584 lines
23 KiB
Python
584 lines
23 KiB
Python
# Copyright (c) OpenMMLab. All rights reserved.
|
|
import math
|
|
from unittest import TestCase
|
|
|
|
import torch
|
|
import torch.nn.functional as F
|
|
import torch.optim as optim
|
|
|
|
from mmengine.optim.scheduler import (ConstantMomentum,
|
|
CosineAnnealingMomentum,
|
|
CosineRestartMomentum,
|
|
ExponentialMomentum, LinearMomentum,
|
|
MultiStepMomentum, PolyMomentum,
|
|
StepMomentum, _ParamScheduler)
|
|
from mmengine.testing import assert_allclose
|
|
|
|
|
|
class ToyModel(torch.nn.Module):
|
|
|
|
def __init__(self):
|
|
super().__init__()
|
|
self.conv1 = torch.nn.Conv2d(1, 1, 1)
|
|
self.conv2 = torch.nn.Conv2d(1, 1, 1)
|
|
|
|
def forward(self, x):
|
|
return self.conv2(F.relu(self.conv1(x)))
|
|
|
|
|
|
class TestMomentumScheduler(TestCase):
|
|
|
|
def setUp(self):
|
|
"""Setup the model and optimizer which are used in every test method.
|
|
|
|
TestCase calls functions in this order: setUp() -> testMethod() ->
|
|
tearDown() -> cleanUp()
|
|
"""
|
|
self.model = ToyModel()
|
|
momentum = 0.05
|
|
self.layer2_mult = 10
|
|
self.optimizer = optim.SGD([{
|
|
'params': self.model.conv1.parameters()
|
|
}, {
|
|
'params': self.model.conv2.parameters(),
|
|
'momentum': momentum * self.layer2_mult
|
|
}],
|
|
lr=0.01,
|
|
momentum=momentum,
|
|
weight_decay=5e-4)
|
|
self.optimizer_with_betas = optim.Adam(
|
|
[{
|
|
'params': self.model.conv1.parameters()
|
|
}, {
|
|
'params': self.model.conv2.parameters(),
|
|
'betas': (momentum * self.layer2_mult, 0.999)
|
|
}],
|
|
lr=0.01,
|
|
betas=(momentum, 0.999),
|
|
weight_decay=5e-4)
|
|
|
|
def test_invalid_optimizer(self):
|
|
with self.assertRaisesRegex(
|
|
ValueError,
|
|
'optimizer must support momentum when using momentum scheduler'
|
|
):
|
|
optimizer = optim.ASGD(
|
|
self.model.parameters(),
|
|
lr=0.01,
|
|
)
|
|
StepMomentum(optimizer, step_size=1)
|
|
|
|
def test_overwrite_optimzer_step(self):
|
|
# raise warning if the counter in optimizer.step() is overwritten
|
|
scheduler = ExponentialMomentum(self.optimizer, gamma=0.9)
|
|
|
|
def overwrite_fun():
|
|
pass
|
|
|
|
self.optimizer.step = overwrite_fun
|
|
self.optimizer.step()
|
|
self.assertWarnsRegex(UserWarning, r'how-to-adjust-learning-rate',
|
|
scheduler.step)
|
|
|
|
def test_resume(self):
|
|
# test invalid case: optimizer and scheduler are not both resumed
|
|
with self.assertRaisesRegex(
|
|
KeyError, "param 'initial_momentum' is not specified"):
|
|
StepMomentum(self.optimizer, gamma=0.1, step_size=3, last_step=10)
|
|
|
|
# test manually resume with ``last_step`` instead of load_state_dict
|
|
epochs = 10
|
|
targets = [0.05 * (0.9**x) for x in range(epochs)]
|
|
scheduler = ExponentialMomentum(self.optimizer, gamma=0.9)
|
|
|
|
results = []
|
|
for epoch in range(5):
|
|
results.append(self.optimizer.param_groups[0]['momentum'])
|
|
# The order should be
|
|
# train_epoch() -> save_checkpoint() -> scheduler.step().
|
|
# Break at here to simulate the checkpoint is saved before
|
|
# the scheduler.step().
|
|
if epoch == 4:
|
|
break
|
|
scheduler.step()
|
|
scheduler2 = ExponentialMomentum(
|
|
self.optimizer, gamma=0.9, last_step=4)
|
|
for epoch in range(6):
|
|
results.append(self.optimizer.param_groups[0]['momentum'])
|
|
scheduler2.step()
|
|
|
|
for epoch in range(epochs):
|
|
assert_allclose(
|
|
targets[epoch],
|
|
results[epoch],
|
|
msg='momentum is wrong in epoch {}: expected {}, got {}'.
