wehub-resource-sync
1947 lines
64 KiB
Python
1947 lines
64 KiB
Python
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
import os
|
|
import random
|
|
import unittest
|
|
from functools import partial
|
|
|
|
import numpy as np
|
|
from op_test import (
|
|
OpTest,
|
|
get_device,
|
|
get_device_place,
|
|
get_devices,
|
|
is_custom_device,
|
|
)
|
|
|
|
import paddle
|
|
from paddle import base, nn
|
|
from paddle.framework import core
|
|
|
|
|
|
def adamw_step(inputs, attributes):
|
|
param = inputs['Param']
|
|
grad = inputs['Grad']
|
|
moment1 = inputs['Moment1']
|
|
moment2 = inputs['Moment2']
|
|
moment2_max = inputs['Moment2Max']
|
|
lr = inputs['LearningRate']
|
|
beta1_pow = inputs['Beta1Pow']
|
|
beta2_pow = inputs['Beta2Pow']
|
|
|
|
epsilon = attributes['epsilon']
|
|
|
|
if 'lr_ratio' in attributes:
|
|
lr = lr * attributes['lr_ratio']
|
|
|
|
if attributes["with_decay"]:
|
|
coeff = attributes["coeff"]
|
|
decay = 1.0 - lr * coeff
|
|
param2 = param * decay
|
|
param = param2.copy()
|
|
|
|
if 'beta1' in attributes:
|
|
beta1 = attributes['beta1']
|
|
else:
|
|
beta1 = inputs['Beta1Tensor'][0]
|
|
if 'beta2' in attributes:
|
|
beta2 = attributes['beta2']
|
|
else:
|
|
beta2 = inputs['Beta2Tensor'][0]
|
|
|
|
amsgrad = attributes['amsgrad']
|
|
|
|
moment1_out = beta1 * moment1 + (1 - beta1) * grad
|
|
moment2_out = beta2 * moment2 + (1 - beta2) * np.square(grad)
|
|
|
|
if amsgrad:
|
|
moment2_max_out = np.maximum(moment2_out, moment2_max)
|
|
denom = (np.sqrt(moment2_max_out) / np.sqrt(1.0 - beta2_pow)) + epsilon
|
|
else:
|
|
moment2_max_out = np.empty_like(moment2_out)
|
|
denom = (np.sqrt(moment2_out) / np.sqrt(1.0 - beta2_pow)) + epsilon
|
|
|
|
param_out = param + ((moment1_out / denom) * (-(lr / (1.0 - beta1_pow))))
|
|
return param_out, moment1_out, moment2_out, moment2_max_out
|
|
|
|
|
|
def adamw_wrapper(
|
|
param,
|
|
grad,
|
|
lr,
|
|
moment1,
|
|
moment2,
|
|
moment2_max,
|
|
beta1_pow,
|
|
beta2_pow,
|
|
master_weight=None,
|
|
found_inf=None,
|
|
beta1=0.78,
|
|
beta2=0.836,
|
|
epsilon=1e-4,
|
|
lr_ratio=1.0,
|
|
weight_decay=0.01,
|
|
with_decay=True,
|
|
lazy_mode=False,
|
|
amsgrad=False,
|
|
):
|
|
_, _, _, _, _, _, _ = paddle._C_ops.adamw_(
|
|
param,
|
|
grad,
|
|
lr,
|
|
moment1,
|
|
moment2,
|
|
moment2_max,
|
|
beta1_pow,
|
|
beta2_pow,
|
|
master_weight,
|
|
found_inf,
|
|
beta1,
|
|
beta2,
|
|
epsilon,
|
|
lr_ratio,
|
|
weight_decay,
|
|
with_decay,
|
|
lazy_mode,
|
|
1000,
|
|
False,
|
|
False,
|
|
amsgrad,
|
|
)
|
|
|
|
|
|
class TestAdamW(OpTest):
|
|
def set_amsgrad(self):
|
|
self.amsgrad = False
|
|
# no check `Moment2MaxOut` with amsgrad is False
|
|
self.no_check_set = ['Moment2MaxOut']
|
|
|
|
def setUp(self):
|
|
'''Test AdamW Op with supplied attributes'''
|
|
self.op_type = "adamw"
|
|
self.python_api = adamw_wrapper
|
|
self.python_out_sig = ['Out']
|
|
param = np.random.uniform(-1, 1, (102, 105)).astype("float32")
|
|
grad = np.random.uniform(-1, 1, (102, 105)).astype("float32")
|
|
moment1 = np.random.uniform(-1, 1, (102, 105)).astype("float32")
|
|
# The second moment is positive
|
|
moment2 = np.random.random((102, 105)).astype("float32")
|
|
moment2_max = np.zeros((102, 105)).astype("float32")
|
|
|
|
learning_rate = 0.004
|
|
beta1 = 0.78
|
|
beta2 = 0.836
|
|
epsilon = 1e-4
|
|
beta1_pow = beta1**10
|
|
beta2_pow = beta2**10
|
|
self.set_amsgrad()
|
|
|
|
self.inputs = {
|
|
'Param': param,
|
|
'Grad': grad,
|
|
'Moment1': moment1,
|
|
'Moment2': moment2,
|
|
'Moment2Max': moment2_max,
|
|
'LearningRate': np.array([learning_rate]).astype("float64"),
|
|
'Beta1Pow': np.array([beta1_pow]).astype("float32"),
|
|
'Beta2Pow': np.array([beta2_pow]).astype("float32"),
|
|
}
|
|
|
|
self.attrs = {
|
|
'epsilon': epsilon,
|
|
'beta1': beta1,
|
|
'beta2': beta2,
|
|
"coeff": 0.5,
|
|
"with_decay": True,
|
|
"amsgrad": self.amsgrad,
|
|
}
|
|
|
|
param_out, moment1_out, moment2_out, moment2_max_out = adamw_step(
|
|
self.inputs, self.attrs
|
|
)
|
|
|
|
self.outputs = {
|
|
'Moment1Out': moment1_out,
|
|
'Moment2Out': moment2_out,
|
|
'Moment2MaxOut': moment2_max_out,
|
|
'ParamOut': param_out,
|
|
'Beta1PowOut': np.array([beta1_pow]).astype("float32") * beta1,
|
|
'Beta2PowOut': np.array([beta2_pow]).astype("float32") * beta2,
|
|
}
|
|
|
|
def test_check_output(self):
|
|
self.check_output(
|
|
no_check_set=self.no_check_set, check_pir=True, rtol=2e-4
|
|
)
|
|
|
|
|
|
class TestAdamWAMSGrad(TestAdamW):
|
|
def set_amsgrad(self):
|
|
# xpu not support `amsgrad`
|
|
if core.is_compiled_with_xpu():
|
|
self.amsgrad = False
|
|
self.no_check_set = ['Moment2MaxOut']
|
|
else:
|
|
self.amsgrad = True
|
|
self.no_check_set = None
|
|
|
|
|
|
@unittest.skipIf(
|
|
not (
|
|
(core.is_compiled_with_cuda() or is_custom_device())
|
|
or core.is_compiled_with_xpu()
|
|
),
|
|
"core is not compiled with CUDA nor XPU",
|
|
)
|
|
class TestAdamW2(OpTest):
|
|
def set_amsgrad(self):
|
|
self.amsgrad = False
|
|
self.no_check_set = ['Moment2MaxOut']
|
|
|
|
def setUp(self):
|
|
'''Test AdamW Op with supplied attributes'''
|
|
self.op_type = "adamw"
|
|
self.python_api = adamw_wrapper
|
|
self.python_out_sig = ['Out']
|
|
param = np.random.uniform(-1, 1, (2, 2)).astype("float32")
|
|
grad = np.random.uniform(-1, 1, (2, 2)).astype("float32")
|
|
moment1 = np.random.uniform(-1, 1, (2, 2)).astype("float32")
|
|
# The second moment is positive
|
|
moment2 = np.random.random((2, 2)).astype("float32")
|
|
moment2_max = np.zeros((2, 2)).astype("float32")
|
|
|
|
learning_rate = 0.004
|
|
beta1 = 0.78
|
|
beta2 = 0.836
|
|
epsilon = 1e-4
|
|
beta1_pow = beta1**10
|
|
beta2_pow = beta2**10
|
|
self.set_amsgrad()
|
|
|
|
self.inputs = {
|
|
'Param': param,
|
|
'Grad': grad,
|
|
'Moment1': moment1,
|
|
'Moment2': moment2,
|
|
'Moment2Max': moment2_max,
|
|
'LearningRate': np.array([learning_rate]).astype("float64"),
|
|
'Beta1Pow': np.array([beta1_pow]).astype("float32"),
|
|
'Beta2Pow': np.array([beta2_pow]).astype("float32"),
|
|
}
|
|
|
|
self.attrs = {
|
|
'epsilon': epsilon,
|
|
'beta1': beta1,
|
|
'beta2': beta2,
|
|
"lr_ratio": 0.1,
|
|
"coeff": 0.5,
|
|
"with_decay": True,
|
|
"amsgrad": self.amsgrad,
|
|
}
|
|
|
|
param_out, moment1_out, moment2_out, moment2_max_out = adamw_step(
|
|
self.inputs, self.attrs
|
|
)
|
|
|
|
self.outputs = {
|
|
'Moment1Out': moment1_out,
|
|
'Moment2Out': moment2_out,
