def test_static_save_dynamic_load(self):
path = tempfile.mkdtemp()
net = MyModel(classifier_activation=None)
inputs = [InputSpec([None, 20], 'float32', 'x')]
labels = [InputSpec([None, 1], 'int64', 'label')]
optim = fluid.optimizer.SGD(learning_rate=0.001,
parameter_list=net.parameters())
model = Model(net, inputs, labels)
model.prepare(optimizer=optim, loss=CrossEntropyLoss(reduction="sum"))
model.save(path + '/test')
device = paddle.set_device('cpu')
fluid.enable_dygraph(device) #if dynamic else None
net = MyModel(classifier_activation=None)
inputs = [InputSpec([None, 20], 'float32', 'x')]
labels = [InputSpec([None, 1], 'int64', 'label')]
optim = fluid.optimizer.SGD(learning_rate=0.001,
parameter_list=net.parameters())
model = Model(net, inputs, labels)
model.prepare(optimizer=optim, loss=CrossEntropyLoss(reduction="sum"))
model.load(path + '/test')
shutil.rmtree(path)
fluid.disable_dygraph()
def test_static_save_dynamic_load(self):
path = os.path.join(tempfile.mkdtemp(),
'.cache_test_static_save_dynamic_load')
if not os.path.exists(path):
os.makedirs(path)
net = MyModel()
inputs = [InputSpec([None, 20], 'float32', 'x')]
labels = [InputSpec([None, 1], 'int64', 'label')]
optim = fluid.optimizer.SGD(learning_rate=0.001,
parameter_list=net.parameters())
model = Model(net, inputs, labels)
model.prepare(optimizer=optim, loss=CrossEntropyLoss(reduction="sum"))
model.save(path)
device = paddle.set_device('cpu')
fluid.enable_dygraph(device) #if dynamic else None
net = MyModel()
inputs = [InputSpec([None, 20], 'float32', 'x')]
labels = [InputSpec([None, 1], 'int64', 'label')]
optim = fluid.optimizer.SGD(learning_rate=0.001,
parameter_list=net.parameters())
model = Model(net, inputs, labels)
model.prepare(optimizer=optim, loss=CrossEntropyLoss(reduction="sum"))
model.load(path)
shutil.rmtree(path)
fluid.disable_dygraph()
def test_fit_by_step(self):
base_lr = 1e-3
boundaries = [5, 8]
def make_optimizer(parameters=None):
momentum = 0.9
weight_decay = 5e-4
values = [base_lr * (0.1**i) for i in range(len(boundaries) + 1)]
learning_rate = paddle.optimizer.lr.PiecewiseDecay(
boundaries=boundaries, values=values)
learning_rate = paddle.optimizer.lr.LinearWarmup(
learning_rate=learning_rate,
warmup_steps=4,
start_lr=base_lr / 5.,
end_lr=base_lr,
verbose=True)
optimizer = paddle.optimizer.Momentum(
learning_rate=learning_rate,
weight_decay=weight_decay,
momentum=momentum,
parameters=parameters)
return optimizer
# dynamic test
device = paddle.set_device('cpu')
fluid.enable_dygraph(device)
net = MyModel()
inputs = [InputSpec([None, 20], 'float32', 'x')]
labels = [InputSpec([None, 1], 'int64', 'label')]
optim = make_optimizer(net.parameters())
model = Model(net, inputs, labels)
model.prepare(optimizer=optim, loss=CrossEntropyLoss(reduction="sum"))
dataset = MyDataset()
model.fit(dataset, dataset, batch_size=4, epochs=10, num_workers=0)
np.testing.assert_allclose(model._optimizer._learning_rate.last_lr,
base_lr * (0.1**len(boundaries)))
# static test
paddle.enable_static()
net = MyModel()
inputs = [InputSpec([None, 20], 'float32', 'x')]
labels = [InputSpec([None, 1], 'int64', 'label')]
optim = make_optimizer(net.parameters())
model = Model(net, inputs, labels)
model.prepare(optimizer=optim, loss=CrossEntropyLoss(reduction="sum"))
dataset = MyDataset()
model.fit(dataset, dataset, batch_size=4, epochs=10, num_workers=0)
np.testing.assert_allclose(model._optimizer._learning_rate.last_lr,