|
|
format(epoch, targets[epoch], results[epoch]),
|
|
atol=1e-5,
|
|
rtol=0)
|
|
|
|
def test_scheduler_before_optim_warning(self):
|
|
"""warns if scheduler is used before optimizer."""
|
|
|
|
def call_sch_before_optim():
|
|
scheduler = StepMomentum(self.optimizer, gamma=0.1, step_size=3)
|
|
scheduler.step()
|
|
self.optimizer.step()
|
|
|
|
# check warning doc link
|
|
self.assertWarnsRegex(UserWarning, r'how-to-adjust-learning-rate',
|
|
call_sch_before_optim)
|
|
|
|
# check warning when resume
|
|
for i, group in enumerate(self.optimizer.param_groups):
|
|
group['initial_momentum'] = 0.01
|
|
|
|
def call_sch_before_optim_resume():
|
|
scheduler = StepMomentum(
|
|
self.optimizer, gamma=0.1, step_size=3, last_step=10)
|
|
scheduler.step()
|
|
self.optimizer.step()
|
|
|
|
# check warning doc link
|
|
self.assertWarnsRegex(UserWarning, r'how-to-adjust-learning-rate',
|
|
call_sch_before_optim_resume)
|
|
|
|
def test_get_last_value(self):
|
|
epochs = 10
|
|
single_targets = [0.05] * 3 + [0.005] * 3 + [0.0005] * 3 + [0.00005]
|
|
targets = [
|
|
single_targets, [t * self.layer2_mult for t in single_targets]
|
|
]
|
|
scheduler = StepMomentum(self.optimizer, 3, gamma=0.1)
|
|
for epoch in range(epochs):
|
|
result = scheduler.get_last_value()
|
|
self.optimizer.step()
|
|
scheduler.step()
|
|
target = [t[epoch] for t in targets]
|
|
for t, r in zip(target, result):
|
|
assert_allclose(
|
|
target,
|
|
result,
|
|
msg='momentum is wrong in epoch {}: expected {}, got {}'.
|
|
format(epoch, t, r),
|
|
atol=1e-5,
|
|
rtol=0)
|
|
|
|
def test_scheduler_step_count(self):
|
|
iteration = 10
|
|
scheduler = StepMomentum(self.optimizer, gamma=0.1, step_size=3)
|
|
self.assertEqual(scheduler.last_step, 0)
|
|
target = [i + 1 for i in range(iteration)]
|
|
step_counts = []
|
|
for i in range(iteration):
|
|
self.optimizer.step()
|
|
scheduler.step()
|
|
step_counts.append(scheduler.last_step)
|
|
self.assertEqual(step_counts, target)
|
|
|
|
def test_effective_interval(self):
|
|
# check invalid begin end
|
|
with self.assertRaisesRegex(ValueError,
|
|
'end should be larger than begin'):
|
|
StepMomentum(
|
|
self.optimizer, gamma=0.1, step_size=3, begin=10, end=5)
|
|
|
|
# momentum = 0.05 if epoch == 0
|
|
# momentum = 0.025 if epoch == 1
|
|
# momentum = 0.03125 if epoch == 2
|
|
# momentum = 0.0375 if epoch == 3
|
|
# momentum = 0.04375 if epoch == 4
|
|
# momentum = 0.005 if epoch > 4
|
|
begin = 1
|
|
epochs = 10
|
|
start_factor = 1.0 / 2
|
|
iters = 4
|
|
interpolation = [
|
|
start_factor + i * (1 - start_factor) / iters for i in range(iters)
|
|
]
|
|
single_targets = [0.05] * begin + [x * 0.05
|
|
for x in interpolation] + [0.05] * (
|
|
epochs - iters - begin)
|
|
targets = [
|
|
single_targets, [x * self.layer2_mult for x in single_targets]
|
|
]
|
|
scheduler = LinearMomentum(
|
|
self.optimizer,
|
|
start_factor=start_factor,
|
|
begin=begin,
|
|
end=begin + iters + 1)
|
|
self._test_scheduler_value(self.optimizer, scheduler, targets, epochs)
|
|
|
|
def _test_scheduler_value(self,
|
|
optimizer,
|
|
schedulers,
|
|
targets,
|
|
epochs=10,
|
|
param_name='momentum'):
|
|
if isinstance(schedulers, _ParamScheduler):
|
|
schedulers = [schedulers]
|
|
for epoch in range(epochs):
|
|
for param_group, target in zip(optimizer.param_groups, targets):
|
|
assert_allclose(
|
|
target[epoch],
|
|
param_group[param_name],
|
|
msg='{} is wrong in epoch {}: expected {}, got {}'.format(
|
|
param_name, epoch, target[epoch],
|
|
param_group[param_name]),
|
|
atol=1e-5,
|
|
rtol=0)
|
|
if 'betas' in optimizer.defaults:
|
|
assert_allclose(
|
|
target[epoch],
|
|
param_group['betas'][0],
|
|
msg='{} is wrong in epoch {}: expected {}, got {}'.