|
|
'Moment2MaxOut': moment2_max_out,
|
|
'ParamOut': param_out,
|
|
'Beta1PowOut': np.array([beta1_pow]).astype("float32") * beta1,
|
|
'Beta2PowOut': np.array([beta2_pow]).astype("float32") * beta2,
|
|
}
|
|
|
|
def test_check_output(self):
|
|
self.check_output_with_place(
|
|
no_check_set=self.no_check_set,
|
|
place=(
|
|
get_device_place()
|
|
if not core.is_compiled_with_xpu()
|
|
else core.XPUPlace(0)
|
|
),
|
|
check_pir=True,
|
|
)
|
|
|
|
|
|
class TestAdamW2AMSGrad(TestAdamW2):
|
|
def set_amsgrad(self):
|
|
# xpu not support `amsgrad`
|
|
if core.is_compiled_with_xpu():
|
|
self.amsgrad = False
|
|
self.no_check_set = ['Moment2MaxOut']
|
|
else:
|
|
self.amsgrad = True
|
|
self.no_check_set = None
|
|
|
|
|
|
class TestAdamWOp(unittest.TestCase):
|
|
def setUp(self):
|
|
self.amsgrad = False
|
|
|
|
def test_adamw_op_dygraph(self):
|
|
paddle.disable_static()
|
|
value = np.arange(26).reshape(2, 13).astype("float32")
|
|
a = paddle.to_tensor(value)
|
|
linear = paddle.nn.Linear(13, 5)
|
|
adam = paddle.optimizer.AdamW(
|
|
learning_rate=0.01,
|
|
parameters=linear.parameters(),
|
|
apply_decay_param_fun=lambda name: True,
|
|
weight_decay=0.01,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
for _ in range(2):
|
|
out = linear(a)
|
|
out.backward()
|
|
adam.step()
|
|
adam.clear_gradients()
|
|
|
|
def test_adamw_op(self):
|
|
paddle.enable_static()
|
|
place = base.CPUPlace()
|
|
shape = [2, 3, 8, 8]
|
|
exe = base.Executor(place)
|
|
train_prog = base.Program()
|
|
startup = base.Program()
|
|
with (
|
|
base.program_guard(train_prog, startup),
|
|
base.unique_name.guard(),
|
|
):
|
|
data = paddle.static.data(name="data", shape=shape)
|
|
conv = paddle.nn.Conv2D(
|
|
in_channels=3,
|
|
out_channels=8,
|
|
kernel_size=3,
|
|
)(data)
|
|
loss = paddle.mean(conv)
|
|
if paddle.framework.in_pir_mode():
|
|
beta1 = paddle.pir.core.create_persistable_value(
|
|
shape=[1],
|
|
dtype='float32',
|
|
initializer=paddle.nn.initializer.Constant(0.85),
|
|
)
|
|
beta2 = paddle.pir.core.create_persistable_value(
|
|
shape=[1],
|
|
dtype='float32',
|
|
initializer=paddle.nn.initializer.Constant(0.95),
|
|
)
|
|
else:
|
|
beta1 = paddle.static.create_global_var(
|
|
shape=[1], value=0.85, dtype='float32', persistable=True
|
|
)
|
|
beta2 = paddle.static.create_global_var(
|
|
shape=[1], value=0.95, dtype='float32', persistable=True
|
|
)
|
|
betas = [beta1, beta2]
|
|
opt = paddle.optimizer.AdamW(
|
|
learning_rate=1e-5,
|
|
beta1=beta1,
|
|
beta2=beta2,
|
|
weight_decay=0.01,
|
|
epsilon=1e-8,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
opt.minimize(loss)
|
|
|
|
exe.run(startup)
|
|
data_np = np.random.random(shape).astype('float32')
|
|
rets = exe.run(train_prog, feed={"data": data_np}, fetch_list=[loss])
|
|
assert rets[0] is not None
|
|
paddle.disable_static()
|
|
|
|
def test_adamw_op_dygraph_bypassing_step(self):
|
|
paddle.disable_static()
|
|
value = np.arange(26).reshape(2, 13).astype("float32")
|
|
a = paddle.to_tensor(value)
|
|
linear = paddle.nn.Linear(13, 5)
|
|
adam = paddle.optimizer.AdamW(
|
|
learning_rate=0.01,
|
|
parameters=linear.parameters(),
|
|
apply_decay_param_fun=lambda name: True,
|
|
weight_decay=0.01,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
os.environ["FLAGS_shard_bypass_dygraph_optimizer"] = "1"
|
|
for _ in range(2):
|
|
out = linear(a)
|
|
out.backward()
|
|
adam.step()
|
|
adam.clear_gradients()
|
|
|
|
def test_adamw_op_coverage(self):
|
|
paddle.disable_static()
|
|
value = np.arange(26).reshape(2, 13).astype("float32")
|
|
a = paddle.to_tensor(value)
|
|
linear = paddle.nn.Linear(13, 5)
|
|
adam = paddle.optimizer.AdamW(
|
|
learning_rate=0.0,
|
|
parameters=linear.parameters(),
|
|
apply_decay_param_fun=lambda name: True,
|
|
weight_decay=0.01,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
assert adam.__str__() is not None
|
|
|
|
def test_pir_adam_op(self):
|
|
with paddle.pir_utils.IrGuard():
|
|
place = base.CPUPlace()
|
|
shape = [2, 3, 8, 8]
|
|
exe = base.Executor(place)
|
|
train_prog = paddle.static.Program()
|
|
startup = paddle.static.Program()
|
|
with (
|
|
paddle.static.program_guard(train_prog, startup),
|
|
base.unique_name.guard(),
|
|
):
|
|
data = paddle.static.data(name="data", shape=shape)
|
|
conv_layer = paddle.nn.Conv2D(3, 8, 3)
|
|
conv = conv_layer(data)
|
|
loss = paddle.mean(conv)
|
|
|
|
beta1 = paddle.pir.core.create_parameter(
|
|
'float32',
|
|
[1],
|
|
initializer=paddle.nn.initializer.Constant(0.85),
|
|
)
|
|
beta2 = paddle.pir.core.create_parameter(
|
|
'float32',
|
|
[1],
|
|
initializer=paddle.nn.initializer.Constant(0.95),
|
|
)
|
|
betas = [beta1, beta2]
|
|
opt = paddle.optimizer.AdamW(
|
|
learning_rate=1e-5,
|
|
beta1=beta1,
|
|
beta2=beta2,
|
|
weight_decay=0.01,
|
|
epsilon=1e-8,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
opt.minimize(loss)
|
|
|
|
exe.run(startup)
|
|
data_np = np.random.random(shape).astype('float32')
|
|
rets = exe.run(
|
|
train_prog, feed={"data": data_np}, fetch_list=[loss]
|
|
)
|
|
assert rets[0] is not None
|
|
|
|
def test_adamw_op_invalid_input(self):
|
|
paddle.disable_static()
|
|
linear = paddle.nn.Linear(10, 10)
|
|
with self.assertRaises(ValueError):
|
|
adam = paddle.optimizer.AdamW(
|
|
0.1,
|
|
beta1=-1,
|
|
parameters=linear.parameters(),
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
with self.assertRaises(ValueError):
|
|
adam = paddle.optimizer.AdamW(
|
|
0.1,
|
|
beta2=-1,
|
|
parameters=linear.parameters(),
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
with self.assertRaises(ValueError):
|
|
adam = paddle.optimizer.AdamW(
|
|
0.1,
|
|
epsilon=-1,
|
|
parameters=linear.parameters(),
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
|
|
class TestAdamWOpAMSGrad(TestAdamWOp):
|
|
def setUp(self):
|
|
# xpu not support `amsgrad`
|
|
if core.is_compiled_with_xpu():
|
|
self.amsgrad = False
|
|
self.no_check_set = ['Moment2MaxOut']
|
|
else:
|
|
self.amsgrad = True
|
|
self.no_check_set = None
|
|
|
|
|
|
class TestAdamWOpGroup(TestAdamWOp):
|
|
def test_adamw_op_dygraph(self):
|
|
paddle.disable_static()
|
|
value = np.arange(26).reshape(2, 13).astype("float32")
|
|
a = paddle.to_tensor(value)
|
|
linear_1 = paddle.nn.Linear(13, 5)
|
|
linear_2 = paddle.nn.Linear(5, 3)
|
|
adam = paddle.optimizer.AdamW(
|
|
learning_rate=0.01,
|
|
parameters=[
|