base_lr * (0.1**len(boundaries)))
def test_static_multiple_gpus(self):
device = set_device('gpu')
im_shape = (-1, 1, 28, 28)
batch_size = 128
inputs = [Input(im_shape, 'float32', 'image')]
labels = [Input([None, 1], 'int64', 'label')]
model = Model(LeNet(), inputs, labels)
optim = fluid.optimizer.Momentum(
learning_rate=0.001, momentum=.9, parameter_list=model.parameters())
model.prepare(optim, CrossEntropyLoss(), Accuracy())
train_dataset = MnistDataset(mode='train')
val_dataset = MnistDataset(mode='test')
test_dataset = MnistDataset(mode='test', return_label=False)
cbk = paddle.callbacks.ProgBarLogger(50)
model.fit(train_dataset,
val_dataset,
epochs=2,
batch_size=batch_size,
callbacks=cbk)
eval_result = model.evaluate(val_dataset, batch_size=batch_size)
output = model.predict(
test_dataset, batch_size=batch_size, stack_outputs=True)
np.testing.assert_equal(output[0].shape[0], len(test_dataset))
acc = compute_accuracy(output[0], val_dataset.labels)
np.testing.assert_allclose(acc, eval_result['acc'])
def test_dynamic_load(self):
mnist_data = MnistDataset(mode='train')
path = os.path.join(tempfile.mkdtemp(), '.cache_dynamic_load')
if not os.path.exists(path):
os.makedirs(path)
for new_optimizer in [True, False]:
paddle.disable_static()
net = LeNet()
inputs = [InputSpec([None, 1, 28, 28], 'float32', 'x')]
labels = [InputSpec([None, 1], 'int64', 'label')]
if new_optimizer:
optim = paddle.optimizer.Adam(
learning_rate=0.001, parameters=net.parameters())
else:
optim = fluid.optimizer.Adam(
learning_rate=0.001, parameter_list=net.parameters())
model = Model(net, inputs, labels)
model.prepare(
optimizer=optim, loss=CrossEntropyLoss(reduction="sum"))
model.fit(mnist_data, batch_size=64, verbose=0)
model.save(path)
model.load(path)
paddle.enable_static()
shutil.rmtree(path)
def test_accumulate(self, ):
dim = 20
data = np.random.random(size=(4, dim)).astype(np.float32)
label = np.random.randint(0, 10, size=(4, 1)).astype(np.int64)
net = MyModel()
optim = fluid.optimizer.SGD(learning_rate=0.001,
parameter_list=net.parameters())
inputs = [InputSpec([None, dim], 'float32', 'x')]
labels = [InputSpec([None, 1], 'int64', 'label')]
for amp_cfg in [None, 'O1']:
model = Model(net, inputs, labels)
model.prepare(
optim,
loss=CrossEntropyLoss(reduction="sum"),
amp_configs=amp_cfg)
losses, grads = [], []
for stat in [False, False, True]:
loss, = model.train_batch([data], [label], update=stat)
losses.append(loss)
grads.append([p.grad.numpy() for p in net.parameters()])
for grad1, grad2, grad3 in zip(*grads):
np.testing.assert_almost_equal(grad1 * 2, grad2, decimal=4)
np.testing.assert_almost_equal(
grad3, np.zeros_like(grad3), decimal=4)
np.testing.assert_almost_equal(losses[0], losses[1], decimal=4)
np.testing.assert_almost_equal(losses[0], losses[2], decimal=4)
def test_dygraph_export_deploy_model_about_inputs(self):
self.set_seed()
np.random.seed(201)
mnist_data = MnistDataset(mode='train')
paddle.disable_static()
# without inputs
save_dir = os.path.join(tempfile.mkdtemp(),
'.cache_test_dygraph_export_deploy')
if not os.path.exists(save_dir):
os.makedirs(save_dir)
for initial in ["fit", "train_batch", "eval_batch", "predict_batch"]:
net = LeNet()
model = Model(net)
optim = fluid.optimizer.Adam(
learning_rate=0.001, parameter_list=model.parameters())
model.prepare(
optimizer=optim, loss=CrossEntropyLoss(reduction="sum"))
if initial == "fit":
model.fit(mnist_data, batch_size=64, verbose=0)
