wehub-resource-sync
609 lines
19 KiB
Python
609 lines
19 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 logging
|
|
import math
|
|
import os
|
|
import tempfile
|
|
import time
|
|
import unittest
|
|
|
|
import numpy as np
|
|
from dygraph_to_static_utils import (
|
|
Dy2StTestBase,
|
|
enable_to_static_guard,
|
|
test_default_mode_only,
|
|
)
|
|
from predictor_utils import PredictorTools
|
|
|
|
import paddle
|
|
from paddle import base
|
|
from paddle.jit.pir_translated_layer import PIR_INFER_MODEL_SUFFIX
|
|
from paddle.jit.translated_layer import INFER_MODEL_SUFFIX, INFER_PARAMS_SUFFIX
|
|
from paddle.nn import BatchNorm, Linear
|
|
from paddle.static import InputSpec
|
|
|
|
SEED = 2020
|
|
np.random.seed(SEED)
|
|
|
|
BATCH_SIZE = 8
|
|
EPOCH_NUM = 1
|
|
PRINT_STEP = 2
|
|
STEP_NUM = 10
|
|
|
|
place = (
|
|
paddle.CUDAPlace(0) if paddle.is_compiled_with_cuda() else paddle.CPUPlace()
|
|
)
|
|
|
|
# Note: Set True to eliminate randomness.
|
|
# 1. For one operation, cuDNN has several algorithms,
|
|
# some algorithm results are non-deterministic, like convolution algorithms.
|
|
if paddle.is_compiled_with_cuda():
|
|
paddle.set_flags({'FLAGS_cudnn_deterministic': True})
|
|
|
|
train_parameters = {
|
|
"learning_strategy": {
|
|
"name": "cosine_decay",
|
|
"batch_size": BATCH_SIZE,
|
|
"epochs": [40, 80, 100],
|
|
"steps": [0.1, 0.01, 0.001, 0.0001],
|
|
},
|
|
"lr": 0.0125,
|
|
"total_images": 6149,
|
|
"momentum_rate": 0.9,
|
|
"l2_decay": 1.2e-4,
|
|
"num_epochs": 1,
|
|
}
|
|
|
|
|
|
def optimizer_setting(params, parameter_list):
|
|
ls = params["learning_strategy"]
|
|
if "total_images" not in params:
|
|
total_images = 6149
|
|
else:
|
|
total_images = params["total_images"]
|
|
|
|
batch_size = ls["batch_size"]
|
|
l2_decay = params["l2_decay"]
|
|
momentum_rate = params["momentum_rate"]
|
|
|
|
step = int(math.ceil(float(total_images) / batch_size))
|
|
bd = [step * e for e in ls["epochs"]]
|
|
lr = params["lr"]
|
|
num_epochs = params["num_epochs"]
|
|
optimizer = paddle.optimizer.Momentum(
|
|
learning_rate=paddle.optimizer.lr.CosineAnnealingDecay(
|
|
learning_rate=lr, T_max=num_epochs
|
|
),
|
|
momentum=momentum_rate,
|
|
weight_decay=paddle.regularizer.L2Decay(l2_decay),
|
|
parameters=parameter_list,
|
|
)
|
|
|
|
return optimizer
|
|
|
|
|
|
class ConvBNLayer(paddle.nn.Layer):
|
|
def __init__(
|
|
self,
|
|
num_channels,
|
|
num_filters,
|
|
filter_size,
|
|
stride=1,
|
|
groups=1,
|
|
act=None,
|
|
):
|
|
super().__init__()
|
|
|
|
self._conv = paddle.nn.Conv2D(
|
|
in_channels=num_channels,
|
|
out_channels=num_filters,
|
|
kernel_size=filter_size,
|
|
stride=stride,
|
|
padding=(filter_size - 1) // 2,
|
|
groups=groups,
|
|
bias_attr=False,
|
|
)
|
|
|
|
self._batch_norm = BatchNorm(num_filters, act=act)
|
|
|
|
def forward(self, inputs):
|
|
y = self._conv(inputs)
|
|
y = self._batch_norm(y)
|
|
|
|
return y
|
|
|
|
|
|
class SqueezeExcitation(paddle.nn.Layer):
|
|
def __init__(self, num_channels, reduction_ratio):
|
|
super().__init__()
|
|
self._num_channels = num_channels
|
|
self._pool = paddle.nn.AdaptiveAvgPool2D(1)