|
{'params': linear_1.parameters()},
|
|
{'params': linear_2.parameters(), 'weight_decay': 0.001},
|
|
],
|
|
apply_decay_param_fun=lambda name: True,
|
|
weight_decay=0.01,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
for _ in range(2):
|
|
out = linear_1(a)
|
|
out = linear_2(out)
|
|
out.backward()
|
|
adam.step()
|
|
adam.clear_gradients()
|
|
|
|
def test_adamw_op_dygraph_bypassing_step(self):
|
|
paddle.disable_static()
|
|
value = np.arange(26).reshape(2, 13).astype("float32")
|
|
a = paddle.to_tensor(value)
|
|
linear_1 = paddle.nn.Linear(13, 5)
|
|
linear_2 = paddle.nn.Linear(5, 3)
|
|
adam = paddle.optimizer.AdamW(
|
|
learning_rate=0.01,
|
|
parameters=[
|
|
{'params': linear_1.parameters()},
|
|
{'params': linear_2.parameters(), 'weight_decay': 0.001},
|
|
],
|
|
apply_decay_param_fun=lambda name: True,
|
|
weight_decay=0.01,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
os.environ["FLAGS_shard_bypass_dygraph_optimizer"] = "1"
|
|
for _ in range(2):
|
|
out = linear_1(a)
|
|
out = linear_2(out)
|
|
out.backward()
|
|
adam.step()
|
|
adam.clear_gradients()
|
|
|
|
|
|
class TestAdamWOpGroupAMSGrad(TestAdamWOpGroup):
|
|
def setUp(self):
|
|
# xpu not support `amsgrad`
|
|
if core.is_compiled_with_xpu():
|
|
self.amsgrad = False
|
|
self.no_check_set = ['Moment2MaxOut']
|
|
else:
|
|
self.amsgrad = True
|
|
self.no_check_set = None
|
|
|
|
|
|
class TestAdamWOpMultiPrecisionWithMainGrad(unittest.TestCase):
|
|
def setUp(self):
|
|
self.amsgrad = False
|
|
|
|
def _test_adamw_op_dygraph_place_amp_with_maingrad(
|
|
self, place, shape, use_main_grad
|
|
):
|
|
paddle.disable_static()
|
|
paddle.seed(10)
|
|
paddle.set_device(place)
|
|
|
|
found_inf = None
|
|
|
|
_weight_decay = 0.1
|
|
with_decay = True
|
|
_lazy_mode = False
|
|
find_master = True
|
|
|
|
_epsilon = 1e-8
|
|
|
|
_beta1 = 0.9
|
|
_beta2 = 0.99
|
|
lr_ratio_ = 1.0
|
|
|
|
lr_rate = 1e-8
|
|
|
|
param = paddle.randn(shape).astype(paddle.bfloat16)
|
|
master_weight = param.astype(paddle.float32)
|
|
grad = paddle.randn(shape).astype(paddle.bfloat16)
|
|
main_grad = grad.astype(paddle.float32)
|
|
moment1 = paddle.randn(shape).astype(paddle.float32)
|
|
moment2 = paddle.randn(shape).astype(paddle.float32).abs()
|
|
moment2_max = paddle.zeros(shape).astype(paddle.float32)
|
|
lr = paddle.zeros([1]).astype(paddle.float64)
|
|
lr[0] = lr_rate
|
|
beta1_pow_acc = paddle.ones([1]).astype(paddle.float32)
|
|
beta1_pow_acc[0] = _beta1**10
|
|
beta2_pow_acc = paddle.ones([1]).astype(paddle.float32)
|
|
beta2_pow_acc[0] = _beta2**10
|
|
|
|
ref_param = param.astype(paddle.float32).clone().detach()
|
|
ref_beta1_pow_acc = (
|
|
beta1_pow_acc.astype(paddle.float32).clone().detach()
|
|
)
|
|
ref_beta2_pow_acc = (
|
|
beta2_pow_acc.astype(paddle.float32).clone().detach()
|
|
)
|
|
ref_moment_1 = moment1.astype(paddle.float32).clone().detach()
|
|
ref_moment_2 = moment2.astype(paddle.float32).clone().detach()
|
|
ref_moment_2_max = moment2_max.astype(paddle.float32).clone().detach()
|
|
|
|
# reference code
|
|
_, _, _, _, _, _, _ = paddle._C_ops.adamw_(
|
|
ref_param,
|
|
main_grad,
|
|
lr,
|
|
ref_moment_1,
|
|
ref_moment_2,
|
|
ref_moment_2_max,
|
|
ref_beta1_pow_acc,
|
|
ref_beta2_pow_acc,
|
|
master_weight,
|
|
found_inf,
|
|
_beta1,
|
|
_beta2,
|
|
_epsilon,
|
|
lr_ratio_,
|
|
_weight_decay,
|
|
with_decay,
|
|
_lazy_mode,
|
|
1000,
|
|
False,
|
|
False,
|
|
self.amsgrad,
|
|
)
|
|
|
|
if use_main_grad:
|
|
_, _, _, _, _, _, _ = paddle._C_ops.adamw_(
|
|
param,
|
|
main_grad,
|
|
lr,
|
|
moment1,
|
|
moment2,
|
|
moment2_max,
|
|
beta1_pow_acc,
|
|
beta2_pow_acc,
|
|
master_weight,
|
|
found_inf,
|
|
_beta1,
|
|
_beta2,
|
|
_epsilon,
|
|
lr_ratio_,
|
|
_weight_decay,
|
|
with_decay,
|
|
_lazy_mode,
|
|
1000,
|
|
find_master,
|
|
False,
|
|
self.amsgrad,
|
|
)
|
|
np.testing.assert_allclose(
|
|
param.astype("float32").numpy(), ref_param.numpy(), rtol=1e-2
|
|
)
|
|
|
|
if self.amsgrad:
|
|
np.testing.assert_allclose(
|
|
master_weight.numpy(), ref_param.numpy(), rtol=1e-4
|
|
)
|
|
else:
|
|
np.testing.assert_allclose(
|
|
master_weight.numpy(), ref_param.numpy(), rtol=1e-6
|
|
)
|
|
|
|
else:
|
|
_, _, _, _, _, _, _ = paddle._C_ops.adamw_(
|
|
param,
|
|
grad,
|
|
lr,
|
|
moment1,
|
|
moment2,
|
|
moment2_max,
|
|
beta1_pow_acc,
|
|
beta2_pow_acc,
|
|
master_weight,
|
|
found_inf,
|
|
_beta1,
|
|
_beta2,
|
|
_epsilon,
|
|
lr_ratio_,
|
|
_weight_decay,
|
|
with_decay,
|
|
_lazy_mode,
|
|
1000,
|
|
find_master,
|
|
False,
|
|
self.amsgrad,
|
|
)
|
|
np.testing.assert_allclose(
|
|
param.astype("float32").numpy(), ref_param.numpy(), rtol=1e-2
|
|
)
|
|
|
|
if self.amsgrad:
|
|
np.testing.assert_allclose(
|
|
master_weight.numpy(), ref_param.numpy(), rtol=1e-4
|
|
)
|
|
else:
|
|
np.testing.assert_allclose(
|
|
master_weight.numpy(), ref_param.numpy(), rtol=1e-6
|
|
)
|
|
|
|
def _get_places(self):
|
|
places = []
|
|
if paddle.is_compiled_with_cuda() or is_custom_device():
|
|
places.append(get_device())
|
|
if paddle.is_compiled_with_xpu():
|
|
places.append('xpu')
|
|
return places
|
|
|
|
def test_main(self):
|
|
for _ in range(10):
|
|
shape = paddle.randint(1, 1024, [2])
|
|
for place in self._get_places():
|
|
use_main_grad_list = [True, False]
|
|
for use_main_grad in use_main_grad_list:
|
|
self._test_adamw_op_dygraph_place_amp_with_maingrad(
|
|
place, shape, use_main_grad
|
|
)
|
|
|
|
|
|
class TestAdamWOpMultiPrecisionWithMainGradAMSGrad(
|
|
TestAdamWOpMultiPrecisionWithMainGrad
|
|
):
|
|
def setUp(self):
|
|
# xpu not support `amsgrad`
|
|
if core.is_compiled_with_xpu():
|
|
self.amsgrad = False
|
|
self.no_check_set = ['Moment2MaxOut']
|
|
else:
|
|
self.amsgrad = True
|
|
self.no_check_set = None
|
|
|
|
|
|
class TestAdamWOpMultiPrecision(unittest.TestCase):
|
|
def setUp(self):
|
|
self.amsgrad = False
|
|
|
|
def _test_adamw_op_dygraph_place_amp(self, place, use_amp=False):
|
|
paddle.disable_static()
|
|
paddle.seed(10)
|
|
paddle.set_device(place)
|
|
|
|
input = paddle.randn((5, 5))
|
|
|
|
model = paddle.nn.Linear(5, 5)
|
|
|
|
optimizer = paddle.optimizer.AdamW(
|
|
parameters=[
|
|
{
|
|
'params': model.parameters(),
|
|
'weight_decay': 0.001,
|
|
'beta1': 0.1,
|
|
'beta2': 0.99,
|
|
}
|
|
],
|
|
multi_precision=use_amp,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
for idx in range(2):
|
|
if (place == get_device() or place == 'xpu') and use_amp:
