564 lines
17 KiB
Python
564 lines
17 KiB
Python
# Copyright (c) 2019 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 unittest
|
|
|
|
import numpy as np
|
|
from op import Operator
|
|
from op_test import (
|
|
OpTest,
|
|
convert_float_to_uint16,
|
|
get_device_place,
|
|
get_places,
|
|
is_custom_device,
|
|
)
|
|
|
|
import paddle
|
|
from paddle.base import core
|
|
|
|
paddle.enable_static()
|
|
|
|
|
|
def lamb_wrapper(
|
|
param,
|
|
grad,
|
|
lr,
|
|
moment1,
|
|
moment2,
|
|
beta1Pow,
|
|
beta2Pow,
|
|
master_weight=None,
|
|
found_inf=None,
|
|
epsilon=1e-8,
|
|
beta1=0.9,
|
|
beta2=0.999,
|
|
weight_decay=0.01,
|
|
always_adapt=False,
|
|
):
|
|
return paddle._C_ops.lamb_(
|
|
param,
|
|
grad,
|
|
lr,
|
|
moment1,
|
|
moment2,
|
|
beta1Pow,
|
|
beta2Pow,
|
|
master_weight,
|
|
found_inf,
|
|
weight_decay,
|
|
beta1,
|
|
beta2,
|
|
epsilon,
|
|
False,
|
|
False,
|
|
)
|
|
|
|
|
|
class TestLambOp1(OpTest):
|
|
def set_attrs(self):
|
|
self.attrs = {
|
|
'epsilon': 1e-4,
|
|
'beta1': 0.78,
|
|
'beta2': 0.836,
|
|
'weight_decay': 0.01,
|
|
'always_adapt': False,
|
|
}
|
|
|
|
def set_dtype(self):
|
|
self.dtype = np.float32
|
|
|
|
def setUp(self):
|
|
'''Test Lamb Op with supplied attributes'''
|
|
self.op_type = "lamb"
|
|
self.set_dtype()
|
|
|
|
if self.is_bfloat16_op():
|
|
param = np.random.uniform(-1, 1, (102, 105)).astype(np.float32)
|
|
grad = np.random.uniform(-1, 1, (102, 105)).astype(np.float32)
|
|
moment1 = np.random.uniform(-1, 1, (102, 105)).astype(np.float32)
|
|
moment2 = np.random.random((102, 105)).astype(np.float32)
|
|
else:
|
|
param = np.random.uniform(-1, 1, (102, 105)).astype(self.dtype)
|
|
grad = np.random.uniform(-1, 1, (102, 105)).astype(self.dtype)
|
|
moment1 = np.random.uniform(-1, 1, (102, 105)).astype(self.dtype)
|
|
moment2 = np.random.random((102, 105)).astype(self.dtype)
|
|
|
|
learning_rate = 0.001
|
|
self.set_attrs()
|
|
|
|
self.python_api = lamb_wrapper
|
|
self.python_out_sig = ['Out']
|
|
|
|
beta1_pow = self.attrs['beta1']
|
|
beta2_pow = self.attrs['beta2']
|
|
|
|
if self.is_bfloat16_op():
|
|
self.inputs = {
|
|
'Param': convert_float_to_uint16(param),
|
|
'Grad': convert_float_to_uint16(grad),
|
|
'Moment1': convert_float_to_uint16(moment1),
|
|
'Moment2': convert_float_to_uint16(moment2),
|
|
'LearningRate': convert_float_to_uint16(
|
|
np.array([learning_rate]).astype(self.dtype)
|
|
),
|
|
'Beta1Pow': convert_float_to_uint16(
|
|
np.array([beta1_pow]).astype(self.dtype)
|
|
),
|
|
'Beta2Pow': convert_float_to_uint16(
|
|
np.array([beta2_pow]).astype(self.dtype)
|
|
),
|
|
}
|
|
|
|
else:
|
|
self.inputs = {
|
|
'Param': param,
|
|
'Grad': grad,
|
|
'Moment1': moment1,
|
|
'Moment2': moment2,
|
|
'LearningRate': np.array([learning_rate]).astype(self.dtype),
|
|
'Beta1Pow': np.array([beta1_pow]).astype(self.dtype),
|
|
'Beta2Pow': np.array([beta2_pow]).astype(self.dtype),
|
|
}
|
|
|
|
(
|
|
param_out,
|
|
moment1_out,
|
|
moment2_out,
|
|
beta1_pow_out,
|
|
beta2_pow_out,
|
|
) = lamb_step(self.inputs, self.attrs)
|
|
|
|
if self.is_bfloat16_op():
|
|