|
|
stdv = 1.0 / math.sqrt(num_channels * 1.0)
|
|
self._fc = Linear(
|
|
num_channels,
|
|
num_channels // reduction_ratio,
|
|
weight_attr=base.ParamAttr(
|
|
initializer=paddle.nn.initializer.Uniform(-stdv, stdv)
|
|
),
|
|
)
|
|
stdv = 1.0 / math.sqrt(num_channels / 16.0 * 1.0)
|
|
self._excitation = Linear(
|
|
num_channels // reduction_ratio,
|
|
num_channels,
|
|
weight_attr=base.ParamAttr(
|
|
initializer=paddle.nn.initializer.Uniform(-stdv, stdv)
|
|
),
|
|
)
|
|
|
|
def forward(self, input):
|
|
y = self._pool(input)
|
|
y = paddle.reshape(y, shape=[-1, self._num_channels])
|
|
y = self._fc(y)
|
|
y = paddle.nn.functional.relu(y)
|
|
y = self._excitation(y)
|
|
y = paddle.nn.functional.sigmoid(y)
|
|
y = paddle.tensor.math._multiply_with_axis(x=input, y=y, axis=0)
|
|
return y
|
|
|
|
|
|
class BottleneckBlock(paddle.nn.Layer):
|
|
def __init__(
|
|
self,
|
|
num_channels,
|
|
num_filters,
|
|
stride,
|
|
cardinality,
|
|
reduction_ratio,
|
|
shortcut=True,
|
|
):
|
|
super().__init__()
|
|
|
|
self.conv0 = ConvBNLayer(
|
|
num_channels=num_channels,
|
|
num_filters=num_filters,
|
|
filter_size=1,
|
|
act="relu",
|
|
)
|
|
self.conv1 = ConvBNLayer(
|
|
num_channels=num_filters,
|
|
num_filters=num_filters,
|
|
filter_size=3,
|
|
stride=stride,
|
|
groups=cardinality,
|
|
act="relu",
|
|
)
|
|
self.conv2 = ConvBNLayer(
|
|
num_channels=num_filters,
|
|
num_filters=num_filters * 2,
|
|
filter_size=1,
|
|
act=None,
|
|
)
|
|
|
|
self.scale = SqueezeExcitation(
|
|
num_channels=num_filters * 2, reduction_ratio=reduction_ratio
|
|
)
|
|
|
|
if not shortcut:
|
|
self.short = ConvBNLayer(
|
|
num_channels=num_channels,
|
|
num_filters=num_filters * 2,
|
|
filter_size=1,
|
|
stride=stride,
|
|
)
|
|
|
|
self.shortcut = shortcut
|
|
|
|
self._num_channels_out = num_filters * 2
|
|
|
|
def forward(self, inputs):
|
|
y = self.conv0(inputs)
|
|
conv1 = self.conv1(y)
|
|
conv2 = self.conv2(conv1)
|
|
scale = self.scale(conv2)
|
|
|
|
if self.shortcut:
|
|
short = inputs
|
|
else:
|
|
short = self.short(inputs)
|
|
|
|
y = paddle.nn.functional.relu(paddle.add(x=short, y=scale))
|
|
return y
|
|
|
|
|
|
class SeResNeXt(paddle.nn.Layer):
|
|
def __init__(self, layers=50, class_dim=102):
|
|
super().__init__()
|
|
|
|
self.layers = layers
|
|
supported_layers = [50, 101, 152]
|
|
assert layers in supported_layers, (
|
|
f"supported layers are {supported_layers} but input layer is {layers}"
|
|
)
|
|
|
|
if layers == 50:
|
|
cardinality = 32
|
|
reduction_ratio = 16
|
|
depth = [3, 4, 6, 3]
|
|
num_filters = [128, 256, 512, 1024]
|
|
self.conv0 = ConvBNLayer(
|
|
num_channels=3,
|
|
num_filters=64,
|
|
filter_size=7,
|
|
stride=2,
|
|
act='relu',
|
|
)
|
|
self.pool = paddle.nn.MaxPool2D(kernel_size=3, stride=2, padding=1)
|
|
elif layers == 101:
|
|
cardinality = 32
|
|
reduction_ratio = 16
|
|
depth = [3, 4, 23, 3]
|
|
num_filters = [128, 256, 512, 1024]
|
|
self.conv0 = ConvBNLayer(
|
|
num_channels=3,
|
|
num_filters=64,
|
|
filter_size=7,
|
|
stride=2,
|
|
act='relu',
|
|
)
|
|
self.pool = paddle.nn.MaxPool2D(kernel_size=3, stride=2, padding=1)