|
|
model = paddle.amp.decorate(models=model, level='O2')
|
|
scaler = paddle.amp.GradScaler(init_loss_scaling=1024)
|
|
|
|
if (place == get_device() or place == 'xpu') and use_amp:
|
|
with paddle.amp.auto_cast(level='O2'):
|
|
output = model(input)
|
|
loss = paddle.mean(output)
|
|
scaled = scaler.scale(loss)
|
|
scaled.backward()
|
|
scaler.step(optimizer)
|
|
optimizer.clear_grad()
|
|
else:
|
|
output = model(input)
|
|
loss = paddle.mean(output)
|
|
loss.backward()
|
|
optimizer.step()
|
|
optimizer.clear_grad()
|
|
|
|
def _get_places(self):
|
|
places = get_devices()
|
|
if paddle.is_compiled_with_xpu():
|
|
places.append('xpu')
|
|
return places
|
|
|
|
def test_main(self):
|
|
for place in self._get_places():
|
|
use_amp_list = [True, False]
|
|
for use_amp in use_amp_list:
|
|
self._test_adamw_op_dygraph_place_amp(place, use_amp)
|
|
|
|
|
|
class TestAdamWOpMultiPrecisionAMSGrad(TestAdamWOpMultiPrecision):
|
|
def setUp(self):
|
|
# xpu not support `amsgrad`
|
|
if core.is_compiled_with_xpu():
|
|
self.amsgrad = False
|
|
self.no_check_set = ['Moment2MaxOut']
|
|
else:
|
|
self.amsgrad = True
|
|
self.no_check_set = None
|
|
|
|
|
|
class TestAdamWOpError(unittest.TestCase):
|
|
def setUp(self):
|
|
self.amsgrad = False
|
|
|
|
def test_api_errors(self):
|
|
def test_parameters_dtype1():
|
|
adam = paddle.optimizer.AdamW(
|
|
learning_rate=0.01,
|
|
parameters=paddle.randn((5, 5)),
|
|
weight_decay=0.1,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
def test_parameters_dtype2():
|
|
linear = paddle.nn.Linear(13, 5)
|
|
adam = paddle.optimizer.AdamW(
|
|
learning_rate=0.01,
|
|
parameters={'params': linear.parameters()},
|
|
weight_decay=0.1,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
def test_parameters_dtype3():
|
|
adam = paddle.optimizer.AdamW(
|
|
learning_rate=0.01,
|
|
parameters=None,
|
|
weight_decay=0.1,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
def test_parameters_dtype4():
|
|
linear = paddle.nn.Linear(13, 5)
|
|
adam = paddle.optimizer.AdamW(
|
|
learning_rate=0.01,
|
|
parameters={'params': set(linear.parameters())},
|
|
weight_decay=0.1,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
def test_learning_rate_dtype():
|
|
linear = paddle.nn.Linear(13, 5)
|
|
adam = paddle.optimizer.AdamW(
|
|
learning_rate=1,
|
|
parameters=linear.parameters(),
|
|
weight_decay=0.1,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
def test_grad_clip_dtype():
|
|
linear = paddle.nn.Linear(13, 5)
|
|
adam = paddle.optimizer.AdamW(
|
|
learning_rate=0.01,
|
|
parameters=linear.parameters(),
|
|
weight_decay=0.1,
|
|
grad_clip=0.1,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
self.assertRaises(TypeError, test_parameters_dtype1)
|
|
self.assertRaises(TypeError, test_parameters_dtype2)
|
|
self.assertRaises(AttributeError, test_parameters_dtype3)
|
|
self.assertRaises(TypeError, test_parameters_dtype4)
|
|
self.assertRaises(TypeError, test_learning_rate_dtype)
|
|
self.assertRaises(TypeError, test_grad_clip_dtype)
|
|
|
|
|
|
class TestAdamWOpErrorAMSGrad(TestAdamWOpError):
|
|
def setUp(self):
|
|
# xpu not support `amsgrad`
|
|
if core.is_compiled_with_xpu():
|
|
self.amsgrad = False
|
|
self.no_check_set = ['Moment2MaxOut']
|
|
else:
|
|
self.amsgrad = True
|
|
self.no_check_set = None
|
|
|
|
|
|
class TestAdamWOpGroupWithLR(TestAdamWOp):
|
|
def test_adamw_op_dygraph(self):
|
|
paddle.disable_static()
|
|
value = np.arange(26).reshape(2, 13).astype("float32")
|
|
a = paddle.to_tensor(value)
|
|
linear_1 = paddle.nn.Linear(13, 5)
|
|
linear_2 = paddle.nn.Linear(5, 3)
|
|
adam = paddle.optimizer.AdamW(
|
|
learning_rate=paddle.optimizer.lr.PiecewiseDecay(
|
|
boundaries=[3, 6], values=[0.1, 0.2, 0.3]
|
|
),
|
|
parameters=[
|
|
{
|
|
'params': linear_1.parameters(),
|
|
'learning_rate': 0.1,
|
|
},
|
|
{
|
|
'params': linear_2.parameters(),
|
|
'weight_decay': 0.001,
|
|
},
|
|
],
|
|
apply_decay_param_fun=lambda name: True,
|
|
weight_decay=0.01,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
for _ in range(2):
|
|
out = linear_1(a)
|
|
out = linear_2(out)
|
|
out.backward()
|
|
adam.step()
|
|
adam.clear_gradients()
|
|
|
|
|
|
class TestAdamWOpGroupWithLRAMSGrad(TestAdamWOpGroupWithLR):
|
|
def setUp(self):
|
|
# xpu not support `amsgrad`
|
|
if core.is_compiled_with_xpu():
|
|
self.amsgrad = False
|
|
self.no_check_set = ['Moment2MaxOut']
|
|
else:
|
|
self.amsgrad = True
|
|
self.no_check_set = None
|
|
|
|
|
|
def simple_lr_setting(param, decay_rate, n_layers):
|
|
if "fc_0" in param.name or "linear_1" in param.name:
|
|
depth = int(param.name.split("_")[2]) + 1
|
|
elif "fc_1" in param.name or "linear_2" in param.name:
|
|
depth = int(param.name.split("_")[2]) + 2
|
|
else:
|
|
depth = 0
|
|
|
|
return decay_rate ** (n_layers + 2 - depth)
|
|
|
|
|
|
@unittest.skipIf(
|
|
not (
|
|
(core.is_compiled_with_cuda() or is_custom_device())
|
|
or core.is_compiled_with_xpu()
|
|
),
|
|
"core is not compiled with CUDA nor XPU",
|
|
)
|
|
class TestAdamWOpLayerwiseLR(TestAdamWOp):
|
|
def setUp(self):
|
|
random.seed(2022)
|
|
np.random.seed(2022)
|
|
paddle.seed(2022)
|
|
self.amsgrad = False
|
|
|
|
def test_adamw_op_dygraph(self):
|
|
paddle.disable_static()
|
|
linear1 = paddle.nn.Linear(
|
|
13, 8, bias_attr=paddle.nn.initializer.Constant(value=1.0)
|
|
)
|
|
linear2 = paddle.nn.Linear(
|
|
8, 5, bias_attr=paddle.nn.initializer.Constant(value=1.0)
|
|
)
|
|
|
|
# fix the linear name, simple_lr_setting function will use the name
|
|
linear1.weight.name = "linear_1.w_0"
|
|
linear1.bias.name = "linear_1.b_0"
|
|
linear2.weight.name = "linear_2.w_0"
|
|
linear2.bias.name = "linear_2.b_0"
|
|
|
|
fc1_w = np.array(linear1.weight)
|
|
fc1_w_mon1 = np.zeros_like(fc1_w)
|
|
fc1_w_mon2 = np.zeros_like(fc1_w)
|
|
fc1_w_mon2_max = np.zeros_like(fc1_w)
|
|
fc1_b = np.array(linear1.bias)
|
|
fc1_b_mon1 = np.zeros_like(fc1_b)
|
|
fc1_b_mon2 = np.zeros_like(fc1_b)
|
|
fc1_b_mon2_max = np.zeros_like(fc1_b)
|
|
|
|
fc2_w = np.array(linear2.weight)
|
|
fc2_w_mon1 = np.zeros_like(fc2_w)
|
|
fc2_w_mon2 = np.zeros_like(fc2_w)
|
|
fc2_w_mon2_max = np.zeros_like(fc2_w)
|
|
fc2_b = np.array(linear2.bias)
|
|
fc2_b_mon1 = np.zeros_like(fc2_b)
|
|
fc2_b_mon2 = np.zeros_like(fc2_b)
|
|
fc2_b_mon2_max = np.zeros_like(fc2_b)
|
|
|
|
simple_lr_fun = partial(simple_lr_setting, decay_rate=0.8, n_layers=2)