self.outputs = {
|
|
'Moment1Out': convert_float_to_uint16(moment1_out),
|
|
'Moment2Out': convert_float_to_uint16(moment2_out),
|
|
'ParamOut': convert_float_to_uint16(param_out),
|
|
'Beta1PowOut': convert_float_to_uint16(beta1_pow_out),
|
|
'Beta2PowOut': convert_float_to_uint16(beta2_pow_out),
|
|
}
|
|
else:
|
|
self.outputs = {
|
|
'Moment1Out': moment1_out,
|
|
'Moment2Out': moment2_out,
|
|
'ParamOut': param_out,
|
|
'Beta1PowOut': beta1_pow_out,
|
|
'Beta2PowOut': beta2_pow_out,
|
|
}
|
|
|
|
def test_check_output(self):
|
|
self.check_output()
|
|
|
|
|
|
class TestLambOp2(TestLambOp1):
|
|
def set_attrs(self):
|
|
self.attrs = {
|
|
'epsilon': 1e-8,
|
|
'beta1': 0.9,
|
|
'beta2': 0.999,
|
|
'weight_decay': 0.01,
|
|
'always_adapt': False,
|
|
}
|
|
|
|
|
|
class TestLambOp3(TestLambOp1):
|
|
def set_attrs(self):
|
|
self.attrs = {
|
|
'epsilon': 1e-8,
|
|
'beta1': 0.9,
|
|
'beta2': 0.999,
|
|
'weight_decay': 0.0,
|
|
'always_adapt': False,
|
|
}
|
|
|
|
|
|
class TestLambOpMultipleSteps(TestLambOp1):
|
|
def set_attrs(self):
|
|
self.attrs = {
|
|
'epsilon': 1e-8,
|
|
'beta1': 0.9,
|
|
'beta2': 0.999,
|
|
'weight_decay': 0.01,
|
|
'always_adapt': False,
|
|
}
|
|
self.num_steps = 10
|
|
|
|
def test_check_output(self):
|
|
for i in range(self.num_steps):
|
|
(
|
|
param_out,
|
|
moment1_out,
|
|
moment2_out,
|
|
beta1_pow_out,
|
|
beta2_pow_out,
|
|
) = lamb_step(self.inputs, self.attrs)
|
|
|
|
self.outputs = {
|
|
'Moment1Out': moment1_out,
|
|
'Moment2Out': moment2_out,
|
|
'ParamOut': param_out,
|
|
'Beta1PowOut': beta1_pow_out,
|
|
'Beta2PowOut': beta2_pow_out,
|
|
}
|
|
|
|
# Verify output for this step
|
|
self.check_output()
|
|
|
|
# Output of this step becomes input for next step
|
|
self.inputs['Param'] = param_out
|
|
self.inputs['Moment1'] = moment1_out
|
|
self.inputs['Moment2'] = moment2_out
|
|
|
|
# Update powers of Beta1 and Beta2 for next time step
|
|
self.inputs['Beta1Pow'] = beta1_pow_out
|
|
self.inputs['Beta2Pow'] = beta2_pow_out
|
|
|
|
# Randomize gradient for next step
|
|
self.inputs['Grad'] = np.random.uniform(-1, 1, (102, 105)).astype(
|
|
self.dtype
|
|
)
|
|
|
|
|
|
class TestLambFP16Op1(TestLambOp1):
|
|
def set_dtype(self):
|
|
self.__class__.op_type = "lamb"
|
|
self.dtype = np.float16
|
|
|
|
def test_check_output(self):
|
|
if core.is_compiled_with_cuda() or is_custom_device():
|
|
place = get_device_place()
|
|
if core.is_float16_supported(place):
|
|
self.check_output_with_place(place)
|
|
|
|
|
|
class TestLambFP16Op2(TestLambOp2):
|
|
def set_dtype(self):
|
|
self.__class__.op_type = "lamb"
|
|
self.dtype = np.float16
|
|
|
|
def test_check_output(self):
|
|
if core.is_compiled_with_cuda() or is_custom_device():
|
|
place = get_device_place()
|
|
if core.is_float16_supported(place):
|
|
self.check_output_with_place(place)
|
|
|
|
|
|
class TestLambFP16Op3(TestLambOp3):
|
|
def set_dtype(self):
|
|
self.__class__.op_type = "lamb"
|
|
self.dtype = np.float16
|
|
|
|
def test_check_output(self):
|
|
if core.is_compiled_with_cuda() or is_custom_device():
|
|
place = get_device_place()