|
|
elif layers == 152:
|
|
cardinality = 64
|
|
reduction_ratio = 16
|
|
depth = [3, 8, 36, 3]
|
|
num_filters = [128, 256, 512, 1024]
|
|
self.conv0 = ConvBNLayer(
|
|
num_channels=3,
|
|
num_filters=64,
|
|
filter_size=3,
|
|
stride=2,
|
|
act='relu',
|
|
)
|
|
self.conv1 = ConvBNLayer(
|
|
num_channels=64,
|
|
num_filters=64,
|
|
filter_size=3,
|
|
stride=1,
|
|
act='relu',
|
|
)
|
|
self.conv2 = ConvBNLayer(
|
|
num_channels=64,
|
|
num_filters=128,
|
|
filter_size=3,
|
|
stride=1,
|
|
act='relu',
|
|
)
|
|
self.pool = paddle.nn.MaxPool2D(kernel_size=3, stride=2, padding=1)
|
|
|
|
self.bottleneck_block_list = []
|
|
num_channels = 64
|
|
if layers == 152:
|
|
num_channels = 128
|
|
for block in range(len(depth)):
|
|
shortcut = False
|
|
for i in range(depth[block]):
|
|
bottleneck_block = self.add_sublayer(
|
|
f'bb_{block}_{i}',
|
|
BottleneckBlock(
|
|
num_channels=num_channels,
|
|
num_filters=num_filters[block],
|
|
stride=2 if i == 0 and block != 0 else 1,
|
|
cardinality=cardinality,
|
|
reduction_ratio=reduction_ratio,
|
|
shortcut=shortcut,
|
|
),
|
|
)
|
|
num_channels = bottleneck_block._num_channels_out
|
|
self.bottleneck_block_list.append(bottleneck_block)
|
|
shortcut = True
|
|
self.pool2d_avg = paddle.nn.AdaptiveAvgPool2D(1)
|
|
stdv = 1.0 / math.sqrt(2048 * 1.0)
|
|
|
|
self.pool2d_avg_output = num_filters[len(num_filters) - 1] * 2 * 1 * 1
|
|
|
|
self.dropout = paddle.nn.Dropout(p=0.5, mode="downscale_in_infer")
|
|
|
|
self.out = Linear(
|
|
self.pool2d_avg_output,
|
|
class_dim,
|
|
weight_attr=base.param_attr.ParamAttr(
|
|
initializer=paddle.nn.initializer.Uniform(-stdv, stdv)
|
|
),
|
|
)
|
|
|
|
def forward(self, inputs, label):
|
|
if self.layers == 50 or self.layers == 101:
|
|
y = self.conv0(inputs)
|
|
y = self.pool(y)
|
|
elif self.layers == 152:
|
|
y = self.conv0(inputs)
|
|
y = self.conv1(y)
|
|
y = self.conv2(y)
|
|
y = self.pool(y)
|
|
|
|
for bottleneck_block in self.bottleneck_block_list:
|
|
y = bottleneck_block(y)
|
|
|
|
y = self.pool2d_avg(y)
|
|
y = self.dropout(y)
|
|
y = paddle.reshape(y, shape=[-1, self.pool2d_avg_output])
|
|
out = self.out(y)
|
|
|
|
softmax_out = paddle.nn.functional.softmax(out)
|
|
loss = paddle.nn.functional.cross_entropy(
|
|
input=softmax_out, label=label, reduction='none', use_softmax=False
|
|
)
|
|
avg_loss = paddle.mean(x=loss)
|
|
|
|
acc_top1 = paddle.static.accuracy(input=softmax_out, label=label, k=1)
|
|
acc_top5 = paddle.static.accuracy(input=softmax_out, label=label, k=5)
|
|
return out, avg_loss, acc_top1, acc_top5
|
|
|
|
|
|
class TestSeResnet(Dy2StTestBase):
|
|
def setUp(self):
|
|
self.train_reader = paddle.batch(
|
|
paddle.dataset.flowers.train(use_xmap=False, cycle=True),
|
|
batch_size=BATCH_SIZE,
|
|
drop_last=True,
|
|
)
|
|
self.temp_dir = tempfile.TemporaryDirectory()
|
|
|
|
self.model_save_dir = os.path.join(self.temp_dir.name, "inference")
|
|
self.model_save_prefix = os.path.join(
|
|
self.temp_dir.name, "inference/se_resnet"
|
|
)
|
|
self.model_filename = "se_resnet" + INFER_MODEL_SUFFIX
|
|
self.pir_model_filename = "se_resnet" + PIR_INFER_MODEL_SUFFIX
