提交 d4f1758d 编写于 作者: Y yukavio

add slim/prune

上级 ed6b2f0c
> 运行示例前请先安装develop版本PaddleSlim
# 模型裁剪压缩教程
## 概述
该示例使用PaddleSlim提供的[裁剪压缩API](https://paddlepaddle.github.io/PaddleSlim/api/prune_api/)对OCR模型进行压缩。
在阅读该示例前,建议您先了解以下内容:
- [OCR模型的常规训练方法](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/doc/doc_ch/detection.md)
- [PaddleSlim使用文档](https://paddlepaddle.github.io/PaddleSlim/)
## 安装PaddleSlim
可按照[PaddleSlim使用文档](https://paddlepaddle.github.io/PaddleSlim/)中的步骤安装PaddleSlim。
## 敏感度分析训练
进入PaddleOCR根目录,通过以下命令对模型进行敏感度分析:
```bash
python deploy/slim/prune/sensitivity_anal.py -c configs/det/det_mv3_db.yml -o Global.pretrain_weights=./deploy/slim/prune/pretrain_models/det_mv3_db/best_accuracy Global.test_batch_size_per_card=1
```
## 裁剪模型与fine-tune
```bash
python deploy/slim/prune/pruning_and_finetune.py -c configs/det/det_mv3_db.yml -o Global.pretrain_weights=./deploy/slim/prune/pretrain_models/det_mv3_db/best_accuracy Global.test_batch_size_per_card=1
```
## 评估并导出
在得到裁剪训练保存的模型后,我们可以将其导出为inference_model,用于预测部署:
```bash
python deploy/slim/prune/export_prune_model.py -c configs/det/det_mv3_db.yml -o Global.pretrain_weights=./output/det_db/best_accuracy Global.test_batch_size_per_card=1 Global.save_inference_dir=inference_model
```
# 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.
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import os
import sys
import logging
import numpy as np
import paddle.fluid as fluid
__dir__ = os.path.dirname(__file__)
sys.path.append(__dir__)
sys.path.append(os.path.join(__dir__, '..', '..', '..'))
__all__ = ['eval_det_run']
import logging
FORMAT = '%(asctime)s-%(levelname)s: %(message)s'
logging.basicConfig(level=logging.INFO, format=FORMAT)
logger = logging.getLogger(__name__)
import cv2
import json
from copy import deepcopy
from ppocr.utils.utility import create_module
from ppocr.data.reader_main import reader_main
from tools.eval_utils.eval_det_iou import DetectionIoUEvaluator
def cal_det_res(exe, config, eval_info_dict):
global_params = config['Global']
save_res_path = global_params['save_res_path']
postprocess_params = deepcopy(config["PostProcess"])
postprocess_params.update(global_params)
postprocess = create_module(postprocess_params['function']) \
(params=postprocess_params)
if not os.path.exists(os.path.dirname(save_res_path)):
os.makedirs(os.path.dirname(save_res_path))
with open(save_res_path, "wb") as fout:
tackling_num = 0
for data in eval_info_dict['reader']():
img_num = len(data)
tackling_num = tackling_num + img_num
logger.info("test tackling num:%d", tackling_num)
img_list = []
ratio_list = []
img_name_list = []
for ino in range(img_num):
img_list.append(data[ino][0])
ratio_list.append(data[ino][1])
img_name_list.append(data[ino][2])
try:
img_list = np.concatenate(img_list, axis=0)
except:
err = "concatenate error usually caused by different input image shapes in evaluation or testing.\n \
Please set \"test_batch_size_per_card\" in main yml as 1\n \
or add \"test_image_shape: [h, w]\" in reader yml for EvalReader."
