import argparse
import os
import random
import shutil
import time
import warnings
import moxing as mox
import apex
import numpy as np
import torch.npu
from apex import amp
from collections import OrderedDict
import torch
import torch.onnx
import torch.nn as nn
import torch.nn.parallel
import torch.backends.cudnn as cudnn
import torch.distributed as dist
import torch.optim
import torch.multiprocessing as mp
import torch.utils.data
import torch.utils.data.distributed
import torchvision.transforms as transforms
import torchvision.datasets as datasets
import senet
CALCULATE_DEVICE = "npu:0"
model_names = ['ResNet', 'resnet18', 'resnet34', 'resnet50', 'resnet101',
'resnet152', 'resnext50_32x4d', 'resnext101_32x8d',
'wide_resnet50_2', 'wide_resnet101_2']
parser = argparse.ArgumentParser(description='PyTorch ImageNet Training')
parser.add_argument('--data', default='', type=str,
help='path to dataset')
parser.add_argument('-a', '--arch', metavar='ARCH', default='resnet18',
choices=model_names,
help='model architecture: ' +
' | '.join(model_names) +
' (default: resnet18)')
parser.add_argument('-j', '--workers', default=4, type=int, metavar='N',
help='number of data loading workers (default: 4)')
parser.add_argument('--epochs', default=90, type=int, metavar='N',
help='number of total epochs to run')
parser.add_argument('--start-epoch', default=0, type=int, metavar='N',
help='manual epoch number (useful on restarts)')
parser.add_argument('-b', '--batch-size', default=256, type=int,
metavar='N',
help='mini-batch size (default: 256), this is the total '
'batch size of all GPUs on the current node when '
'using Data Parallel or Distributed Data Parallel')
parser.add_argument('--lr', '--learning-rate', default=0.1, type=float,
metavar='LR', help='initial learning rate', dest='lr')
parser.add_argument('--momentum', default=0.9, type=float, metavar='M',
help='momentum')
parser.add_argument('--wd', '--weight-decay', default=1e-4, type=float,
metavar='W', help='weight decay (default: 1e-4)',
dest='weight_decay')
parser.add_argument('-p', '--print-freq', default=10, type=int,
metavar='N', help='print frequency (default: 10)')
parser.add_argument('--resume', default='', type=str, metavar='PATH',
help='path to latest checkpoint (default: none)')
parser.add_argument('-e', '--evaluate', dest='evaluate', action='store_true',
help='evaluate model on validation set')
parser.add_argument('--pretrained', default=False, dest='pretrained', action='store_true',
help='use pre-trained model')
parser.add_argument('--world-size', default=-1, type=int,
help='number of nodes for distributed training')
parser.add_argument('--rank', default=-1, type=int,
help='node rank for distributed training')
parser.add_argument('--dist-url', default='tcp://224.66.41.62:23456', type=str,
help='url used to set up distributed training')
parser.add_argument('--dist-backend', default='nccl', type=str,
help='distributed backend')
parser.add_argument('--seed', default=None, type=int,
help='seed for initializing training. ')
parser.add_argument('--gpu', default=None, type=int,
help='GPU id to use.')
parser.add_argument('--npu', default=None, type=int,
help='NPU id to use.')
