"""
BSD 3-Clause License
Copyright (c) Soumith Chintala 2016,
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright notice, this
list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright notice,
this list of conditions and the following disclaimer in the documentation
and/or other materials provided with the distribution.
* Neither the name of the copyright holder nor the names of its
contributors may be used to endorse or promote products derived from
this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
Copyright 2020 Huawei Technologies Co., Ltd
Licensed under the BSD 3-Clause License (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
https://spdx.org/licenses/BSD-3-Clause.html
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 print_function, absolute_import
import argparse
import os.path as osp
import numpy as np
import sys
import torch
from torch import nn
from torch.backends import cudnn
from torch.utils.data import DataLoader
from reid import datasets
from reid import models
from reid.trainers_partloss import Trainer
from reid.evaluators import Evaluator
from reid.utils.data import transforms as T
from reid.utils.data.preprocessor import Preprocessor
from reid.utils.logging import Logger
from reid.utils.serialization import load_checkpoint, save_checkpoint
from apex import amp
import os
def get_data(name, data_dir, height, width, batch_size, workers, device_num):
root = osp.join(data_dir, name)
root = data_dir
dataset = datasets.create(name, root)
normalizer = T.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
num_classes = dataset.num_train_ids
train_transformer = T.Compose([
T.RectScale(height, width),
T.RandomHorizontalFlip(),
T.ToTensor(),
normalizer,
])
test_transformer = T.Compose([
T.RectScale(height, width),
T.ToTensor(),
normalizer,
])
train_sampler = None
if device_num > 1:
train_sampler = torch.utils.data.distributed.DistributedSampler(dataset.train)
train_loader = DataLoader(
Preprocessor(dataset.train, root=osp.join(dataset.images_dir,dataset.train_path),
transform=train_transformer),
batch_size=batch_size, num_workers=workers,
shuffle=False, pin_memory=True, drop_last=True, sampler=train_sampler)
else:
train_loader = DataLoader(
Preprocessor(dataset.train, root=osp.join(dataset.images_dir,dataset.train_path),
transform=train_transformer),
batch_size=batch_size, num_workers=workers,
shuffle=True, pin_memory=True, drop_last=True)
query_loader = DataLoader(
Preprocessor(dataset.query, root=osp.join(dataset.images_dir,dataset.query_path),
transform=test_transformer),
batch_size=batch_size, num_workers=workers,
shuffle=False, pin_memory=True)
gallery_loader = DataLoader(
Preprocessor(dataset.gallery, root=osp.join(dataset.images_dir,dataset.gallery_path),
transform=test_transformer),
batch_size=batch_size, num_workers=workers,
shuffle=False, pin_memory=True)
return dataset, num_classes, train_loader, query_loader, gallery_loader, train_sampler
def main(args):
np.random.seed(args.seed)
torch.manual_seed(args.seed)
if args.device_num == -1:
os.environ["CUDA_VISIBLE_DEVICES"] = '0'
else:
environ_str = '0'
for i in range(1, args.device_num):
environ_str = environ_str + ',%d' % i
os.environ["CUDA_VISIBLE_DEVICES"] = environ_str
if args.npu:
os.environ['MASTER_ADDR'] = args.addr
os.environ['MASTER_PORT'] = '29688'
if args.device_num > 1:
torch.distributed.init_process_group(backend="hccl", rank=args.local_rank, world_size=args.device_num)
torch.npu.manual_seed_all(args.seed)
torch.npu.set_device(args.local_rank)
os.environ['device'] = 'npu'
else:
if args.device_num > 1:
torch.distributed.init_process_group(backend='nccl', init_method='env://')
torch.cuda.manual_seed_all(args.seed)
cudnn.benchmark = True
torch.cuda.set_device(args.local_rank)
os.environ['device'] = 'gpu'
if not args.evaluate:
sys.stdout = Logger(osp.join(args.logs_dir, 'log.txt'))
if args.height is None or args.width is None:
args.height, args.width = (144, 56) if args.arch == 'inception' else \
(256, 128)
dataset, num_classes, train_loader, query_loader, gallery_loader, train_sampler = \
get_data(args.dataset, args.data_dir, args.height,
args.width, args.batch_size, args.workers, args.device_num
)
model = models.create(args.arch, num_features=args.features,
dropout=args.dropout, num_classes=num_classes,cut_at_pooling=False, FCN=True)
start_epoch = best_top1 = 0
if args.resume:
checkpoint = load_checkpoint(args.resume)
model_dict = model.state_dict()
checkpoint_load = {k: v for k, v in (checkpoint['state_dict']).items() if k in model_dict}
model_dict.update(checkpoint_load)
model.load_state_dict(model_dict)
start_epoch = checkpoint['epoch']
best_top1 = checkpoint['best_top1']
amp.load_state_dict(checkpoint['amp'])
print("=> Start epoch {} best top1 {:.1%}"
.format(start_epoch, best_top1))