else:
img = np.array(
np.random.random((1, 1, 28, 28)), dtype=np.float32)
label = np.array(np.random.rand(1, 1), dtype=np.int64)
if initial == "train_batch":
model.train_batch([img], [label])
elif initial == "eval_batch":
model.eval_batch([img], [label])
else:
model.predict_batch([img])
model.save(save_dir, training=False)
shutil.rmtree(save_dir)
# with inputs, and the type of inputs is InputSpec
save_dir = os.path.join(tempfile.mkdtemp(),
'.cache_test_dygraph_export_deploy_2')
if not os.path.exists(save_dir):
os.makedirs(save_dir)
net = LeNet()
inputs = InputSpec([None, 1, 28, 28], 'float32', 'x')
model = Model(net, inputs)
optim = fluid.optimizer.Adam(
learning_rate=0.001, parameter_list=model.parameters())
model.prepare(optimizer=optim, loss=CrossEntropyLoss(reduction="sum"))
model.save(save_dir, training=False)
shutil.rmtree(save_dir)
def func_warn_or_error(self):
with self.assertRaises(ValueError):
paddle.callbacks.ReduceLROnPlateau(factor=2.0)
# warning
paddle.callbacks.ReduceLROnPlateau(mode='1', patience=3, verbose=1)
transform = T.Compose([T.Transpose(), T.Normalize([127.5], [127.5])])
train_dataset = CustomMnist(mode='train', transform=transform)
val_dataset = CustomMnist(mode='test', transform=transform)
net = LeNet()
optim = paddle.optimizer.Adam(learning_rate=0.001,
parameters=net.parameters())
inputs = [InputSpec([None, 1, 28, 28], 'float32', 'x')]
labels = [InputSpec([None, 1], 'int64', 'label')]
model = Model(net, inputs=inputs, labels=labels)
model.prepare(optim, loss=CrossEntropyLoss(), metrics=[Accuracy()])
callbacks = paddle.callbacks.ReduceLROnPlateau(monitor='miou',
patience=3,
verbose=1)
model.fit(train_dataset,
val_dataset,
batch_size=8,
log_freq=1,
save_freq=10,
epochs=1,
callbacks=[callbacks])
optim = paddle.optimizer.Adam(
learning_rate=paddle.optimizer.lr.PiecewiseDecay([0.001, 0.0001],
[5, 10]),
parameters=net.parameters())
model.prepare(optim, loss=CrossEntropyLoss(), metrics=[Accuracy()])
callbacks = paddle.callbacks.ReduceLROnPlateau(monitor='acc',
mode='max',
patience=3,
verbose=1,
cooldown=1)
model.fit(train_dataset,
val_dataset,
batch_size=8,
log_freq=1,
save_freq=10,
epochs=3,
callbacks=[callbacks])
def dynamic_train(model, dataloader):
optim = fluid.optimizer.Adam(
learning_rate=0.001, parameter_list=model.parameters())
model.train()
for inputs, labels in dataloader:
outputs = model(inputs)
loss = CrossEntropyLoss(reduction="sum")(outputs, labels)
avg_loss = fluid.layers.reduce_sum(loss)
avg_loss.backward()
optim.minimize(avg_loss)
model.clear_gradients()
def fit(self, dynamic, num_replicas=None, rank=None, num_iters=None):
fluid.enable_dygraph(self.device) if dynamic else None
seed = 333
paddle.seed(seed)
paddle.framework.random._manual_program_seed(seed)
net = LeNet()
optim_new = fluid.optimizer.Adam(
learning_rate=0.001, parameter_list=net.parameters())
model = Model(net, inputs=self.inputs, labels=self.labels)
model.prepare(
optim_new,
loss=CrossEntropyLoss(reduction="sum"),
metrics=Accuracy())
model.fit(self.train_dataset, batch_size=64, shuffle=False)
result = model.evaluate(self.val_dataset, batch_size=64)
np.testing.assert_allclose(result['acc'], self.acc1)
model.fit(self.train_dataset,
batch_size=64,
shuffle=False,
num_iters=num_iters)
result = model.evaluate(
self.val_dataset, batch_size=64, num_iters=num_iters)
train_sampler = DistributedBatchSampler(
self.train_dataset,
batch_size=64,
shuffle=False,
num_replicas=num_replicas,