|
|
format('betas_0', epoch, target[epoch],
|
|
param_group['betas'][0]),
|
|
atol=1e-5,
|
|
rtol=0)
|
|
[scheduler.step() for scheduler in schedulers]
|
|
|
|
def test_step_scheduler(self):
|
|
# momentum = 0.05 if epoch < 3
|
|
# momentum = 0.005 if 3 <= epoch < 6
|
|
# momentum = 0.0005 if 6 <= epoch < 9
|
|
# momentum = 0.00005 if epoch >=9
|
|
epochs = 10
|
|
single_targets = [0.05] * 3 + [0.005] * 3 + [0.0005] * 3 + [0.00005
|
|
] * 3
|
|
targets = [
|
|
single_targets, [x * self.layer2_mult for x in single_targets]
|
|
]
|
|
scheduler = StepMomentum(
|
|
self.optimizer, gamma=0.1, step_size=3, verbose=True)
|
|
self._test_scheduler_value(self.optimizer, scheduler, targets, epochs)
|
|
|
|
scheduler = StepMomentum(
|
|
self.optimizer_with_betas, gamma=0.1, step_size=3, verbose=True)
|
|
self._test_scheduler_value(self.optimizer_with_betas, scheduler,
|
|
targets, epochs)
|
|
|
|
def test_multi_step_scheduler(self):
|
|
# momentum = 0.05 if epoch < 2
|
|
# momentum = 0.005 if 2 <= epoch < 5
|
|
# momentum = 0.0005 if 5 <= epoch < 9
|
|
# momentum = 0.00005 if epoch >= 9
|
|
epochs = 10
|
|
single_targets = [0.05] * 2 + [0.005] * 3 + [0.0005] * 4 + [0.00005
|
|
] * 3
|
|
targets = [
|
|
single_targets, [x * self.layer2_mult for x in single_targets]
|
|
]
|
|
scheduler = MultiStepMomentum(
|
|
self.optimizer, gamma=0.1, milestones=[2, 5, 9])
|
|
self._test_scheduler_value(self.optimizer, scheduler, targets, epochs)
|
|
|
|
scheduler = MultiStepMomentum(
|
|
self.optimizer_with_betas, gamma=0.1, milestones=[2, 5, 9])
|
|
self._test_scheduler_value(self.optimizer_with_betas, scheduler,
|
|
targets, epochs)
|
|
|
|
def test_constant_scheduler(self):
|
|
# factor should between 0~1
|
|
with self.assertRaises(ValueError):
|
|
ConstantMomentum(self.optimizer, factor=99)
|
|
|
|
# momentum = 0.025 if epoch < 5
|
|
# momentum = 0.005 if 5 <= epoch
|
|
epochs = 10
|
|
single_targets = [0.025] * 4 + [0.05] * 6
|
|
targets = [
|
|
single_targets, [x * self.layer2_mult for x in single_targets]
|
|
]
|
|
scheduler = ConstantMomentum(self.optimizer, factor=1.0 / 2, end=5)
|
|
self._test_scheduler_value(self.optimizer, scheduler, targets, epochs)
|
|
|
|
scheduler = ConstantMomentum(
|
|
self.optimizer_with_betas, factor=1.0 / 2, end=5)
|
|
self._test_scheduler_value(self.optimizer_with_betas, scheduler,
|
|
targets, epochs)
|
|
|
|
def test_linear_scheduler(self):
|
|
with self.assertRaises(ValueError):
|
|
LinearMomentum(self.optimizer, start_factor=10, end=900)
|
|
with self.assertRaises(ValueError):
|
|
LinearMomentum(self.optimizer, start_factor=-1, end=900)
|
|
with self.assertRaises(ValueError):
|
|
LinearMomentum(self.optimizer, end_factor=1.001, end=900)
|
|
with self.assertRaises(ValueError):
|
|
LinearMomentum(self.optimizer, end_factor=-0.00001, end=900)
|
|
# momentum = 0.025 if epoch == 0
|
|
# momentum = 0.03125 if epoch == 1
|
|
# momentum = 0.0375 if epoch == 2
|
|
# momentum = 0.04375 if epoch == 3
|
|
# momentum = 0.005 if epoch >= 4
|
|
epochs = 10
|
|
start_factor = 1.0 / 2
|
|
iters = 4
|
|