|
|
if core.is_float16_supported(place):
|
|
self.check_output_with_place(place)
|
|
|
|
|
|
class TestLambFP16OpMultipleSteps(TestLambOpMultipleSteps):
|
|
def set_dtype(self):
|
|
self.__class__.op_type = "lamb"
|
|
self.dtype = np.float16
|
|
|
|
|
|
class TestLambBF16Op1(TestLambOp1):
|
|
def set_dtype(self):
|
|
self.__class__.op_type = "lamb"
|
|
self.dtype = np.uint16
|
|
|
|
def test_check_output(self):
|
|
if core.is_compiled_with_cuda() or is_custom_device():
|
|
place = get_device_place()
|
|
if core.is_bfloat16_supported(place):
|
|
self.check_output_with_place(place)
|
|
|
|
|
|
class TestLambBF16Op2(TestLambOp2):
|
|
def set_dtype(self):
|
|
self.__class__.op_type = "lamb"
|
|
self.dtype = np.uint16
|
|
|
|
def test_check_output(self):
|
|
if core.is_compiled_with_cuda() or is_custom_device():
|
|
place = get_device_place()
|
|
if core.is_bfloat16_supported(place):
|
|
self.check_output_with_place(place)
|
|
|
|
|
|
class TestLambBF16Op3(TestLambOp3):
|
|
def set_dtype(self):
|
|
self.__class__.op_type = "lamb"
|
|
self.dtype = np.uint16
|
|
|
|
def test_check_output(self):
|
|
if core.is_compiled_with_cuda() or is_custom_device():
|
|
place = get_device_place()
|
|
if core.is_bfloat16_supported(place):
|
|
self.check_output_with_place(place)
|
|
|
|
|
|
class TestLambBF16OpMultipleSteps(TestLambOpMultipleSteps):
|
|
def set_dtype(self):
|
|
self.__class__.op_type = "lamb"
|
|
self.dtype = np.uint16
|
|
|
|
def test_check_output(self):
|
|
for i in range(self.num_steps):
|
|
(
|
|
param_out,
|
|
moment1_out,
|
|
moment2_out,
|
|
beta1_pow_out,
|
|
beta2_pow_out,
|
|
) = lamb_step(self.inputs, self.attrs)
|
|
|
|
self.outputs = {
|
|
'Moment1Out': convert_float_to_uint16(moment1_out),
|
|
'Moment2Out': convert_float_to_uint16(moment2_out),
|
|
'ParamOut': convert_float_to_uint16(param_out),
|
|
'Beta1PowOut': convert_float_to_uint16(beta1_pow_out),
|
|
'Beta2PowOut': convert_float_to_uint16(beta2_pow_out),
|
|
}
|
|
|
|
# Verify output for this step
|
|
if core.is_compiled_with_cuda() or is_custom_device():
|
|
place = get_device_place()
|
|
if core.is_bfloat16_supported(place):
|
|
self.check_output_with_place(place)
|
|
|
|
# Output of this step becomes input for next step
|
|
self.inputs['Param'] = convert_float_to_uint16(param_out)
|
|
self.inputs['Moment1'] = convert_float_to_uint16(moment1_out)
|
|
self.inputs['Moment2'] = convert_float_to_uint16(moment2_out)
|
|
|
|
# Update powers of Beta1 and Beta2 for next time step
|
|
self.inputs['Beta1Pow'] = convert_float_to_uint16(beta1_pow_out)
|
|
self.inputs['Beta2Pow'] = convert_float_to_uint16(beta2_pow_out)
|
|
|
|
# Randomize gradient for next step
|
|
self.inputs['Grad'] = convert_float_to_uint16(
|
|
np.random.uniform(-1, 1, (102, 105)).astype(np.float32)
|
|
)
|
|
|
|
|
|
def lamb_step(inputs, attributes):
|
|
'''
|
|
Simulate one step of the lamb optimizer
|
|
:param inputs: dict of inputs
|
|
:param attributes: dict of attributes
|
|
:return tuple: tuple of output param, moment1, moment2,
|
|
beta1 power accumulator and beta2 power accumulator
