activation="relu"))
from dnet import datasets
from dnet.archs import LeNet5
from dnet.layers import FC
(x_train, y_train), (x_val, y_val) = datasets.tiny_mnist(flatten=False,
one_hot_encoding=True)
model = LeNet5()()
model.add(FC(units=10, activation="softmax"))
model.compile(loss="categorical_crossentropy",
optimizer="adam",
lr=1e-03,
bs=512)
model.fit(inputs=x_train,
targets=y_train,
epochs=10,
validation_data=(x_val, y_val))
model.plot_losses()
model.plot_accuracy()
from dnet import datasets
from dnet.layers import Conv2D, MaxPool2D, Flatten, FC
from dnet.models import Sequential
(x_train, y_train), (x_val, y_val) = datasets.tiny_mnist(flatten=False,
one_hot_encoding=True)
model = Sequential([
Conv2D(filters=6, kernel_size=(5, 5), activation="relu"),
MaxPool2D(pool_size=(2, 2)),
Conv2D(filters=16, kernel_size=(5, 5), activation="relu"),
MaxPool2D(pool_size=(2, 2)),
Flatten(),
FC(units=120, activation="relu"),
FC(units=84, activation="relu"),
FC(units=10, activation="softmax")
])
model.compile(loss="categorical_crossentropy",
optimizer="adam",
lr=1e-03,
bs=512)
model.fit(inputs=x_train,
targets=y_train,
epochs=10,
validation_data=(x_val, y_val))
model.plot_losses()
model.plot_accuracy()
def __init__(self) -> None:
self.model: Sequential = Sequential()
self.model.add(
Conv2D(filters=64,
kernel_size=(3, 3),
padding="same",
activation="relu"))
self.model.add(
Conv2D(filters=64,
kernel_size=(3, 3),
padding="same",
activation="relu"))
self.model.add(MaxPool2D(pool_size=(2, 2)))
self.model.add(
Conv2D(filters=128,
kernel_size=(3, 3),
padding="same",
activation="relu"))
self.model.add(
Conv2D(filters=128,
kernel_size=(3, 3),
padding="same",
activation="relu"))
self.model.add(MaxPool2D(pool_size=(2, 2)))
self.model.add(
Conv2D(filters=256,
kernel_size=(3, 3),
padding="same",
activation="relu"))
self.model.add(
Conv2D(filters=256,
kernel_size=(3, 3),
padding="same",
activation="relu"))
self.model.add(
Conv2D(filters=256,
kernel_size=(3, 3),
padding="same",
activation="relu"))
self.model.add(MaxPool2D(pool_size=(2, 2)))
self.model.add(
Conv2D(filters=512,
kernel_size=(3, 3),
padding="same",
activation="relu"))
self.model.add(
Conv2D(filters=512,
kernel_size=(3, 3),
padding="same",
activation="relu"))
self.model.add(
Conv2D(filters=512,
kernel_size=(3, 3),
padding="same",
activation="relu"))
self.model.add(MaxPool2D(pool_size=(2, 2)))
self.model.add(
Conv2D(filters=512,
kernel_size=(3, 3),
padding="same",
activation="relu"))
self.model.add(
Conv2D(filters=512,
kernel_size=(3, 3),
padding="same",
activation="relu"))
self.model.add(
Conv2D(filters=512,
kernel_size=(3, 3),
padding="same",
activation="relu"))
self.model.add(MaxPool2D(pool_size=(2, 2)))
self.model.add(Flatten())
self.model.add(FC(units=4096, activation="relu"))
self.model.add(FC(units=4096, activation="relu"))
from dnet import datasets from dnet.layers import FC from dnet.models import Sequential (x_train, y_train), (x_val, y_val) = datasets.tiny_mnist(flatten=True, one_hot_encoding=True) model = Sequential() model.add(FC(units=500, activation="relu")) model.add(FC(units=50, activation="relu")) model.add(FC(units=10, activation="softmax")) model.compile(loss="categorical_crossentropy", optimizer="rmsprop", lr=1e-03, bs=512) model.fit(inputs=x_train, targets=y_train, epochs=20, validation_data=(x_val, y_val)) model.plot_losses() model.plot_accuracy()