rank=rank)
val_sampler = DistributedBatchSampler(
self.val_dataset,
batch_size=64,
shuffle=False,
num_replicas=num_replicas,
rank=rank)
train_loader = fluid.io.DataLoader(
self.train_dataset,
batch_sampler=train_sampler,
places=self.device,
return_list=True)
val_loader = fluid.io.DataLoader(
self.val_dataset,
batch_sampler=val_sampler,
places=self.device,
return_list=True)
model.fit(train_loader, val_loader)
fluid.disable_dygraph() if dynamic else None
def get_model(self, amp_config):
net = LeNet()
inputs = InputSpec([None, 1, 28, 28], "float32", 'x')
labels = InputSpec([None, 1], "int64", "y")
model = Model(net, inputs, labels)
optim = paddle.optimizer.Adam(learning_rate=0.001,
parameters=model.parameters())
model.prepare(optimizer=optim,
loss=CrossEntropyLoss(reduction="sum"),
amp_configs=amp_config)
return model
def get_expect():
fluid.enable_dygraph(fluid.CPUPlace())
self.set_seed()
m = MyModel()
optim = fluid.optimizer.SGD(learning_rate=0.001,
parameter_list=m.parameters())
m.train()
output = m(to_tensor(data))
loss = CrossEntropyLoss(reduction='sum')(output, to_tensor(label))
avg_loss = fluid.layers.reduce_sum(loss)
avg_loss.backward()
optim.minimize(avg_loss)
m.clear_gradients()
fluid.disable_dygraph()
return avg_loss.numpy()
def test_static_check_input(self):
paddle.enable_static()
amp_configs = {"level": "O2", "use_pure_fp16": True}
if not fluid.is_compiled_with_cuda():
self.skipTest('module not tested when ONLY_CPU compling')
paddle.set_device('gpu')
net = LeNet()
inputs = InputSpec([None, 1, 28, 28], "float32", 'x')
labels = InputSpec([None, 1], "int64", "y")
model = Model(net, inputs, labels)
optim = paddle.optimizer.Adam(learning_rate=0.001,
parameters=model.parameters())
loss = CrossEntropyLoss(reduction="sum")
with self.assertRaises(ValueError):
model.prepare(optimizer=optim, loss=loss, amp_configs=amp_configs)
def test_amp_training_purefp16(self):
if not fluid.is_compiled_with_cuda():
self.skipTest('module not tested when ONLY_CPU compling')
data = np.random.random(size=(4, 1, 28, 28)).astype(np.float32)
label = np.random.randint(0, 10, size=(4, 1)).astype(np.int64)
paddle.enable_static()
paddle.set_device('gpu')
net = LeNet()
amp_level = "O2"
inputs = InputSpec([None, 1, 28, 28], "float32", 'x')
labels = InputSpec([None, 1], "int64", "y")
model = Model(net, inputs, labels)
optim = paddle.optimizer.Adam(learning_rate=0.001,
parameters=model.parameters(),
multi_precision=True)
amp_configs = {"level": amp_level, "use_fp16_guard": False}
model.prepare(optimizer=optim,
loss=CrossEntropyLoss(reduction="sum"),
amp_configs=amp_configs)
model.train_batch([data], [label])
def func_reduce_lr_on_plateau(self):
transform = T.Compose([T.Transpose(), T.Normalize([127.5], [127.5])])
train_dataset = CustomMnist(mode='train', transform=transform)
val_dataset = CustomMnist(mode='test', transform=transform)
net = LeNet()
optim = paddle.optimizer.Adam(learning_rate=0.001,
parameters=net.parameters())
inputs = [InputSpec([None, 1, 28, 28], 'float32', 'x')]
labels = [InputSpec([None, 1], 'int64', 'label')]
model = Model(net, inputs=inputs, labels=labels)
model.prepare(optim, loss=CrossEntropyLoss(), metrics=[Accuracy()])
callbacks = paddle.callbacks.ReduceLROnPlateau(patience=1,
verbose=1,
cooldown=1)
model.fit(train_dataset,
val_dataset,
batch_size=8,
log_freq=1,
save_freq=10,
epochs=10,
callbacks=[callbacks])
def test_train_batch(self, dynamic=True):
dim = 20