interpolation = [
|
|
start_factor + i * (1 - start_factor) / iters for i in range(iters)
|
|
]
|
|
single_targets = [x * 0.05 for x in interpolation] + [0.05] * (
|
|
epochs - iters)
|
|
targets = [
|
|
single_targets, [x * self.layer2_mult for x in single_targets]
|
|
]
|
|
scheduler = LinearMomentum(
|
|
self.optimizer, start_factor=start_factor, end=iters + 1)
|
|
self._test_scheduler_value(self.optimizer, scheduler, targets, epochs)
|
|
|
|
scheduler = LinearMomentum(
|
|
self.optimizer_with_betas,
|
|
start_factor=start_factor,
|
|
end=iters + 1)
|
|
self._test_scheduler_value(self.optimizer_with_betas, scheduler,
|
|
targets, epochs)
|
|
|
|
def test_exp_scheduler(self):
|
|
epochs = 10
|
|
single_targets = [0.05 * (0.9**x) for x in range(epochs)]
|
|
targets = [
|
|
single_targets, [x * self.layer2_mult for x in single_targets]
|
|
]
|
|
scheduler = ExponentialMomentum(self.optimizer, gamma=0.9)
|
|
self._test_scheduler_value(self.optimizer, scheduler, targets, epochs)
|
|
|
|
scheduler = ExponentialMomentum(self.optimizer_with_betas, gamma=0.9)
|
|
self._test_scheduler_value(self.optimizer_with_betas, scheduler,
|
|
targets, epochs)
|
|
|
|
def test_cos_anneal_scheduler(self):
|
|
epochs = 12
|
|
t = 10
|
|
eta_min = 1e-10
|
|
single_targets = [
|
|
eta_min + (0.05 - eta_min) * (1 + math.cos(math.pi * x / t)) / 2
|
|
for x in range(epochs)
|
|
]
|
|
targets = [
|
|
single_targets, [x * self.layer2_mult for x in single_targets]
|
|
]
|
|
scheduler = CosineAnnealingMomentum(
|
|
self.optimizer, T_max=t, eta_min=eta_min)
|
|
self._test_scheduler_value(self.optimizer, scheduler, targets, epochs)
|
|
|
|
scheduler = CosineAnnealingMomentum(
|
|
self.optimizer_with_betas, T_max=t, eta_min=eta_min)
|
|
self._test_scheduler_value(self.optimizer_with_betas, scheduler,
|
|
targets, epochs)
|
|
|
|
# Test default `T_max`
|
|
scheduler = CosineAnnealingMomentum(
|
|
self.optimizer, begin=5, end=100, eta_min=eta_min)
|
|
self.assertEqual(scheduler.T_max, 100 - 5)
|
|
|
|
def test_poly_scheduler(self):
|
|
epochs = 10
|
|
power = 0.9
|
|
min_lr = 0.001
|
|
iters = 4
|
|
layer1_targets = [
|
|
min_lr + (0.05 - min_lr) * (1 - i / iters)**power
|
|
for i in range(iters)
|
|
] + [min_lr] * (
|
|
epochs - iters)
|
|
layer2_targets = [
|
|
min_lr + (0.05 * self.layer2_mult - min_lr) *
|
|
(1 - i / iters)**power for i in range(iters)
|
|
] + [min_lr] * (
|
|
epochs - iters)
|
|
targets = [layer1_targets, layer2_targets]
|
|
scheduler = PolyMomentum(
|
|
self.optimizer, power=power, eta_min=min_lr, end=iters + 1)
|
|
self._test_scheduler_value(
|
|
self.optimizer, scheduler, targets, epochs=10)
|
|
|
|
scheduler = PolyMomentum(
|
|
self.optimizer_with_betas,
|
|
power=power,
|
|
eta_min=min_lr,
|
|
end=iters + 1)
|
|
self._test_scheduler_value(
|
|
self.optimizer_with_betas, scheduler, targets, epochs=10)
|
|
|
|
def test_cosine_restart_scheduler(self):
|
|
with self.assertRaises(AssertionError):
|
|
CosineRestartMomentum(
|
|
self.optimizer,
|
|
periods=[4, 5],
|
|
restart_weights=[1, 0.5],
|
|
eta_min=0,
|
|
eta_min_ratio=0.1)