|
|
'''
|
|
param = inputs['Param']
|
|
grad = inputs['Grad']
|
|
moment1 = inputs['Moment1']
|
|
moment2 = inputs['Moment2']
|
|
lr = inputs['LearningRate']
|
|
beta1_pow = inputs['Beta1Pow']
|
|
beta2_pow = inputs['Beta2Pow']
|
|
|
|
beta1 = attributes['beta1']
|
|
beta2 = attributes['beta2']
|
|
epsilon = attributes['epsilon']
|
|
weight_decay = attributes['weight_decay']
|
|
always_adapt = attributes['always_adapt']
|
|
|
|
moment1_out = beta1 * moment1 + (1 - beta1) * grad
|
|
moment2_out = beta2 * moment2 + (1 - beta2) * np.square(grad)
|
|
|
|
moment1_unbiased = moment1_out / (1 - beta1_pow)
|
|
moment2_unbiased = moment2_out / (1 - beta2_pow)
|
|
|
|
if weight_decay > 0 or always_adapt:
|
|
r_1 = np.linalg.norm(param)
|
|
r_2 = np.linalg.norm(
|
|
moment1_unbiased / (np.sqrt(moment2_unbiased) + epsilon)
|
|
+ weight_decay * param
|
|
)
|
|
lr_t = lr * r_1 / r_2
|
|
else:
|
|
lr_t = lr
|
|
|
|
param_out = param - lr_t * (
|
|
moment1_unbiased / (np.sqrt(moment2_unbiased) + epsilon)
|
|
+ weight_decay * param
|
|
)
|
|
|
|
beta1_pow_out = beta1_pow * beta1
|
|
beta2_pow_out = beta2_pow * beta2
|
|
|
|
return param_out, moment1_out, moment2_out, beta1_pow_out, beta2_pow_out
|
|
|
|
|
|
def lamb_step_sparse(inputs, attributes, height, rows, row_numel, np_grad):
|
|
'''
|
|
Simulate one step of the lamb optimizer
|
|
:param inputs: dict of inputs
|
|
:param attributes: dict of attributes
|
|
:return tuple: tuple of output param, moment1, moment2,
|
|
beta1 power accumulator and beta2 power accumulator
|
|
'''
|
|
param = inputs['Param']
|
|
# grad = inputs['Grad']
|
|
moment1 = inputs['Moment1']
|
|
moment2 = inputs['Moment2']
|
|
lr = inputs['LearningRate']
|
|
beta1_pow = inputs['Beta1Pow']
|
|
beta2_pow = inputs['Beta2Pow']
|
|
|
|
beta1 = attributes['beta1']
|
|
beta2 = attributes['beta2']
|
|
epsilon = attributes['epsilon']
|
|
weight_decay = attributes['weight_decay']
|
|
always_adapt = attributes['always_adapt']
|
|
|
|
moment1_out = np.zeros(shape=[height, row_numel])
|
|
moment2_out = np.zeros(shape=[height, row_numel])
|
|
param_out = np.zeros(shape=[height, row_numel])
|
|
moment1_unbiased = np.zeros(shape=[height, row_numel])
|
|
moment2_unbiased = np.zeros(shape=[height, row_numel])
|
|
|
|
def update_mom(row_id, update_value):
|
|
moment1_out[row_id] = (
|
|
beta1 * moment1[row_id] + (1 - beta1) * update_value
|
|
)
|
|
moment2_out[row_id] = beta2 * moment2[row_id] + (1 - beta2) * np.square(
|
|
update_value
|
|
)
|
|
|
|
moment1_out[row_id] = (
|
|
beta1 * moment1[row_id] + (1 - beta1) * update_value
|
|
)
|
|
moment2_out[row_id] = beta2 * moment2[row_id] + (1 - beta2) * np.square(
|
|
update_value
|
|
)
|
|
|
|
def update_param(weight_decay, always_adapt):
|
|
if weight_decay > 0 or always_adapt:
|
|
r_1 = np.linalg.norm(param)
|
|
r_2 = np.linalg.norm(
|
|
moment1_out / (np.sqrt(moment2_out) + epsilon)
|
|
+ weight_decay * param
|
|
)
|
|
lr_t = lr * r_1 / r_2
|
|
else:
|
|
lr_t = lr
|
|
|
|
param_out = param - lr_t * (
|
|
moment1_out / (np.sqrt(moment2_out) + epsilon)
|
|
+ weight_decay * param
|