|
|
learning_rate = 0.001
|
|
weight_decay = 0.01
|
|
beta1 = 0.9
|
|
beta2 = 0.999
|
|
|
|
opt = paddle.optimizer.AdamW(
|
|
learning_rate=learning_rate,
|
|
parameters=[
|
|
{'params': linear1.parameters()},
|
|
{
|
|
'params': linear2.parameters(),
|
|
},
|
|
],
|
|
apply_decay_param_fun=lambda name: True,
|
|
weight_decay=weight_decay,
|
|
lr_ratio=simple_lr_fun,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
def get_numpy_output(
|
|
param, grad, moment1, moment2, moment2_max, lr_ratio, t
|
|
):
|
|
np_inputs = {
|
|
'Param': param,
|
|
'Grad': grad,
|
|
'Moment1': moment1,
|
|
'Moment2': moment2,
|
|
'Moment2Max': moment2_max,
|
|
'LearningRate': np.array([learning_rate]).astype("float64"),
|
|
'Beta1Pow': np.array([beta1**t]).astype("float32"),
|
|
'Beta2Pow': np.array([beta2**t]).astype("float32"),
|
|
}
|
|
|
|
np_attrs = {
|
|
'epsilon': 1e-8,
|
|
'beta1': beta1,
|
|
'beta2': beta2,
|
|
"lr_ratio": lr_ratio,
|
|
"coeff": weight_decay,
|
|
"with_decay": True,
|
|
"amsgrad": self.amsgrad,
|
|
}
|
|
param_out, moment1_out, moment2_out, moment2_max_out = adamw_step(
|
|
np_inputs, np_attrs
|
|
)
|
|
return param_out, moment1_out, moment2_out, moment2_max_out
|
|
|
|
for i in range(5):
|
|
a = paddle.to_tensor(
|
|
np.random.uniform(-1, 1, (2, 13)).astype("float32")
|
|
)
|
|
a1 = linear1(a)
|
|
out = linear2(a1)
|
|
out = paddle.mean(out)
|
|
out.backward()
|
|
|
|
fc1_w, fc1_w_mon1, fc1_w_mon2, fc1_w_mon2_max = get_numpy_output(
|
|
fc1_w,
|
|
np.array(linear1.weight.grad),
|
|
fc1_w_mon1,
|
|
fc1_w_mon2,
|
|
fc1_w_mon2_max,
|
|
simple_lr_fun(linear1.weight),
|
|
i + 1,
|
|
)
|
|
fc1_b, fc1_b_mon1, fc1_b_mon2, fc1_b_mon2_max = get_numpy_output(
|
|
fc1_b,
|
|
np.array(linear1.bias.grad),
|
|
fc1_b_mon1,
|
|
fc1_b_mon2,
|
|
fc1_b_mon2_max,
|
|
simple_lr_fun(linear1.bias),
|
|
i + 1,
|
|
)
|
|
fc2_w, fc2_w_mon1, fc2_w_mon2, fc2_w_mon2_max = get_numpy_output(
|
|
fc2_w,
|
|
np.array(linear2.weight.grad),
|
|
fc2_w_mon1,
|
|
fc2_w_mon2,
|
|
fc2_w_mon2_max,
|
|
simple_lr_fun(linear2.weight),
|
|
i + 1,
|
|
)
|
|
fc2_b, fc2_b_mon1, fc2_b_mon2, fc2_b_mon2_max = get_numpy_output(
|
|
fc2_b,
|
|
np.array(linear2.bias.grad),
|
|
fc2_b_mon1,
|
|
fc2_b_mon2,
|
|
fc2_b_mon2_max,
|
|
simple_lr_fun(linear2.bias),
|
|
i + 1,
|
|
)
|
|
|
|
opt.step()
|
|
opt.clear_gradients()
|
|
|
|
np.testing.assert_allclose(
|
|
linear1.weight.numpy(),
|
|
fc1_w,
|
|
atol=1e-8 if core.is_compiled_with_xpu() else 0,
|
|
rtol=1e-5,
|
|
)
|
|
np.testing.assert_allclose(linear1.bias.numpy(), fc1_b, rtol=1e-6)
|
|
np.testing.assert_allclose(linear2.weight.numpy(), fc2_w, rtol=1e-6)
|
|
np.testing.assert_allclose(linear2.bias.numpy(), fc2_b, rtol=1e-6)
|
|
|
|
def test_adamw_op(self):
|
|
with paddle.pir_utils.OldIrGuard():
|
|
paddle.enable_static()
|
|
place = (
|
|
get_device_place()
|
|
if not core.is_compiled_with_xpu()
|
|
else base.XPUPlace(0)
|
|
)
|
|
|
|
learning_rate = 0.0001
|
|
beta1 = 0.85
|
|
beta2 = 0.95
|
|
weight_decay = 0.01
|
|
epsilon = 1e-8
|
|
|
|
train_prog = paddle.static.Program()
|
|
startup = paddle.static.Program()
|
|
with (
|
|
paddle.static.program_guard(train_prog, startup),
|
|
base.unique_name.guard(),
|
|
):
|
|
x = paddle.static.data(
|
|
name='x', shape=[None, 10], dtype='float32'
|
|
)
|
|
y = paddle.static.data(
|
|
name='y', shape=[None, 1], dtype='float32'
|
|
)
|
|
|
|
weight_attr1 = paddle.framework.ParamAttr(name="linear_0.w_0")
|
|
bias_attr1 = paddle.framework.ParamAttr(
|
|
name="linear_0.b_0",
|
|
initializer=paddle.nn.initializer.Constant(value=1.0),
|
|
)
|
|
weight_attr2 = paddle.framework.ParamAttr(name="linear_1.w_0")
|
|
bias_attr2 = paddle.framework.ParamAttr(
|
|
name="linear_1.b_0",
|
|
initializer=paddle.nn.initializer.Constant(value=1.0),
|
|
)
|
|
linear1 = paddle.nn.Linear(
|
|
10, 32, weight_attr=weight_attr1, bias_attr=bias_attr1
|
|
)
|
|
linear2 = paddle.nn.Linear(
|
|
32, 1, weight_attr=weight_attr2, bias_attr=bias_attr2
|
|
)
|
|
|
|
out = linear1(x)
|
|
out = linear2(out)
|
|
|
|
fc1_w_mon1 = np.zeros(linear1.weight.shape).astype("float32")
|
|
fc1_w_mon2 = np.zeros(linear1.weight.shape).astype("float32")
|
|
fc1_w_mon2_max = np.zeros(linear1.weight.shape).astype(
|
|
"float32"
|
|
)
|
|
fc1_b_mon1 = np.zeros(linear1.bias.shape).astype("float32")
|
|
fc1_b_mon2 = np.zeros(linear1.bias.shape).astype("float32")
|
|
fc1_b_mon2_max = np.zeros(linear1.bias.shape).astype("float32")
|
|
fc2_w_mon1 = np.zeros(linear2.weight.shape).astype("float32")
|
|
fc2_w_mon2 = np.zeros(linear2.weight.shape).astype("float32")
|
|
fc2_w_mon2_max = np.zeros(linear2.weight.shape).astype(
|
|
"float32"
|
|
)
|
|
fc2_b_mon1 = np.zeros(linear2.bias.shape).astype("float32")
|
|
fc2_b_mon2 = np.zeros(linear2.bias.shape).astype("float32")
|
|
fc2_b_mon2_max = np.zeros(linear2.bias.shape).astype("float32")
|
|
|
|
cost = paddle.nn.functional.square_error_cost(
|
|
input=out, label=y
|
|
)
|
|
avg_cost = paddle.mean(cost)
|
|
|
|
simple_lr_fun = partial(
|
|
simple_lr_setting, decay_rate=0.8, n_layers=2
|
|
)
|
|
|
|
opt = paddle.optimizer.AdamW(
|
|
learning_rate=learning_rate,
|
|
beta1=beta1,
|
|
beta2=beta2,
|
|
weight_decay=weight_decay,
|
|
epsilon=epsilon,
|
|
lr_ratio=simple_lr_fun,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
opt.minimize(avg_cost)
|
|
|
|
def get_numpy_output(
|
|
param, grad, moment1, moment2, moment2_max, lr_ratio, t
|
|
):
|
|
np_inputs = {
|
|
'Param': param,
|
|
'Grad': grad,
|
|
'Moment1': moment1,
|
|
'Moment2': moment2,
|
|
'Moment2Max': moment2_max,
|
|
'LearningRate': np.array([learning_rate]).astype("float64"),
|
|
'Beta1Pow': np.array([beta1**t]).astype("float32"),
|
|
'Beta2Pow': np.array([beta2**t]).astype("float32"),
|
|
}
|
|
|
|
np_attrs = {
|
|
'epsilon': epsilon,
|
|
'beta1': beta1,
|
|
'beta2': beta2,
|
|
"lr_ratio": lr_ratio,
|
|
"coeff": weight_decay,
|
|
"with_decay": True,
|
|
"amsgrad": self.amsgrad,
|
|
}
|
|
param_out, moment1_out, moment2_out, moment2_max_out = (
|
|
adamw_step(np_inputs, np_attrs)
|
|
)
|
|
return param_out, moment1_out, moment2_out, moment2_max_out
|
|
|
|
fetch_list1 = [
|
|
"linear_0.w_0",
|
|
"linear_0.b_0",
|
|
"linear_1.w_0",
|
|
"linear_1.b_0",
|
|
]
|
|
fetch_list2 = [
|
|
"linear_0.w_0",
|
|
"linear_0.w_0@GRAD",
|
|
"linear_0.b_0",
|
|
"linear_0.b_0@GRAD",
|
|
"linear_1.w_0",
|
|
"linear_1.w_0@GRAD",
|
|