raise Exception(err)
outs = exe.run(eval_info_dict['program'], \
feed={'image': img_list}, \
fetch_list=eval_info_dict['fetch_varname_list'])
outs_dict = {}
for tno in range(len(outs)):
fetch_name = eval_info_dict['fetch_name_list'][tno]
fetch_value = np.array(outs[tno])
outs_dict[fetch_name] = fetch_value
dt_boxes_list = postprocess(outs_dict, ratio_list)
for ino in range(img_num):
dt_boxes = dt_boxes_list[ino]
img_name = img_name_list[ino]
dt_boxes_json = []
for box in dt_boxes:
tmp_json = {"transcription": ""}
tmp_json['points'] = box.tolist()
dt_boxes_json.append(tmp_json)
otstr = img_name + "\t" + json.dumps(dt_boxes_json) + "\n"
fout.write(otstr.encode())
return
def load_label_infor(label_file_path, do_ignore=False):
img_name_label_dict = {}
with open(label_file_path, "rb") as fin:
lines = fin.readlines()
for line in lines:
substr = line.decode().strip("\n").split("\t")
bbox_infor = json.loads(substr[1])
bbox_num = len(bbox_infor)
for bno in range(bbox_num):
text = bbox_infor[bno]['transcription']
ignore = False
if text == "###" and do_ignore:
ignore = True
bbox_infor[bno]['ignore'] = ignore
img_name_label_dict[os.path.basename(substr[0])] = bbox_infor
return img_name_label_dict
def cal_det_metrics(gt_label_path, save_res_path):
"""
calculate the detection metrics
Args:
gt_label_path(string): The groundtruth detection label file path
save_res_path(string): The saved predicted detection label path
return:
claculated metrics including Hmean, precision and recall
"""
evaluator = DetectionIoUEvaluator()
gt_label_infor = load_label_infor(gt_label_path, do_ignore=True)
dt_label_infor = load_label_infor(save_res_path)
results = []
for img_name in gt_label_infor:
gt_label = gt_label_infor[img_name]
if img_name not in dt_label_infor:
dt_label = []
else:
dt_label = dt_label_infor[img_name]
result = evaluator.evaluate_image(gt_label, dt_label)
results.append(result)
methodMetrics = evaluator.combine_results(results)
return methodMetrics
def eval_det_run(eval_args, mode='eval'):
exe = eval_args['exe']
config = eval_args['config']
eval_info_dict = eval_args['eval_info_dict']
cal_det_res(exe, config, eval_info_dict)
save_res_path = config['Global']['save_res_path']
if mode == "eval":
gt_label_path = config['EvalReader']['label_file_path']
metrics = cal_det_metrics(gt_label_path, save_res_path)
else:
gt_label_path = config['TestReader']['label_file_path']
do_eval = config['TestReader']['do_eval']
if do_eval:
metrics = cal_det_metrics(gt_label_path, save_res_path)
else:
metrics = {}
return metrics['hmean']
# 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.
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import os
import sys
__dir__ = os.path.dirname(os.path.abspath(__file__))
sys.path.append(__dir__)
sys.path.append(os.path.join(__dir__, '..', '..', '..'))
sys.path.append(os.path.join(__dir__, '..', '..', '..', 'tools'))
def set_paddle_flags(**kwargs):
for key, value in kwargs.items():
if os.environ.get(key, None) is None:
os.environ[key] = str(value)
# NOTE(paddle-dev): All of these flags should be
# set before `import paddle`. Otherwise, it would
# not take any effect.
set_paddle_flags(
FLAGS_eager_delete_tensor_gb=0, # enable GC to save memory
)
import program
from paddle import fluid
from ppocr.utils.utility import initial_logger
logger = initial_logger()
from ppocr.utils.save_load import init_model
from paddleslim.prune import load_model
def main():
startup_prog, eval_program, place, config, _ = program.preprocess()
feeded_var_names, target_vars, fetches_var_name = program.build_export(
config, eval_program, startup_prog)
eval_program = eval_program.clone(for_test=True)
exe = fluid.Executor(place)
exe.run(startup_prog)
if config['Global']['checkpoints'] is not None:
path = config['Global']['checkpoints']
else:
path = config['Global']['pretrain_weights']
load_model(exe, eval_program, path)
save_inference_dir = config['Global']['save_inference_dir']
if not os.path.exists(save_inference_dir):
os.makedirs(save_inference_dir)
fluid.io.save_inference_model(
dirname=save_inference_dir,
feeded_var_names=feeded_var_names,
main_program=eval_program,
target_vars=target_vars,
executor=exe,
model_filename='model',
params_filename='params')
print("inference model saved in {}/model and {}/params".format(
save_inference_dir, save_inference_dir))
print("save success, output_name_list:", fetches_var_name)
if __name__ == '__main__':
main()
# 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.
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import os
import sys
import numpy as np
__dir__ = os.path.dirname(__file__)
sys.path.append(__dir__)
sys.path.append(os.path.join(__dir__, '..', '..', '..'))