data = np.random.random(size=(4, dim)).astype(np.float32)
label = np.random.randint(0, 10, size=(4, 1)).astype(np.int64)
def get_expect():
fluid.enable_dygraph(fluid.CPUPlace())
self.set_seed()
m = MyModel()
optim = fluid.optimizer.SGD(learning_rate=0.001,
parameter_list=m.parameters())
m.train()
output = m(to_tensor(data))
loss = CrossEntropyLoss(reduction='sum')(output, to_tensor(label))
avg_loss = fluid.layers.reduce_sum(loss)
avg_loss.backward()
optim.minimize(avg_loss)
m.clear_gradients()
fluid.disable_dygraph()
return avg_loss.numpy()
ref = get_expect()
for dynamic in [True, False]:
device = paddle.set_device('cpu')
fluid.enable_dygraph(device) if dynamic else None
self.set_seed()
net = MyModel()
optim2 = fluid.optimizer.SGD(learning_rate=0.001,
parameter_list=net.parameters())
inputs = [InputSpec([None, dim], 'float32', 'x')]
labels = [InputSpec([None, 1], 'int64', 'label')]
model = Model(net, inputs, labels)
model.prepare(optim2, loss=CrossEntropyLoss(reduction="sum"))
loss, = model.train_batch([data], [label])
np.testing.assert_allclose(loss.flatten(), ref.flatten())
fluid.disable_dygraph() if dynamic else None
def test_dynamic_check_input(self):
paddle.disable_static()
amp_configs_list = [
{
"level": "O3"
},
{
"level": "O1",
"test": 0
},
{
"level": "O1",
"use_fp16_guard": True
},
"O3",
]
if not fluid.is_compiled_with_cuda():
self.skipTest('module not tested when ONLY_CPU compling')
paddle.set_device('gpu')
net = LeNet()
model = Model(net)
optim = paddle.optimizer.Adam(learning_rate=0.001,
parameters=model.parameters())
loss = CrossEntropyLoss(reduction="sum")
with self.assertRaises(ValueError):
for amp_configs in amp_configs_list:
model.prepare(optimizer=optim,
loss=loss,
amp_configs=amp_configs)
model.prepare(optimizer=optim, loss=loss, amp_configs="O2")
model.prepare(optimizer=optim,
loss=loss,
amp_configs={
"custom_white_list": {"matmul"},
"init_loss_scaling": 1.0
})
def test_earlystopping(self):
paddle.seed(2020)
for dynamic in [True, False]:
paddle.enable_static if not dynamic else None
device = paddle.set_device('cpu')
sample_num = 100
train_dataset = MnistDataset(mode='train', sample_num=sample_num)
val_dataset = MnistDataset(mode='test', sample_num=sample_num)
net = LeNet()
optim = paddle.optimizer.Adam(learning_rate=0.001,
parameters=net.parameters())
inputs = [InputSpec([None, 1, 28, 28], 'float32', 'x')]
labels = [InputSpec([None, 1], 'int64', 'label')]
model = Model(net, inputs=inputs, labels=labels)
model.prepare(optim,
loss=CrossEntropyLoss(reduction="sum"),
metrics=[Accuracy()])
callbacks_0 = paddle.callbacks.EarlyStopping('loss',
mode='min',
patience=1,
verbose=1,
min_delta=0,
baseline=None,
save_best_model=True)
callbacks_1 = paddle.callbacks.EarlyStopping('acc',
mode='auto',
patience=1,
verbose=1,
min_delta=0,
baseline=0,
save_best_model=True)
callbacks_2 = paddle.callbacks.EarlyStopping('loss',
mode='auto_',
patience=1,
verbose=1,
min_delta=0,
baseline=None,
save_best_model=True)
callbacks_3 = paddle.callbacks.EarlyStopping('acc_',
mode='max',
patience=1,
verbose=1,
min_delta=0,
baseline=0,
save_best_model=True)
model.fit(
train_dataset,
val_dataset,
batch_size=64,
save_freq=10,
save_dir=self.save_dir,
epochs=10,
verbose=0,
callbacks=[callbacks_0, callbacks_1, callbacks_2, callbacks_3])
# Test for no val_loader
model.fit(train_dataset,
batch_size=64,
save_freq=10,
save_dir=self.save_dir,
epochs=10,
verbose=0,
callbacks=[callbacks_0])