"linear_1.b_0",
|
|
"linear_1.b_0@GRAD",
|
|
]
|
|
|
|
exe = base.Executor(place)
|
|
exe.run(startup)
|
|
test_prog = train_prog.clone(for_test=True)
|
|
|
|
for i in range(5):
|
|
inputs = np.random.random(size=[8, 10]).astype('float32')
|
|
outputs = np.random.random(size=[8, 1]).astype('float32')
|
|
|
|
param = exe.run(
|
|
test_prog,
|
|
feed={"x": inputs, "y": outputs},
|
|
fetch_list=fetch_list1,
|
|
)
|
|
params_and_gras = exe.run(
|
|
train_prog,
|
|
feed={"x": inputs, "y": outputs},
|
|
fetch_list=fetch_list2,
|
|
)
|
|
|
|
fc1_w = param[0]
|
|
fc1_w_grad = params_and_gras[1]
|
|
fc1_b = param[1]
|
|
fc1_b_grad = params_and_gras[3]
|
|
fc2_w = param[2]
|
|
fc2_w_grad = params_and_gras[5]
|
|
fc2_b = param[3]
|
|
fc2_b_grad = params_and_gras[7]
|
|
|
|
fc1_w, fc1_w_mon1, fc1_w_mon2, fc1_w_mon2_max = (
|
|
get_numpy_output(
|
|
fc1_w,
|
|
fc1_w_grad,
|
|
fc1_w_mon1,
|
|
fc1_w_mon2,
|
|
fc1_w_mon2_max,
|
|
simple_lr_fun(linear1.weight),
|
|
i + 1,
|
|
)
|
|
)
|
|
fc1_b, fc1_b_mon1, fc1_b_mon2, fc1_b_mon2_max = (
|
|
get_numpy_output(
|
|
fc1_b,
|
|
fc1_b_grad,
|
|
fc1_b_mon1,
|
|
fc1_b_mon2,
|
|
fc1_b_mon2_max,
|
|
simple_lr_fun(linear1.bias),
|
|
i + 1,
|
|
)
|
|
)
|
|
fc2_w, fc2_w_mon1, fc2_w_mon2, fc2_w_mon2_max = (
|
|
get_numpy_output(
|
|
fc2_w,
|
|
fc2_w_grad,
|
|
fc2_w_mon1,
|
|
fc2_w_mon2,
|
|
fc2_w_mon2_max,
|
|
simple_lr_fun(linear2.weight),
|
|
i + 1,
|
|
)
|
|
)
|
|
fc2_b, fc2_b_mon1, fc2_b_mon2, fc2_b_mon2_max = (
|
|
get_numpy_output(
|
|
fc2_b,
|
|
fc2_b_grad,
|
|
fc2_b_mon1,
|
|
fc2_b_mon2,
|
|
fc2_b_mon2_max,
|
|
simple_lr_fun(linear2.bias),
|
|
i + 1,
|
|
)
|
|
)
|
|
|
|
np.testing.assert_allclose(params_and_gras[0], fc1_w, rtol=1e-6)
|
|
np.testing.assert_allclose(params_and_gras[2], fc1_b, rtol=1e-6)
|
|
np.testing.assert_allclose(
|
|
params_and_gras[4],
|
|
fc2_w,
|
|
rtol=1e-6 if not core.is_compiled_with_xpu() else 1e-5,
|
|
)
|
|
np.testing.assert_allclose(params_and_gras[6], fc2_b, rtol=1e-6)
|
|
|
|
paddle.disable_static()
|
|
|
|
def test_adamw_op_with_pir(self):
|
|
with paddle.pir_utils.IrGuard():
|
|
paddle.enable_static()
|
|
place = (
|
|
get_device_place()
|
|
if not core.is_compiled_with_xpu()
|
|
else base.XPUPlace(0)
|
|
)
|
|
|
|
learning_rate = 0.0001
|
|
beta1 = 0.85
|
|
beta2 = 0.95
|
|
weight_decay = 0.01
|
|
epsilon = 1e-8
|
|
|
|
train_prog = paddle.static.Program()
|
|
train_startup = paddle.static.Program()
|
|
with (
|
|
paddle.static.program_guard(train_prog, train_startup),
|
|
base.unique_name.guard(),
|
|
):
|
|
x = paddle.static.data(
|
|
name='x', shape=[None, 10], dtype='float32'
|
|
)
|
|
y = paddle.static.data(
|
|
name='y', shape=[None, 1], dtype='float32'
|
|
)
|
|
|
|
weight_attr1 = paddle.framework.ParamAttr(name="linear_0.w_0")
|
|
bias_attr1 = paddle.framework.ParamAttr(
|
|
name="linear_0.b_0",
|
|
initializer=paddle.nn.initializer.Constant(value=1.0),
|
|
)
|
|
weight_attr2 = paddle.framework.ParamAttr(name="linear_1.w_0")
|
|
bias_attr2 = paddle.framework.ParamAttr(
|
|
name="linear_1.b_0",
|
|
initializer=paddle.nn.initializer.Constant(value=1.0),
|
|
)
|
|
linear1 = paddle.nn.Linear(
|
|
10, 32, weight_attr=weight_attr1, bias_attr=bias_attr1
|
|
)
|
|
linear2 = paddle.nn.Linear(
|
|
32, 1, weight_attr=weight_attr2, bias_attr=bias_attr2
|
|
)
|
|
|
|
out = linear1(x)
|
|
out = linear2(out)
|
|
|
|
fc1_w_mon1 = np.zeros(linear1.weight.shape).astype("float32")
|
|
fc1_w_mon2 = np.zeros(linear1.weight.shape).astype("float32")
|
|
fc1_w_mon2_max = np.zeros(linear1.weight.shape).astype(
|
|
"float32"
|
|
)
|
|
fc1_b_mon1 = np.zeros(linear1.bias.shape).astype("float32")
|
|
fc1_b_mon2 = np.zeros(linear1.bias.shape).astype("float32")
|
|
fc1_b_mon2_max = np.zeros(linear1.bias.shape).astype("float32")
|
|
fc2_w_mon1 = np.zeros(linear2.weight.shape).astype("float32")
|
|
fc2_w_mon2 = np.zeros(linear2.weight.shape).astype("float32")
|
|
fc2_w_mon2_max = np.zeros(linear2.weight.shape).astype(
|
|
"float32"
|
|
)
|
|
fc2_b_mon1 = np.zeros(linear2.bias.shape).astype("float32")
|
|
fc2_b_mon2 = np.zeros(linear2.bias.shape).astype("float32")
|
|
fc2_b_mon2_max = np.zeros(linear2.bias.shape).astype("float32")
|
|
|
|
cost = paddle.nn.functional.square_error_cost(
|
|
input=out, label=y
|
|
)
|
|
avg_cost = paddle.mean(cost)
|
|
|
|
simple_lr_fun = partial(
|
|
simple_lr_setting, decay_rate=0.8, n_layers=2
|
|
)
|
|
|
|
opt = paddle.optimizer.AdamW(
|
|
learning_rate=learning_rate,
|
|
beta1=beta1,
|
|
beta2=beta2,
|
|
weight_decay=weight_decay,
|
|
epsilon=epsilon,
|
|
lr_ratio=simple_lr_fun,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
_, params_grads = opt.minimize(avg_cost)
|
|
|
|
def get_numpy_output(
|
|
param, grad, moment1, moment2, moment2_max, lr_ratio, t
|
|
):
|
|
np_inputs = {
|
|
'Param': param,
|
|
'Grad': grad,
|
|
'Moment1': moment1,
|
|
'Moment2': moment2,
|
|
'Moment2Max': moment2_max,
|
|
'LearningRate': np.array([learning_rate]).astype("float64"),
|
|
'Beta1Pow': np.array([beta1**t]).astype("float32"),
|
|
'Beta2Pow': np.array([beta2**t]).astype("float32"),
|
|
}
|
|
|
|
np_attrs = {
|
|
'epsilon': epsilon,
|
|
'beta1': beta1,
|
|
'beta2': beta2,
|
|
"lr_ratio": lr_ratio,
|
|
"coeff": weight_decay,
|
|
"with_decay": True,
|
|
"amsgrad": self.amsgrad,
|
|
}
|
|
param_out, moment1_out, moment2_out, moment2_out_max = (
|
|
adamw_step(np_inputs, np_attrs)
|
|
)
|
|
return param_out, moment1_out, moment2_out, moment2_out_max
|
|
|
|
exe = base.Executor(place)
|
|
exe.run(train_startup)
|
|
|
|
test_prog = paddle.static.Program()
|
|
test_startup = paddle.static.Program()
|
|
with (
|
|
paddle.static.program_guard(test_prog, test_startup),
|
|
base.unique_name.guard(),
|
|
):
|
|
x = paddle.static.data(
|
|
name='x', shape=[None, 10], dtype='float32'
|
|
)
|
|
y = paddle.static.data(
|
|
name='y', shape=[None, 1], dtype='float32'
|
|
)
|
|
|
|
weight_attr1 = paddle.framework.ParamAttr(name="linear_0.w_0")
|
|
bias_attr1 = paddle.framework.ParamAttr(
|
|
name="linear_0.b_0",
|
|
initializer=paddle.nn.initializer.Constant(value=1.0),
|
|
)
|
|
weight_attr2 = paddle.framework.ParamAttr(name="linear_1.w_0")
|
|
bias_attr2 = paddle.framework.ParamAttr(
|
|
name="linear_1.b_0",
|
|
initializer=paddle.nn.initializer.Constant(value=1.0),
|
|
)
|
|
linear1_2 = paddle.nn.Linear(
|
|
10, 32, weight_attr=weight_attr1, bias_attr=bias_attr1
|
|
)
|
|
linear2_2 = paddle.nn.Linear(
