output = output.view(output.shape[0], -1)
return self.fc(output)
def score(self, x, y):
y_pred = self.forward(x)
acc = (y_pred.argmax(dim=1) == y).sum().item() / len(y)
return acc
if __name__ == "__main__":
model = VGG16()
model = model.cuda()
loss = nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(model.parameters(), lr=0.0003)
# load data
mnist_train, mnist_test = download_data_fashion_mnist(resize=224)
params = {
"epoch_num": 2,
"data_num": len(mnist_train),
"batch_size": 16,
"gpu": True,
"model": model,
"loss": loss,
"optimizer": optimizer,
"draw": True,
"evaluate": model.score,
# "accum_step": 8,
}
train_iter, test_iter = load_data_fashion_mnist(batch_size=params["batch_size"],
def score(self, x, y):
y_pred = self.forward(x)
acc = (y_pred.argmax(dim=1) == y).sum().item() / len(y)
return acc
if __name__ == "__main__":
# define model
model = LeNet5()
model = model.cuda()
# loss and optimizer
loss = nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(model.parameters(), lr=0.01, momentum=0.9)
# load data
mnist_train, mnist_test = download_data_fashion_mnist()
params = {
"epoch_num": 2,
"model": model,
"loss": loss,
"data_num": len(mnist_train),
"optimizer": optimizer,
"draw": True,
"gpu": True,
"batch_size": 256,
"evaluate": model.score,
"test_iter": Data.DataLoader(mnist_test, batch_size=len(mnist_test),
num_workers=8, pin_memory=True),
# "save_fig": True,
# "save_path": "../result/BN对比试验/img/"
y_pred = self.forward(x)
return (y_pred.argmax(dim=1) == y).sum().item() / len(y)
if __name__ == "__main__":
# define model
model = DropoutModel2()
# change to cuda if cuda is available
if torch.cuda.is_available():
model = model.cuda()
# loss
loss = nn.CrossEntropyLoss()
# optimizer
optimizer = torch.optim.SGD(model.parameters(), lr=0.3)
# download data
train_mnist, test_mnist = download_data_fashion_mnist()
# parameter
params = {
"model": model,
"loss": loss,
"epoch_num": 10,
"data_num": len(train_mnist),
"batch_size": 512,
"optimizer": optimizer,
# "test_iter": Data.DataLoader(test_mnist, len(test_mnist), shuffle=True),
"test_iter": Data.DataLoader(test_mnist, batch_size=512, shuffle=True),
# "evaluate": model.score,
"draw": True,
"gpu": True,
# "sample_rate": 0.5,