|
|
with self.assertRaises(AssertionError):
|
|
CosineRestartMomentum(
|
|
self.optimizer,
|
|
periods=[4, 5],
|
|
restart_weights=[1, 0.5, 0.0],
|
|
eta_min=0)
|
|
single_targets = [
|
|
0.05, 0.0426776, 0.025, 0.00732233, 0.025, 0.022612712, 0.01636271,
|
|
0.0086372, 0.0023872, 0.0023872
|
|
]
|
|
targets = [
|
|
single_targets, [t * self.layer2_mult for t in single_targets]
|
|
]
|
|
scheduler = CosineRestartMomentum(
|
|
self.optimizer,
|
|
periods=[4, 5],
|
|
restart_weights=[1, 0.5],
|
|
eta_min=0)
|
|
self._test_scheduler_value(
|
|
self.optimizer, scheduler, targets, epochs=10)
|
|
|
|
scheduler = CosineRestartMomentum(
|
|
self.optimizer_with_betas,
|
|
periods=[4, 5],
|
|
restart_weights=[1, 0.5],
|
|
eta_min=0)
|
|
self._test_scheduler_value(
|
|
self.optimizer_with_betas, scheduler, targets, epochs=10)
|
|
|
|
def _check_scheduler_state_dict(self, construct, construct2, epochs=10):
|
|
scheduler = construct()
|
|
for _ in range(epochs):
|
|
scheduler.optimizer.step()
|
|
scheduler.step()
|
|
scheduler_copy = construct2()
|
|
scheduler_copy.load_state_dict(scheduler.state_dict())
|
|
for key in scheduler.__dict__.keys():
|
|
if key != 'optimizer':
|
|
self.assertEqual(scheduler.__dict__[key],
|
|
scheduler_copy.__dict__[key])
|
|
self.assertEqual(scheduler.get_last_value(),
|
|
scheduler_copy.get_last_value())
|
|
|
|
def test_step_scheduler_state_dict(self):
|
|
self._check_scheduler_state_dict(
|
|
lambda: StepMomentum(self.optimizer, gamma=0.1, step_size=3),
|
|
lambda: StepMomentum(self.optimizer, gamma=0.01 / 2, step_size=1))
|
|
|
|
def test_multi_step_scheduler_state_dict(self):
|
|
self._check_scheduler_state_dict(
|
|
lambda: MultiStepMomentum(
|
|
self.optimizer, gamma=0.1, milestones=[2, 5, 9]),
|
|
lambda: MultiStepMomentum(
|
|
self.optimizer, gamma=0.01, milestones=[1, 4, 6]))
|
|
|
|
def test_exp_scheduler_state_dict(self):
|
|
self._check_scheduler_state_dict(
|
|
lambda: ExponentialMomentum(self.optimizer, gamma=0.1),
|
|
lambda: ExponentialMomentum(self.optimizer, gamma=0.01))
|
|
|
|
def test_cosine_scheduler_state_dict(self):
|
|
epochs = 10
|
|
eta_min = 1e-10
|
|
self._check_scheduler_state_dict(
|
|
lambda: CosineAnnealingMomentum(
|
|
self.optimizer, T_max=epochs, eta_min=eta_min),
|
|
lambda: CosineAnnealingMomentum(
|
|
self.optimizer, T_max=epochs // 2, eta_min=eta_min / 2),
|
|
epochs=epochs)
|
|
|
|
def test_linear_scheduler_state_dict(self):
|
|
epochs = 10
|
|
self._check_scheduler_state_dict(
|
|
lambda: LinearMomentum(self.optimizer, start_factor=1 / 3),
|
|
lambda: LinearMomentum(
|
|
self.optimizer, start_factor=0, end_factor=0.3),
|
|
epochs=epochs)
|
|
|
|
def test_poly_scheduler_state_dict(self):
|
|
self._check_scheduler_state_dict(
|
|
lambda: PolyMomentum(self.optimizer, power=0.5, eta_min=0.001),
|
|
lambda: PolyMomentum(self.optimizer, power=0.8, eta_min=0.002),
|
|
epochs=10)
|
|
|
|
def test_cosine_restart_scheduler_state_dict(self):
|
|
self._check_scheduler_state_dict(
|
|
lambda: CosineRestartMomentum(
|
|
self.optimizer,
|
|
periods=[4, 5],