|
|
32, 1, weight_attr=weight_attr2, bias_attr=bias_attr2
|
|
)
|
|
|
|
out = linear1_2(x)
|
|
out = linear2_2(out)
|
|
|
|
cost = paddle.nn.functional.square_error_cost(
|
|
input=out, label=y
|
|
)
|
|
avg_cost = paddle.mean(cost)
|
|
|
|
simple_lr_fun = partial(
|
|
simple_lr_setting, decay_rate=0.8, n_layers=2
|
|
)
|
|
|
|
random.seed(2022)
|
|
np.random.seed(2022)
|
|
paddle.seed(2022)
|
|
exe.run(test_startup)
|
|
|
|
test_fetch_list = [
|
|
linear1_2.weight,
|
|
linear1_2.bias,
|
|
linear2_2.weight,
|
|
linear2_2.bias,
|
|
]
|
|
|
|
train_fetch_list = [
|
|
item for sublist in params_grads for item in sublist
|
|
]
|
|
|
|
for i in range(5):
|
|
inputs = np.random.random(size=[8, 10]).astype('float32')
|
|
outputs = np.random.random(size=[8, 1]).astype('float32')
|
|
|
|
param = exe.run(
|
|
test_prog,
|
|
feed={"x": inputs, "y": outputs},
|
|
fetch_list=test_fetch_list,
|
|
)
|
|
params_and_gras = exe.run(
|
|
train_prog,
|
|
feed={"x": inputs, "y": outputs},
|
|
fetch_list=train_fetch_list,
|
|
)
|
|
|
|
fc1_w = param[0]
|
|
fc1_w_grad = params_and_gras[7]
|
|
fc1_b = param[1]
|
|
fc1_b_grad = params_and_gras[5]
|
|
fc2_w = param[2]
|
|
fc2_w_grad = params_and_gras[3]
|
|
fc2_b = param[3]
|
|
fc2_b_grad = params_and_gras[1]
|
|
|
|
fc1_w, fc1_w_mon1, fc1_w_mon2, fc1_w_mon2_max = (
|
|
get_numpy_output(
|
|
fc1_w,
|
|
fc1_w_grad,
|
|
fc1_w_mon1,
|
|
fc1_w_mon2,
|
|
fc1_w_mon2_max,
|
|
simple_lr_fun(linear1.weight),
|
|
i + 1,
|
|
)
|
|
)
|
|
fc1_b, fc1_b_mon1, fc1_b_mon2, fc1_b_mon2_max = (
|
|
get_numpy_output(
|
|
fc1_b,
|
|
fc1_b_grad,
|
|
fc1_b_mon1,
|
|
fc1_b_mon2,
|
|
fc1_b_mon2_max,
|
|
simple_lr_fun(linear1.bias),
|
|
i + 1,
|
|
)
|
|
)
|
|
fc2_w, fc2_w_mon1, fc2_w_mon2, fc2_w_mon2_max = (
|
|
get_numpy_output(
|
|
fc2_w,
|
|
fc2_w_grad,
|
|
fc2_w_mon1,
|
|
fc2_w_mon2,
|
|
fc2_w_mon2_max,
|
|
simple_lr_fun(linear2.weight),
|
|
i + 1,
|
|
)
|
|
)
|
|
fc2_b, fc2_b_mon1, fc2_b_mon2, fc2_b_mon2_max = (
|
|
get_numpy_output(
|
|
fc2_b,
|
|
fc2_b_grad,
|
|
fc2_b_mon1,
|
|
fc2_b_mon2,
|
|
fc2_b_mon2_max,
|
|
simple_lr_fun(linear2.bias),
|
|
i + 1,
|
|
)
|
|
)
|
|
|
|
np.testing.assert_allclose(params_and_gras[6], fc1_w, rtol=1e-6)
|
|
np.testing.assert_allclose(params_and_gras[4], fc1_b, rtol=1e-6)
|
|
np.testing.assert_allclose(
|
|
params_and_gras[2],
|
|
fc2_w,
|
|
rtol=1e-6 if not core.is_compiled_with_xpu() else 1e-5,
|
|
)
|
|
np.testing.assert_allclose(params_and_gras[0], fc2_b, rtol=1e-6)
|
|
|
|
paddle.disable_static()
|
|
|
|
def test_weight_decay_int(self):
|
|
paddle.disable_static()
|
|
linear1 = paddle.nn.Linear(
|
|
13, 8, bias_attr=paddle.nn.initializer.Constant(value=1.0)
|
|
)
|
|
linear2 = paddle.nn.Linear(
|
|
8, 5, bias_attr=paddle.nn.initializer.Constant(value=1.0)
|
|
)
|
|
|
|
# fix the linear name, simple_lr_setting function will use the name
|
|
linear1.weight.name = "linear_1.w_0"
|
|
linear1.bias.name = "linear_1.b_0"
|
|
linear2.weight.name = "linear_2.w_0"
|
|
linear2.bias.name = "linear_2.b_0"
|
|
|
|
fc1_w = np.array(linear1.weight)
|
|
fc1_w_mon1 = np.zeros_like(fc1_w)
|
|
fc1_w_mon2 = np.zeros_like(fc1_w)
|
|
fc1_w_mon2_max = np.zeros_like(fc1_w)
|
|
fc1_b = np.array(linear1.bias)
|
|
fc1_b_mon1 = np.zeros_like(fc1_b)
|
|
fc1_b_mon2 = np.zeros_like(fc1_b)
|
|
fc1_b_mon2_max = np.zeros_like(fc1_b)
|
|
|
|
fc2_w = np.array(linear2.weight)
|
|
fc2_w_mon1 = np.zeros_like(fc2_w)
|
|
fc2_w_mon2 = np.zeros_like(fc2_w)
|
|
fc2_w_mon2_max = np.zeros_like(fc2_w)
|
|
fc2_b = np.array(linear2.bias)
|
|
fc2_b_mon1 = np.zeros_like(fc2_b)
|
|
fc2_b_mon2 = np.zeros_like(fc2_b)
|
|
fc2_b_mon2_max = np.zeros_like(fc2_b)
|
|
|
|
simple_lr_fun = partial(simple_lr_setting, decay_rate=0.8, n_layers=2)
|
|
learning_rate = 0.001
|
|
weight_decay = 0
|
|
beta1 = 0.9
|
|
beta2 = 0.999
|
|
|
|
opt = paddle.optimizer.AdamW(
|
|
learning_rate=learning_rate,
|
|
parameters=[
|
|
{'params': linear1.parameters()},
|
|
{
|
|
'params': linear2.parameters(),
|
|
},
|
|
],
|
|
apply_decay_param_fun=lambda name: True,
|
|
weight_decay=weight_decay,
|
|
lr_ratio=simple_lr_fun,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
def get_numpy_output(
|
|
param, grad, moment1, moment2, moment2_max, lr_ratio, t
|
|
):
|
|
np_inputs = {
|
|
'Param': param,
|
|
'Grad': grad,
|
|
'Moment1': moment1,
|
|
'Moment2': moment2,
|
|
'Moment2Max': moment2_max,
|
|
'LearningRate': np.array([learning_rate]).astype("float64"),
|
|
'Beta1Pow': np.array([beta1**t]).astype("float32"),
|
|
'Beta2Pow': np.array([beta2**t]).astype("float32"),
|
|
}
|
|
|
|
np_attrs = {
|
|
'epsilon': 1e-8,
|
|
'beta1': beta1,
|
|
'beta2': beta2,
|
|
"lr_ratio": lr_ratio,
|
|
"coeff": float(weight_decay),
|
|
"with_decay": True,
|
|
"amsgrad": self.amsgrad,
|
|
}
|
|
param_out, moment1_out, moment2_out, moment2_out_max = adamw_step(
|
|
np_inputs, np_attrs
|
|
)
|
|
return param_out, moment1_out, moment2_out, moment2_out_max
|
|
|
|
for i in range(5):
|
|
a = paddle.to_tensor(
|
|
np.random.uniform(-1, 1, (2, 13)).astype("float32")
|
|
)
|
|
a1 = linear1(a)
|
|
out = linear2(a1)
|
|
out = paddle.mean(out)
|
|
out.backward()
|
|
|
|
fc1_w, fc1_w_mon1, fc1_w_mon2, fc1_w_mon2_max = get_numpy_output(
|
|
fc1_w,
|
|
np.array(linear1.weight.grad),
|
|
fc1_w_mon1,
|
|
fc1_w_mon2,
|
|
fc1_w_mon2_max,
|
|
simple_lr_fun(linear1.weight),
|
|
i + 1,
|
|
)
|
|
fc1_b, fc1_b_mon1, fc1_b_mon2, fc1_b_mon2_max = get_numpy_output(
|
|
fc1_b,
|
|
np.array(linear1.bias.grad),
|
|
fc1_b_mon1,
|
|
fc1_b_mon2,
|
|
fc1_b_mon2_max,
|
|
simple_lr_fun(linear1.bias),
|
|
i + 1,
|
|
)
|
|
fc2_w, fc2_w_mon1, fc2_w_mon2, fc2_w_mon2_max = get_numpy_output(
|
|
fc2_w,
|
|
np.array(linear2.weight.grad),
|
|
fc2_w_mon1,
|
|
fc2_w_mon2,
|
|
fc2_w_mon2_max,
|
|
simple_lr_fun(linear2.weight),
|
|
i + 1,
|
|
)
|
|
fc2_b, fc2_b_mon1, fc2_b_mon2, fc2_b_mon2_max = get_numpy_output(
|
|
fc2_b,
|
|
np.array(linear2.bias.grad),
|
|
fc2_b_mon1,
|
|
fc2_b_mon2,
|
|
fc2_b_mon2_max,
|
|
simple_lr_fun(linear2.bias),
|
|
i + 1,
|
|
)
|
|
|
|
opt.step()
|
|
opt.clear_gradients()
|
|
|
|
np.testing.assert_allclose(
|
|
linear1.weight.numpy(),
|
|
fc1_w,
|
|
atol=1e-8 if core.is_compiled_with_xpu() else 0,
|
|
rtol=1e-5,
|
|
)
|
|
np.testing.assert_allclose(linear1.bias.numpy(), fc1_b, rtol=1e-6)
|
|
np.testing.assert_allclose(linear2.weight.numpy(), fc2_w, rtol=1e-6)
|
|
np.testing.assert_allclose(linear2.bias.numpy(), fc2_b, rtol=1e-6)