|
)
|
|
|
|
for row_id in range(param_out.shape[0]):
|
|
update_value = np.zeros(np_grad[0].shape).astype("float32")
|
|
if row_id in rows:
|
|
update_value = np_grad[rows.index(row_id)]
|
|
update_mom(row_id, update_value)
|
|
|
|
update_param(weight_decay, always_adapt)
|
|
beta1_pow_out = beta1_pow * beta1
|
|
beta2_pow_out = beta2_pow * beta2
|
|
|
|
return param_out, moment1_out, moment2_out, beta1_pow_out, beta2_pow_out
|
|
|
|
|
|
class TestSparseLambOp(unittest.TestCase):
|
|
def setup(self, scope, place):
|
|
beta1 = 0.78
|
|
beta2 = 0.836
|
|
epsilon = 1e-4
|
|
|
|
height = 10
|
|
rows = [0, 4, 7]
|
|
self.rows = rows
|
|
row_numel = 12
|
|
self.row_numel = row_numel
|
|
self.dense_inputs = {
|
|
"Param": np.full((height, row_numel), 5.0).astype("float32"),
|
|
"Moment1": np.full((height, row_numel), 5.0).astype("float32"),
|
|
"Moment2": np.full((height, row_numel), 5.0).astype("float32"),
|
|
'Beta1Pow': np.array([beta1]).astype("float32"),
|
|
'Beta2Pow': np.array([beta2]).astype("float32"),
|
|
"LearningRate": np.full((1), 2.0).astype("float32"),
|
|
}
|
|
self.init_output = np.full((height, row_numel), 0.0).astype("float32")
|
|
self.attrs = {
|
|
'epsilon': epsilon,
|
|
'beta1': beta1,
|
|
'beta2': beta2,
|
|
'weight_decay': 0.05,
|
|
'always_adapt': False,
|
|
}
|
|
|
|
grad_selected_rows = scope.var('Grad').get_selected_rows()
|
|
grad_selected_rows.set_height(height)
|
|
grad_selected_rows.set_rows(rows)
|
|
np_array = np.ones((len(rows), row_numel)).astype("float32")
|
|
np_array[0, 0] = 2.0
|
|
np_array[2, 8] = 4.0
|
|
|
|
grad_tensor = grad_selected_rows.get_tensor()
|
|
grad_tensor.set(np_array, place)
|
|
|
|
self.sparse_inputs = ["Grad"]
|
|
|
|
param_out, mom1, mom2, beta1_pow_out, beta2_pow_out = lamb_step_sparse(
|
|
self.dense_inputs, self.attrs, height, rows, row_numel, np_array
|
|
)
|
|
self.outputs = {
|
|
"ParamOut": param_out,
|
|
"Moment1Out": mom1,
|
|
"Moment2Out": mom2,
|
|
'Beta1PowOut': beta1_pow_out,
|
|
'Beta2PowOut': beta2_pow_out,
|
|
}
|
|
|
|
def check_with_place(self, place):
|
|
scope = core.Scope()
|
|
self.setup(scope, place)
|
|
|
|
op_args = {}
|
|
for key, np_array in self.dense_inputs.items():
|
|
var = scope.var(key).get_tensor()
|
|
var.set(np_array, place)
|
|
op_args[key] = key
|
|
for s in self.sparse_inputs:
|
|
op_args[s] = s
|
|
for s in self.outputs:
|
|
var = scope.var(s).get_tensor()
|
|
var.set(self.init_output, place)
|
|
op_args[s] = s
|
|
for k in self.attrs:
|
|
op_args[k] = self.attrs[k]
|
|
|
|
# create and run sgd operator
|
|
lamb_op = Operator("lamb", **op_args)
|
|
lamb_op.run(scope, place)
|
|
|
|
for key, np_array in self.outputs.items():
|
|
out_var = scope.var(key).get_tensor()
|
|
actual = np.array(out_var)
|
|
actual = actual.reshape([actual.size])
|
|
np_array = np_array.reshape([np_array.size])
|
|
|
|
for i in range(np_array.size):
|
|
self.assertLess((actual[i] - np_array[i]), 0.00001)
|
|
|
|
def test_sparse_lamb(self):
|
|
for place in get_places():
|
|
self.check_with_place(place)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|