|
|
self.params_filename = "se_resnet" + INFER_PARAMS_SUFFIX
|
|
self.dy_state_dict_save_path = os.path.join(
|
|
self.temp_dir.name, "se_resnet.dygraph"
|
|
)
|
|
|
|
def tearDown(self):
|
|
self.temp_dir.cleanup()
|
|
|
|
def train(self, train_reader, to_static):
|
|
np.random.seed(SEED)
|
|
|
|
with base.dygraph.guard(place):
|
|
paddle.seed(SEED)
|
|
paddle.framework.random._manual_program_seed(SEED)
|
|
se_resnext = SeResNeXt()
|
|
se_resnext = paddle.jit.to_static(se_resnext, full_graph=True)
|
|
optimizer = optimizer_setting(
|
|
train_parameters, se_resnext.parameters()
|
|
)
|
|
|
|
for epoch_id in range(EPOCH_NUM):
|
|
total_loss = 0.0
|
|
total_acc1 = 0.0
|
|
total_acc5 = 0.0
|
|
total_sample = 0
|
|
step_idx = 0
|
|
speed_list = []
|
|
for step_id, data in enumerate(train_reader()):
|
|
dy_x_data = np.array(
|
|
[x[0].reshape(3, 224, 224) for x in data]
|
|
).astype('float32')
|
|
y_data = (
|
|
np.array([x[1] for x in data])
|
|
.astype('int64')
|
|
.reshape(BATCH_SIZE, 1)
|
|
)
|
|
|
|
img = paddle.to_tensor(dy_x_data)
|
|
label = paddle.to_tensor(y_data)
|
|
label.stop_gradient = True
|
|
|
|
pred, avg_loss, acc_top1, acc_top5 = se_resnext(img, label)
|
|
|
|
dy_out = avg_loss.numpy()
|
|
avg_loss.backward()
|
|
|
|
optimizer.minimize(avg_loss)
|
|
se_resnext.clear_gradients()
|
|
|
|
lr = optimizer._global_learning_rate().numpy()
|
|
total_loss += dy_out
|
|
total_acc1 += acc_top1.numpy()
|
|
total_acc5 += acc_top5.numpy()
|
|
total_sample += 1
|
|
if step_id % PRINT_STEP == 0:
|
|
if step_id == 0:
|
|
logging.info(
|
|
f"epoch {epoch_id} | step {step_id}, "
|
|
f"loss {total_loss / total_sample:0.3f}, "
|
|
f"acc1 {total_acc1 / total_sample:0.3f}, "
|
|
f"acc5 {total_acc5 / total_sample:0.3f}"
|
|
)
|
|
avg_batch_time = time.time()
|
|
else:
|
|
speed = PRINT_STEP / (time.time() - avg_batch_time)
|
|
speed_list.append(speed)
|
|
logging.info(
|
|
f"epoch {epoch_id} | step {step_id}, "
|
|
f"loss {total_loss / total_sample:0.3f}, "
|
|
f"acc1 {total_acc1 / total_sample:0.3f}, "
|
|
f"acc5 {total_acc5 / total_sample:0.3f}, "
|
|
f"speed {speed:.3f} steps/s"
|
|
)
|
|
avg_batch_time = time.time()
|
|
|
|
step_idx += 1
|
|
if step_idx == STEP_NUM:
|
|
if to_static:
|
|
output_spec = [0]
|
|
|
|
paddle.jit.save(
|
|
se_resnext,
|
|
self.model_save_prefix,
|
|
output_spec=output_spec,
|
|
input_names_after_prune=['x'],
|
|
input_spec=[
|
|
InputSpec(
|
|
shape=[None, 3, 224, 224], name='x'
|
|
),
|
|
InputSpec(shape=[None, 1], name='y'),
|
|
],
|
|
clip_extra=False,
|
|
)
|
|
else:
|
|
paddle.save(
|
|
se_resnext.state_dict(),
|
|
self.dy_state_dict_save_path + '.pdparams',
|
|
)
|
|
break
|
|
return (
|
|
pred.numpy(),
|
|
avg_loss.numpy(),
|
|
acc_top1.numpy(),
|
|
acc_top5.numpy(),
|
|
)
|
|
|
|
def predict_dygraph(self, data):
|
|
with (
|
|
enable_to_static_guard(False),
|
|
base.dygraph.guard(place),
|
|
):
|
|
se_resnext = SeResNeXt()
|
|
|
|
model_dict = paddle.load(self.dy_state_dict_save_path + '.pdparams')
|
|
se_resnext.set_dict(model_dict)
|
|
se_resnext.eval()
|
|
|
|
label = np.random.random([1, 1]).astype("int64")