|
|
|
|
|
|
class TestAdamWOpLayerwiseLRAMSGrad(TestAdamWOpLayerwiseLR):
|
|
def setUp(self):
|
|
random.seed(2022)
|
|
np.random.seed(2022)
|
|
paddle.seed(2022)
|
|
|
|
# xpu not support `amsgrad`
|
|
if core.is_compiled_with_xpu():
|
|
self.amsgrad = False
|
|
self.no_check_set = ['Moment2MaxOut']
|
|
else:
|
|
self.amsgrad = True
|
|
self.no_check_set = None
|
|
|
|
|
|
@unittest.skipIf(
|
|
core.is_compiled_with_xpu(),
|
|
"core is not compiled with XPU",
|
|
)
|
|
class TestAdamwMomentBfloat16Amp(unittest.TestCase):
|
|
def setUp(self):
|
|
self.amsgrad = False
|
|
self.num_steps = 10 # 增加训练的 step 数
|
|
|
|
def test_adamw_moment_bfloat16_amp(self):
|
|
paddle.disable_static()
|
|
value = np.arange(26).reshape(2, 13).astype("float32")
|
|
a = paddle.to_tensor(value)
|
|
a.stop_gradient = True
|
|
|
|
paddle.seed(111)
|
|
linear1 = nn.Linear(13, 5)
|
|
paddle.seed(111)
|
|
linear2 = nn.Linear(13, 5)
|
|
|
|
adam1 = paddle.optimizer.AdamW(
|
|
learning_rate=0.01,
|
|
parameters=linear1.parameters(),
|
|
apply_decay_param_fun=lambda name: True,
|
|
weight_decay=0.01,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
adam2 = paddle.optimizer.AdamW(
|
|
learning_rate=0.01,
|
|
parameters=linear2.parameters(),
|
|
apply_decay_param_fun=lambda name: True,
|
|
weight_decay=0.01,
|
|
amsgrad=self.amsgrad,
|
|
)
|
|
|
|
for _ in range(self.num_steps):
|
|
output1 = linear1(a)
|
|
loss1 = paddle.mean(output1)
|
|
loss1.backward()
|
|
adam1.step()
|
|
adam1.clear_grad()
|
|
|
|
model, optimizer = paddle.amp.decorate(
|
|
models=linear2, optimizers=adam2, level='O2', dtype="bfloat16"
|
|
)
|
|
|
|
for _ in range(self.num_steps):
|
|
with paddle.amp.auto_cast(
|
|
dtype="bfloat16",
|
|
enable=True,
|
|
custom_white_list=None,
|
|
custom_black_list=None,
|
|
level="O2",
|
|
):
|
|
output2 = model(a)
|
|
loss2 = paddle.mean(output2)
|
|
loss2.backward()
|
|
optimizer.step()
|
|
optimizer.clear_grad()
|
|
|
|
np.testing.assert_allclose(
|
|
loss1.astype(paddle.float32).numpy(),
|
|
loss2.astype(paddle.float32).numpy(),
|
|
rtol=1e-2,
|
|
atol=1e-2,
|
|
)
|
|
|
|
for param1, param2 in zip(linear1.parameters(), linear2.parameters()):
|
|
np.testing.assert_allclose(
|
|
param1.astype(paddle.float32).numpy(),
|
|
param2.astype(paddle.float32).numpy(),
|
|
rtol=1e-2,
|
|
atol=1e-2,
|
|
)
|
|
|
|
def _get_places(self):
|
|
places = []
|
|
if paddle.is_compiled_with_cuda() or is_custom_device():
|
|
places.append(get_device())
|
|
if paddle.is_compiled_with_xpu():
|
|
places.append('xpu')
|
|
return places
|
|
|
|
def _test_adamw_op_dygraph_place_amp_with_maingrad_with_moment_bf16(
|
|
self, place, shape, use_main_grad
|
|
):
|
|
paddle.disable_static()
|
|
paddle.seed(10)
|
|
paddle.set_device(place)
|
|
|
|
found_inf = None
|
|
|
|
_weight_decay = 0.1
|
|
with_decay = True
|
|
_lazy_mode = False
|
|
find_master = True
|
|
|
|
_epsilon = 1e-8
|
|
|
|
_beta1 = 0.9
|
|
_beta2 = 0.99
|
|
lr_ratio_ = 1.0
|
|
|
|
lr_rate = 1e-8
|
|
|
|
param = paddle.randn(shape).astype(paddle.bfloat16)
|
|
master_weight = param.astype(paddle.float32)
|
|
grad = paddle.randn(shape).astype(paddle.bfloat16)
|
|
main_grad = grad.astype(paddle.float32)
|
|
moment1 = paddle.randn(shape).astype(paddle.bfloat16)
|
|
moment2 = paddle.randn(shape).astype(paddle.bfloat16).abs()
|
|
moment2_max = paddle.zeros(shape).astype(paddle.bfloat16)
|
|
lr = paddle.zeros([1]).astype(paddle.float64)
|
|
lr[0] = lr_rate
|
|
beta1_pow_acc = paddle.ones([1]).astype(paddle.float32)
|
|
beta1_pow_acc[0] = _beta1**10
|
|
beta2_pow_acc = paddle.ones([1]).astype(paddle.float32)
|
|
beta2_pow_acc[0] = _beta2**10
|
|
|
|
ref_param = param.astype(paddle.float32).clone().detach()
|
|
ref_beta1_pow_acc = (
|
|
beta1_pow_acc.astype(paddle.float32).clone().detach()
|
|
)
|
|
ref_beta2_pow_acc = (
|
|
beta2_pow_acc.astype(paddle.float32).clone().detach()
|
|
)
|
|
ref_moment_1 = moment1.astype(paddle.bfloat16).clone().detach()
|
|
ref_moment_2 = moment2.astype(paddle.bfloat16).clone().detach()
|
|
ref_moment_2_max = moment2_max.astype(paddle.bfloat16).clone().detach()
|
|
|
|
# reference code
|
|
_, _, _, _, _, _, _ = paddle._C_ops.adamw_(
|
|
ref_param,
|
|
main_grad,
|
|
lr,
|
|
ref_moment_1,
|
|
ref_moment_2,
|
|
ref_moment_2_max,
|
|
ref_beta1_pow_acc,
|
|
ref_beta2_pow_acc,
|
|
master_weight,
|
|
found_inf,
|
|
_beta1,
|
|
_beta2,
|
|
_epsilon,
|
|
lr_ratio_,
|
|
_weight_decay,
|
|
with_decay,
|
|
_lazy_mode,
|
|
1000,
|
|
False,
|
|
False,
|
|
self.amsgrad,
|
|
)
|
|
|
|
if use_main_grad:
|
|
_, _, _, _, _, _, _ = paddle._C_ops.adamw_(
|
|
param,
|
|
main_grad,
|
|
lr,
|
|
moment1,
|
|
moment2,
|
|
moment2_max,
|
|
beta1_pow_acc,
|
|
beta2_pow_acc,
|
|
master_weight,
|
|
found_inf,
|
|
_beta1,
|
|
_beta2,
|
|
_epsilon,
|
|
lr_ratio_,
|
|
_weight_decay,
|
|
with_decay,
|
|
_lazy_mode,
|
|
1000,
|
|
find_master,
|
|
False,
|
|
self.amsgrad,
|
|
)
|
|
np.testing.assert_allclose(
|
|
param.astype("float32").numpy(), ref_param.numpy(), rtol=1e-2
|
|
)
|
|
|
|
if self.amsgrad:
|
|
np.testing.assert_allclose(
|
|
master_weight.numpy(), ref_param.numpy(), rtol=1e-4
|
|
)
|
|
else:
|
|
np.testing.assert_allclose(
|
|
master_weight.numpy(), ref_param.numpy(), rtol=1e-6
|
|
)
|
|
|
|
else:
|
|
_, _, _, _, _, _, _ = paddle._C_ops.adamw_(
|
|
param,
|
|
grad,
|
|
lr,
|
|
moment1,
|
|
moment2,
|
|
moment2_max,
|
|
beta1_pow_acc,
|
|
beta2_pow_acc,
|
|
master_weight,
|
|
found_inf,
|
|
_beta1,
|
|
_beta2,
|
|
_epsilon,
|
|
lr_ratio_,
|
|
_weight_decay,
|
|
with_decay,
|
|
_lazy_mode,
|
|
1000,
|
|
find_master,
|
|
False,
|
|
self.amsgrad,
|
|
)
|
|
np.testing.assert_allclose(
|
|
param.astype("float32").numpy(), ref_param.numpy(), rtol=1e-2
|
|
)
|
|
|
|
if self.amsgrad:
|
|
np.testing.assert_allclose(
|
|
master_weight.numpy(), ref_param.numpy(), rtol=1e-4
|
|
)
|
|
else:
|
|
np.testing.assert_allclose(
|
|
master_weight.numpy(), ref_param.numpy(), rtol=1e-6
|
|
)
|
|
|
|
def test_adamw_op_dygraph_place_amp_with_maingrad_with_moment_bf16(self):
|
|
for _ in range(10):
|
|
shape = paddle.randint(1, 1024, [2])
|
|
for place in self._get_places():
|
|
use_main_grad_list = [True, False]
|
|
for use_main_grad in use_main_grad_list:
|
|
self._test_adamw_op_dygraph_place_amp_with_maingrad_with_moment_bf16(
|
|
place, shape, use_main_grad
|
|
)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|