sys.path.append(os.path.join(__dir__, '..', '..', '..', 'tools'))
def set_paddle_flags(**kwargs):
for key, value in kwargs.items():
if os.environ.get(key, None) is None:
os.environ[key] = str(value)
# NOTE(paddle-dev): All of these flags should be
# set before `import paddle`. Otherwise, it would
# not take any effect.
set_paddle_flags(
FLAGS_eager_delete_tensor_gb=0, # enable GC to save memory
)
import tools.program as program
from paddle import fluid
from ppocr.utils.utility import initial_logger
logger = initial_logger()
from ppocr.data.reader_main import reader_main
from ppocr.utils.save_load import init_model
from ppocr.utils.character import CharacterOps
from ppocr.utils.utility import initial_logger
from paddleslim.prune import Pruner, save_model
from paddleslim.analysis import flops
from paddleslim.core.graph_wrapper import *
from paddleslim.prune import load_sensitivities, get_ratios_by_loss, merge_sensitive
logger = initial_logger()
skip_list = [
'conv10_linear_weights', 'conv11_linear_weights', 'conv12_expand_weights',
'conv12_linear_weights', 'conv12_se_2_weights', 'conv13_linear_weights',
'conv2_linear_weights', 'conv4_linear_weights', 'conv5_expand_weights',
'conv5_linear_weights', 'conv5_se_2_weights', 'conv6_linear_weights',
'conv7_linear_weights', 'conv8_expand_weights', 'conv8_linear_weights',
'conv9_expand_weights', 'conv9_linear_weights'
]
def main():
config = program.load_config(FLAGS.config)
program.merge_config(FLAGS.opt)
logger.info(config)
# check if set use_gpu=True in paddlepaddle cpu version
use_gpu = config['Global']['use_gpu']
program.check_gpu(use_gpu)
alg = config['Global']['algorithm']
assert alg in ['EAST', 'DB', 'Rosetta', 'CRNN', 'STARNet', 'RARE']
if alg in ['Rosetta', 'CRNN', 'STARNet', 'RARE']:
config['Global']['char_ops'] = CharacterOps(config['Global'])
place = fluid.CUDAPlace(0) if use_gpu else fluid.CPUPlace()
startup_program = fluid.Program()
train_program = fluid.Program()
train_build_outputs = program.build(
config, train_program, startup_program, mode='train')
train_loader = train_build_outputs[0]
train_fetch_name_list = train_build_outputs[1]
train_fetch_varname_list = train_build_outputs[2]
train_opt_loss_name = train_build_outputs[3]
eval_program = fluid.Program()
eval_build_outputs = program.build(
config, eval_program, startup_program, mode='eval')
eval_fetch_name_list = eval_build_outputs[1]
eval_fetch_varname_list = eval_build_outputs[2]
eval_program = eval_program.clone(for_test=True)
train_reader = reader_main(config=config, mode="train")
train_loader.set_sample_list_generator(train_reader, places=place)
eval_reader = reader_main(config=config, mode="eval")
exe = fluid.Executor(place)
exe.run(startup_program)
# compile program for multi-devices
init_model(config, train_program, exe)
# params = get_pruned_params(train_program)
'''
sens_file = ['sensitivities_'+ str(x) for x in range(0,4)]
sens = []
for f in sens_file:
sens.append(load_sensitivities(f+'.data'))
sen = merge_sensitive(sens)
'''
sen = load_sensitivities("sensitivities_0.data")
for i in skip_list:
sen.pop(i)
back_bone_list = ['conv' + str(x) for x in range(1, 5)]
for i in back_bone_list:
for key in list(sen.keys()):
if i + '_' in key:
sen.pop(key)
ratios = get_ratios_by_loss(sen, 0.03)
logger.info("FLOPs before pruning: {}".format(flops(eval_program)))
pruner = Pruner(criterion='geometry_median')
print("ratios: {}".format(ratios))
pruned_val_program, _, _ = pruner.prune(
eval_program,
fluid.global_scope(),
params=ratios.keys(),
ratios=ratios.values(),
place=place,
only_graph=True)
pruned_program, _, _ = pruner.prune(
train_program,
fluid.global_scope(),
params=ratios.keys(),
ratios=ratios.values(),
place=place)