|
|
restart_weights=[1, 0.5],
|
|
eta_min=0),
|
|
lambda: CosineRestartMomentum(
|
|
self.optimizer,
|
|
periods=[4, 6],
|
|
restart_weights=[1, 0.5],
|
|
eta_min=0),
|
|
epochs=10)
|
|
|
|
def test_multi_scheduler_without_overlap_linear_multi_step(self):
|
|
# use Linear in the first 5 epochs and then use MultiStep
|
|
epochs = 12
|
|
single_targets = [0.025, 0.03125, 0.0375, 0.04375
|
|
] + [0.05] * 4 + [0.005] * 3 + [0.0005] * 1
|
|
targets = [
|
|
single_targets, [x * self.layer2_mult for x in single_targets]
|
|
]
|
|
scheduler1 = LinearMomentum(
|
|
self.optimizer, start_factor=1 / 2, begin=0, end=5)
|
|
scheduler2 = MultiStepMomentum(
|
|
self.optimizer, gamma=0.1, milestones=[3, 6], begin=5, end=12)
|
|
self._test_scheduler_value(self.optimizer, [scheduler1, scheduler2],
|
|
targets, epochs)
|
|
|
|
def test_multi_scheduler_without_overlap_exp_cosine(self):
|
|
# use Exp in the first 5 epochs and then use Cosine
|
|
epochs = 10
|
|
single_targets1 = [0.05 * (0.9**x) for x in range(5)]
|
|
scheduler1 = ExponentialMomentum(
|
|
self.optimizer, gamma=0.9, begin=0, end=5)
|
|
|
|
eta_min = 1e-10
|
|
single_targets2 = [
|
|
eta_min + (single_targets1[-1] - eta_min) *
|
|
(1 + math.cos(math.pi * x / 5)) / 2 for x in range(5)
|
|
]
|
|
single_targets = single_targets1 + single_targets2
|
|
targets = [
|
|
single_targets, [x * self.layer2_mult for x in single_targets]
|
|
]
|
|
scheduler2 = CosineAnnealingMomentum(
|
|
self.optimizer, T_max=5, eta_min=eta_min, begin=5, end=10)
|
|
|
|
self._test_scheduler_value(self.optimizer, [scheduler1, scheduler2],
|
|
targets, epochs)
|
|
|
|
def test_multi_scheduler_with_overlap(self):
|
|
# use Linear at first 5 epochs together with MultiStep
|
|
epochs = 10
|
|
single_targets = [0.025, 0.03125, 0.0375, 0.004375
|
|
] + [0.005] * 2 + [0.0005] * 3 + [0.00005] * 1
|
|
targets = [
|
|
single_targets, [x * self.layer2_mult for x in single_targets]
|
|
]
|
|
scheduler1 = LinearMomentum(
|
|
self.optimizer, start_factor=1 / 2, begin=0, end=5)
|
|
scheduler2 = MultiStepMomentum(
|
|
self.optimizer, gamma=0.1, milestones=[3, 6, 9])
|
|
self._test_scheduler_value(self.optimizer, [scheduler1, scheduler2],
|
|
targets, epochs)
|
|
|
|
def test_multi_scheduler_with_gap(self):
|
|
# use Exp in the first 5 epochs and the last 5 epochs use Cosine
|
|
# no scheduler in the middle 5 epochs
|
|
epochs = 15
|
|
single_targets1 = [0.05 * (0.9**x) for x in range(5)]
|
|
scheduler1 = ExponentialMomentum(
|
|
self.optimizer, gamma=0.9, begin=0, end=5)
|
|
|
|
eta_min = 1e-10
|
|
single_targets2 = [
|
|
eta_min + (single_targets1[-1] - eta_min) *
|
|
(1 + math.cos(math.pi * x / 5)) / 2 for x in range(5)
|
|
]
|
|
single_targets = single_targets1 + [single_targets1[-1]
|
|
] * 5 + single_targets2
|
|
targets = [
|
|
single_targets, [x * self.layer2_mult for x in single_targets]
|
|
]
|
|
scheduler2 = CosineAnnealingMomentum(
|
|
self.optimizer, T_max=5, eta_min=eta_min, begin=10, end=15)
|
|
|
|
self._test_scheduler_value(self.optimizer, [scheduler1, scheduler2],
|
|
targets, epochs)
|