if hasattr(model, 'base'):
base_param_ids = set(map(id, model.base.parameters()))
new_params = [p for p in model.parameters() if
id(p) not in base_param_ids]
param_groups = [
{'params': model.base.parameters(), 'lr_mult': 0.1},
{'params': new_params, 'lr_mult': 1.0}]
else:
param_groups = model.parameters()
if args.npu:
from apex.optimizers import NpuFusedSGD
optimizer = NpuFusedSGD(param_groups, lr=args.lr,
momentum=args.momentum,
weight_decay=args.weight_decay,
nesterov=True)
model = model.npu()
model, optimizer = amp.initialize(model, optimizer, opt_level="O2", loss_scale=128.0, combine_grad=True)
else:
optimizer = torch.optim.SGD(param_groups, lr=args.lr,
momentum=args.momentum,
weight_decay=args.weight_decay,
nesterov=True)
model = model.cuda()
model, optimizer = amp.initialize(model, optimizer, opt_level="O2", loss_scale=128.0)
if args.device_num > 1:
model = nn.parallel.DistributedDataParallel(model,
device_ids=[args.local_rank],
output_device=args.local_rank,
find_unused_parameters=True,
broadcast_buffers=False
)
else:
model = nn.DataParallel(model)
evaluator = Evaluator(model)
if args.evaluate:
print("Test:")
checkpoint = load_checkpoint(osp.join(args.logs_dir, 'checkpoint.pth.tar'))
model.module.load_state_dict(checkpoint['state_dict'])
amp.load_state_dict(checkpoint['amp'])
evaluator.evaluate(query_loader, gallery_loader, dataset.query, dataset.gallery)
return
if args.npu:
criterion = nn.CrossEntropyLoss().npu()
else:
criterion = nn.CrossEntropyLoss().cuda()
trainer = Trainer(model, criterion, 0, 0, SMLoss_mode=0)
def adjust_lr(epoch):
step_size = 60 if args.arch == 'inception' else args.step_size
lr = args.lr * (0.1 ** (epoch // step_size))
for g in optimizer.param_groups:
g['lr'] = lr * g.get('lr_mult', 1)
total_avg = 0.0
for epoch in range(start_epoch, args.epochs):
if args.device_num > 1:
train_sampler.set_epoch(epoch)
adjust_lr(epoch)
use_time = trainer.train(epoch, train_loader, optimizer)
total_avg += use_time
is_best = True
if args.local_rank == 0:
save_checkpoint({
'state_dict': model.module.state_dict(),
'epoch': epoch + 1,
'best_top1': best_top1,
'amp': amp.state_dict(),
}, is_best, fpath=osp.join(args.logs_dir, 'checkpoint.pth.tar'))
avg_time = total_avg / (args.epochs - start_epoch)
if not args.performance:
print('Test with best model:')
evaluator.evaluate(query_loader, gallery_loader, dataset.query, dataset.gallery)
device_num = 1 if args.device_num == -1 else args.device_num
print('FPS@all {:.3f}, TIME@all {:.3f}'.format(device_num * args.batch_size / avg_time, avg_time))
if __name__ == '__main__':
parser = argparse.ArgumentParser(description="Softmax loss classification")
parser.add_argument('-d', '--dataset', type=str, default='cuhk03',
choices=datasets.names())
parser.add_argument('-b', '--batch-size', type=int, default=256)
parser.add_argument('-j', '--workers', type=int, default=4)
parser.add_argument('--split', type=int, default=0)
parser.add_argument('--height', type=int,
help="input height, default: 256 for resnet*, "
"144 for inception")
parser.add_argument('--width', type=int,
help="input width, default: 128 for resnet*, "
"56 for inception")
parser.add_argument('--combine-trainval', action='store_true',
help="train and val sets together for training, "
"val set alone for validation")
parser.add_argument('-a', '--arch', type=str, default='resnet50',
choices=models.names())
parser.add_argument('--features', type=int, default=128)
parser.add_argument('--dropout', type=float, default=0.5)
parser.add_argument('--lr', type=float, default=0.1,
help="learning rate of new parameters, for pretrained "
"parameters it is 10 times smaller than this")
parser.add_argument('--momentum', type=float, default=0.9)
parser.add_argument('--weight-decay', type=float, default=5e-4)
parser.add_argument('--resume', type=str, default='', metavar='PATH')
parser.add_argument('--evaluate', action='store_true',
help="evaluation only")
parser.add_argument('--epochs', type=int, default=50)
parser.add_argument('--step-size',type=int, default=40)
parser.add_argument('--seed', type=int, default=1)
parser.add_argument('--print-freq', type=int, default=1)
working_dir = osp.dirname(osp.abspath(__file__))
parser.add_argument('--data-dir', type=str, metavar='PATH',
default=osp.join(working_dir, 'data'))
parser.add_argument('--logs-dir', type=str, metavar='PATH',
default=osp.join(working_dir, 'logs'))
parser.add_argument("--local_rank", type=int, default=0)
parser.add_argument('--device_num', default=-1, type=int,
help='device_num')
parser.add_argument('--npu', action='store_true',
help="npu")
parser.add_argument('--addr', default='127.0.0.1',
type=str, help='master addr')
parser.add_argument('--performance', action='store_true',
help="performance")
main(parser.parse_args())