logger.info("FLOPs after pruning: {}".format(flops(pruned_val_program)))
train_compile_program = program.create_multi_devices_program(
pruned_program, train_opt_loss_name)
train_info_dict = {'compile_program':train_compile_program,\
'train_program':pruned_program,\
'reader':train_loader,\
'fetch_name_list':train_fetch_name_list,\
'fetch_varname_list':train_fetch_varname_list}
eval_info_dict = {'program':pruned_val_program,\
'reader':eval_reader,\
'fetch_name_list':eval_fetch_name_list,\
'fetch_varname_list':eval_fetch_varname_list}
if alg in ['EAST', 'DB']:
program.train_eval_det_run(
config, exe, train_info_dict, eval_info_dict, is_pruning=True)
else:
program.train_eval_rec_run(config, exe, train_info_dict, eval_info_dict)
def test_reader():
config = program.load_config(FLAGS.config)
program.merge_config(FLAGS.opt)
print(config)
train_reader = reader_main(config=config, mode="train")
import time
starttime = time.time()
count = 0
try:
for data in train_reader():
count += 1
if count % 1 == 0:
batch_time = time.time() - starttime
starttime = time.time()
print("reader:", count, len(data), batch_time)
except Exception as e:
logger.info(e)
logger.info("finish reader: {}, Success!".format(count))
if __name__ == '__main__':
parser = program.ArgsParser()
FLAGS = parser.parse_args()
main()
# test_reader()
# 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.
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import os
import sys
__dir__ = os.path.dirname(__file__)
sys.path.append(__dir__)
sys.path.append(os.path.join(__dir__, '..', '..', '..'))
sys.path.append(os.path.join(__dir__, '..', '..', '..', 'tools'))
def set_paddle_flags(**kwargs):
for key, value in kwargs.items():
if os.environ.get(key, None) is None:
os.environ[key] = str(value)
# NOTE(paddle-dev): All of these flags should be
# set before `import paddle`. Otherwise, it would
# not take any effect.
set_paddle_flags(
FLAGS_eager_delete_tensor_gb=0, # enable GC to save memory
)
import json
import cv2
from paddle import fluid
import paddleslim as slim
from copy import deepcopy
from eval_det_utils import eval_det_run
from tools import program
from ppocr.utils.utility import initial_logger
from ppocr.data.reader_main import reader_main
from ppocr.utils.save_load import init_model
from ppocr.utils.character import CharacterOps
from ppocr.utils.utility import create_module
from ppocr.data.reader_main import reader_main
logger = initial_logger()
def get_pruned_params(program):
params = []
for param in program.global_block().all_parameters():
if len(
param.shape
) == 4 and 'depthwise' not in param.name and 'transpose' not in param.name:
params.append(param.name)
return params
def main():
config = program.load_config(FLAGS.config)
program.merge_config(FLAGS.opt)
logger.info(config)
# check if set use_gpu=True in paddlepaddle cpu version
use_gpu = config['Global']['use_gpu']
program.check_gpu(use_gpu)
alg = config['Global']['algorithm']
assert alg in ['EAST', 'DB', 'Rosetta', 'CRNN', 'STARNet', 'RARE']
if alg in ['Rosetta', 'CRNN', 'STARNet', 'RARE']:
config['Global']['char_ops'] = CharacterOps(config['Global'])
place = fluid.CUDAPlace(0) if use_gpu else fluid.CPUPlace()
startup_prog = fluid.Program()
eval_program = fluid.Program()
eval_build_outputs = program.build(
config, eval_program, startup_prog, mode='test')
eval_fetch_name_list = eval_build_outputs[1]
eval_fetch_varname_list = eval_build_outputs[2]
eval_program = eval_program.clone(for_test=True)
exe = fluid.Executor(place)
exe.run(startup_prog)
init_model(config, eval_program, exe)
eval_reader = reader_main(config=config, mode="eval")
eval_info_dict = {'program':eval_program,\
'reader':eval_reader,\
'fetch_name_list':eval_fetch_name_list,\