|
|
img = paddle.to_tensor(data)
|
|
label = paddle.to_tensor(label)
|
|
pred_res, _, _, _ = se_resnext(img, label)
|
|
|
|
return pred_res.numpy()
|
|
|
|
def predict_static(self, data):
|
|
paddle.enable_static()
|
|
model_filename = self.pir_model_filename
|
|
|
|
exe = base.Executor(place)
|
|
[
|
|
inference_program,
|
|
feed_target_names,
|
|
fetch_targets,
|
|
] = paddle.static.io.load_inference_model(
|
|
self.model_save_dir,
|
|
executor=exe,
|
|
model_filename=model_filename,
|
|
params_filename=self.params_filename,
|
|
)
|
|
|
|
pred_res = exe.run(
|
|
inference_program,
|
|
feed={feed_target_names[0]: data},
|
|
fetch_list=fetch_targets,
|
|
)
|
|
|
|
return pred_res[0]
|
|
|
|
def predict_dygraph_jit(self, data):
|
|
with base.dygraph.guard(place):
|
|
se_resnext = paddle.jit.load(self.model_save_prefix)
|
|
se_resnext.eval()
|
|
|
|
pred_res = se_resnext(data)
|
|
|
|
return pred_res.numpy()
|
|
|
|
def predict_analysis_inference(self, data):
|
|
model_filename = self.pir_model_filename
|
|
|
|
output = PredictorTools(
|
|
self.model_save_dir,
|
|
model_filename,
|
|
self.params_filename,
|
|
[data],
|
|
)
|
|
(out,) = output()
|
|
return out
|
|
|
|
def verify_predict(self):
|
|
image = np.random.random([1, 3, 224, 224]).astype('float32')
|
|
dy_pre = self.predict_dygraph(image)
|
|
st_pre = self.predict_static(image)
|
|
dy_jit_pre = self.predict_dygraph_jit(image)
|
|
|
|
np.testing.assert_allclose(
|
|
dy_pre,
|
|
st_pre,
|
|
rtol=1e-05,
|
|
err_msg=f'dy_pre:\n {dy_pre}\n, st_pre: \n{st_pre}.',
|
|
)
|
|
np.testing.assert_allclose(
|
|
dy_jit_pre,
|
|
st_pre,
|
|
rtol=1e-05,
|
|
err_msg=f'dy_jit_pre:\n {dy_jit_pre}\n, st_pre: \n{st_pre}.',
|
|
)
|
|
|
|
predictor_pre = self.predict_analysis_inference(image)
|
|
flat_st_pre = st_pre.flatten()
|
|
flat_predictor_pre = np.array(predictor_pre).flatten()
|
|
for i in range(len(flat_predictor_pre)):
|
|
# modify precision to 1e-6, avoid unittest failed
|
|
self.assertAlmostEqual(
|
|
flat_predictor_pre[i],
|
|
flat_st_pre[i],
|
|
delta=1e-5,
|
|
msg=f"predictor_pre:\n {flat_predictor_pre[i]}\n, st_pre: \n{flat_st_pre[i]}.",
|
|
)
|
|
|
|
@test_default_mode_only
|
|
def test_check_result(self):
|
|
with enable_to_static_guard(False):
|
|
pred_1, loss_1, acc1_1, acc5_1 = self.train(
|
|
self.train_reader, to_static=False
|
|
)
|
|
pred_2, loss_2, acc1_2, acc5_2 = self.train(
|
|
self.train_reader, to_static=True
|
|
)
|
|
|
|
np.testing.assert_allclose(
|
|
pred_1,
|
|
pred_2,
|
|
rtol=1e-05,
|
|
err_msg=f'static pred: {pred_1} \ndygraph pred: {pred_2}',
|
|
)
|
|
np.testing.assert_allclose(
|
|
loss_1,
|
|
loss_2,
|
|
rtol=1e-05,
|
|
err_msg=f'static loss: {loss_1} \ndygraph loss: {loss_2}',
|
|
)
|
|
np.testing.assert_allclose(
|
|
acc1_1,
|
|
acc1_2,
|
|
rtol=1e-05,
|
|
err_msg=f'static acc1: {acc1_1} \ndygraph acc1: {acc1_2}',
|
|
)
|
|
np.testing.assert_allclose(
|
|
acc5_1,
|
|
acc5_2,
|
|
rtol=1e-05,
|
|
err_msg=f'static acc5: {acc5_1} \ndygraph acc5: {acc5_2}',
|
|
)
|
|
|
|
self.verify_predict()
|
|
|
|
|
|
if __name__ == '__main__':
|
|
unittest.main()
|