'fetch_varname_list':eval_fetch_varname_list}
eval_args = dict()
eval_args = {'exe': exe, 'config': config, 'eval_info_dict': eval_info_dict}
metrics = eval_det_run(eval_args)
print("Baseline: {}".format(metrics))
params = get_pruned_params(eval_program)
print('Start to analyze')
sens_0 = slim.prune.sensitivity(
eval_program,
place,
params,
eval_det_run,
sensitivities_file="sensitivities_0.data",
pruned_ratios=[0.1],
eval_args=eval_args,
criterion='geometry_median')
if __name__ == '__main__':
parser = program.ArgsParser()
FLAGS = parser.parse_args()
main()
...@@ -33,6 +33,7 @@ from eval_utils.eval_rec_utils import eval_rec_run ...@@ -33,6 +33,7 @@ from eval_utils.eval_rec_utils import eval_rec_run
from ppocr.utils.save_load import save_model from ppocr.utils.save_load import save_model
import numpy as np import numpy as np
from ppocr.utils.character import cal_predicts_accuracy, cal_predicts_accuracy_srn, CharacterOps from ppocr.utils.character import cal_predicts_accuracy, cal_predicts_accuracy_srn, CharacterOps
import paddleslim as slim
class ArgsParser(ArgumentParser): class ArgsParser(ArgumentParser):
...@@ -238,7 +239,11 @@ def create_multi_devices_program(program, loss_var_name): ...@@ -238,7 +239,11 @@ def create_multi_devices_program(program, loss_var_name):
return compile_program return compile_program
def train_eval_det_run(config, exe, train_info_dict, eval_info_dict): def train_eval_det_run(config,
exe,
train_info_dict,
eval_info_dict,
is_pruning=False):
train_batch_id = 0 train_batch_id = 0
log_smooth_window = config['Global']['log_smooth_window'] log_smooth_window = config['Global']['log_smooth_window']
epoch_num = config['Global']['epoch_num'] epoch_num = config['Global']['epoch_num']
...@@ -294,7 +299,13 @@ def train_eval_det_run(config, exe, train_info_dict, eval_info_dict): ...@@ -294,7 +299,13 @@ def train_eval_det_run(config, exe, train_info_dict, eval_info_dict):
best_batch_id = train_batch_id best_batch_id = train_batch_id
best_epoch = epoch best_epoch = epoch
save_path = save_model_dir + "/best_accuracy" save_path = save_model_dir + "/best_accuracy"
save_model(train_info_dict['train_program'], save_path) if is_pruning:
slim.prune.save_model(
exe, train_info_dict['train_program'],
save_path)
else:
save_model(train_info_dict['train_program'],
save_path)
strs = 'Test iter: {}, metrics:{}, best_hmean:{:.6f}, best_epoch:{}, best_batch_id:{}'.format( strs = 'Test iter: {}, metrics:{}, best_hmean:{:.6f}, best_epoch:{}, best_batch_id:{}'.format(
train_batch_id, metrics, best_eval_hmean, best_epoch, train_batch_id, metrics, best_eval_hmean, best_epoch,
best_batch_id) best_batch_id)
...@@ -305,9 +316,17 @@ def train_eval_det_run(config, exe, train_info_dict, eval_info_dict): ...@@ -305,9 +316,17 @@ def train_eval_det_run(config, exe, train_info_dict, eval_info_dict):
train_loader.reset() train_loader.reset()
if epoch == 0 and save_epoch_step == 1: if epoch == 0 and save_epoch_step == 1:
save_path = save_model_dir + "/iter_epoch_0" save_path = save_model_dir + "/iter_epoch_0"
if is_pruning:
slim.prune.save_model(exe, train_info_dict['train_program'],
save_path)
else:
save_model(train_info_dict['train_program'], save_path) save_model(train_info_dict['train_program'], save_path)
if epoch > 0 and epoch % save_epoch_step == 0: if epoch > 0 and epoch % save_epoch_step == 0:
save_path = save_model_dir + "/iter_epoch_%d" % (epoch) save_path = save_model_dir + "/iter_epoch_%d" % (epoch)
if is_pruning:
slim.prune.save_model(exe, train_info_dict['train_program'],
save_path)
else:
save_model(train_info_dict['train_program'], save_path) save_model(train_info_dict['train_program'], save_path)
return return
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册