parser.add_argument('--multiprocessing-distributed', action='store_true',
help='Use multi-processing distributed training to launch '
'N processes per node, which has N GPUs. This is the '
'fastest way to use PyTorch for either single node or '
'multi node data parallel training')
parser.add_argument('--device', default='npu', type=str, help='npu or gpu')
parser.add_argument('--addr', default='10.136.181.115',
type=str, help='master addr')
parser.add_argument('--amp', default=False, action='store_true',
help='use amp to train the model')
parser.add_argument('--warm_up_epochs', default=0, type=int,
help='warm up')
parser.add_argument('--loss-scale', default=1024., type=float,
help='loss scale using in amp, default -1 means dynamic')
parser.add_argument('--opt-level', default='O2', type=str,
help='loss scale using in amp, default -1 means dynamic')
parser.add_argument('--prof', default=False, action='store_true',
help='use profiling to evaluate the performance of model')
parser.add_argument('--save_path', default='', type=str,
help='path to save models')
parser.add_argument('--num_classes', default=1000, type=int,
help='path to save models')
parser.add_argument('--train_url',
default='obs://mindx-user/csl-shi/wideresnet101/result/',
type=str,
help="setting dir of training output")
parser.add_argument('--data_url',
default='obs://mindx-user/csl-shi/wideresnet101/data/',
type=str,
help='path to dataset')
parser.add_argument('--model_url',
metavar='DIR',
default='',
help='path to pretrained model')
parser.add_argument('--onnx', default=True, action='store_true',
help="convert pth model to onnx")
cur_step = 0
CACHE_TRAINING_URL = "/cache/training/"
CACHE_DATA_URL = "/cache/data_url"
CACHE_MODEL_URL = "/cache/model"
best_acc1 = 0
def main():
args = parser.parse_args()
global CALCULATE_DEVICE
CALCULATE_DEVICE = "npu:{}".format(args.npu)
if 'npu' in CALCULATE_DEVICE:
torch.npu.set_device(CALCULATE_DEVICE)
if args.data_url:
import moxing as mox
if args.seed is not None:
random.seed(args.seed)
torch.manual_seed(args.seed)
cudnn.deterministic = True
warnings.warn('You have chosen to seed training. '
'This will turn on the CUDNN deterministic setting, '
'which can slow down your training considerably! '
'You may see unexpected behavior when restarting '
'from checkpoints.')
if args.gpu is not None:
warnings.warn('You have chosen a specific GPU. This will completely '
'disable data parallelism.')
if args.dist_url == "env://" and args.world_size == -1:
args.world_size = int(os.environ["WORLD_SIZE"])
args.distributed = args.world_size > 1 or args.multiprocessing_distributed
ngpus_per_node = torch.cuda.device_count()
if args.multiprocessing_distributed:
args.world_size = ngpus_per_node * args.world_size
mp.spawn(main_worker, nprocs=ngpus_per_node, args=(ngpus_per_node, args))
else:
main_worker(args.gpu, ngpus_per_node, args)
def main_worker(gpu, ngpus_per_node, args):
global best_acc1
args.gpu = None
if args.gpu is not None:
print("Use GPU: {} for training".format(args.gpu))
if args.distributed:
if args.dist_url == "env://" and args.rank == -1:
args.rank = int(os.environ["RANK"])
if args.multiprocessing_distributed:
args.rank = args.rank * ngpus_per_node + gpu
if args.device == 'npu':
dist.init_process_group(backend=args.dist_backend,
world_size=args.world_size, rank=args.rank)
else:
dist.init_process_group(backend=args.dist_backend, init_method=args.dist_url,
world_size=args.world_size, rank=args.rank)
if args.pretrained:
print("=> using pre-trained model wide_resnet101_2")
model = senet.se_resnet50()
print("loading model of yours...")
model_path = "./checkpoint.pth.tar"
if args.model_url:
real_path = CACHE_MODEL_URL
if not os.path.exists(real_path):
os.makedirs(real_path)
mox.file.copy_parallel(args.model_url, real_path)
print("training data finish copy to %s." % real_path)
model_path = os.path.join(CACHE_MODEL_URL, 'checkpoint.pth.tar')
pretrained_dict = torch.load(model_path, map_location="cpu")["state_dict"]
model.load_state_dict({k.replace('module.', ''): v for k, v in pretrained_dict.items()})
if "fc.weight" in pretrained_dict:
pretrained_dict.pop('fc.weight')
pretrained_dict.pop('fc.bias')
for param in model.parameters():
param.requires_grad = False
model.fc = nn.Linear(2048, args.num_classes)
else:
print("=> creating model wide_resnet101_2")
model = senet.se_resnet50()
if args.distributed:
if args.gpu is not None:
torch.cuda.set_device(args.gpu)
model.cuda(args.gpu)
args.batch_size = int(args.batch_size / ngpus_per_node)
args.workers = int((args.workers + ngpus_per_node - 1) / ngpus_per_node)
model = torch.nn.parallel.DistributedDataParallel(model, device_ids=[args.gpu])
else:
model.cuda()
model = torch.nn.parallel.DistributedDataParallel(model)
elif args.gpu is not None:
torch.cuda.set_device(args.gpu)
model = model.cuda(args.gpu)
else:
if args.arch.startswith('alexnet') or args.arch.startswith('vgg'):
model.features = torch.nn.DataParallel(model.features)
model.cuda()
else:
model = model.to(CALCULATE_DEVICE)
criterion = nn.CrossEntropyLoss().to(CALCULATE_DEVICE)
optimizer = apex.optimizers.NpuFusedSGD(model.parameters(), args.lr,
momentum=args.momentum,
nesterov=True,
weight_decay=args.weight_decay)
if args.amp:
model, optimizer = amp.initialize(
model, optimizer, opt_level=args.opt_level, loss_scale=args.loss_scale)
if args.resume:
if os.path.isfile(args.resume):
print("=> loading checkpoint '{}'".format(args.resume))
if args.gpu is None:
checkpoint = torch.load(args.resume)
else:
loc = 'cuda:{}'.format(args.gpu)
checkpoint = torch.load(args.resume, map_location=loc)
args.start_epoch = checkpoint['epoch']
best_acc1 = checkpoint['best_acc1']
if args.gpu is not None:
best_acc1 = best_acc1.to(args.gpu)
model.load_state_dict(checkpoint['state_dict'])
optimizer.load_state_dict(checkpoint['optimizer'])
print("=> loaded checkpoint '{}' (epoch {})"
.format(args.resume, checkpoint['epoch']))
else:
print("=> no checkpoint found at '{}'".format(args.resume))
cudnn.benchmark = True
if args.data_url:
real_path = CACHE_DATA_URL
if not os.path.exists(real_path):
os.makedirs(real_path)
mox.file.copy_parallel(args.data_url, real_path)
print("training data finish copy to %s." % real_path)
args.data = real_path
traindir = os.path.join(args.data, 'train')
valdir = os.path.join(args.data, 'val')
normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
train_dataset = datasets.ImageFolder(
traindir,
transforms.Compose([
transforms.RandomResizedCrop(224),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
normalize,
]))
if args.distributed:
train_sampler = torch.utils.data.distributed.DistributedSampler(train_dataset)
else:
train_sampler = None
train_loader = torch.utils.data.DataLoader(
train_dataset, batch_size=args.batch_size, shuffle=(
train_sampler is None),
num_workers=args.workers, pin_memory=False, sampler=train_sampler, drop_last=True)
val_loader = torch.utils.data.DataLoader(
datasets.ImageFolder(valdir, transforms.Compose([
transforms.Resize(256),
transforms.CenterCrop(224),
transforms.ToTensor(),
normalize,
])),
batch_size=args.batch_size, shuffle=False,
num_workers=args.workers, pin_memory=True)
if args.evaluate:
validate(val_loader, model, criterion, args)
return
if args.prof:
profiling(train_loader, model, criterion, optimizer, args)
return
for epoch in range(args.start_epoch, args.epochs):
if args.distributed:
train_sampler.set_epoch(epoch)
adjust_learning_rate(optimizer, epoch, args)
train(train_loader, model, criterion, optimizer, epoch, args, ngpus_per_node)
acc1 = validate(val_loader, model, criterion, args)
is_best = acc1 > best_acc1
best_acc1 = max(acc1, best_acc1)
if not args.multiprocessing_distributed or (args.multiprocessing_distributed
and args.rank % ngpus_per_node == 0):
save_checkpoint({
'epoch': epoch + 1,
'arch': args.arch,
'state_dict': model.state_dict(),
'best_acc1': best_acc1,
'optimizer': optimizer.state_dict(),
}, is_best)
if args.train_url:
mox.file.copy_parallel(CACHE_TRAINING_URL, args.train_url)
def proc_node_module(checkpoint, AttrName):
new_state_dict = OrderedDict()
for k, v in checkpoint[AttrName].items():
if(k[0:7] == "module."):
name = k[7:]
else:
name = k[0:]
new_state_dict[name] = v
return new_state_dict
def convert(model_path, onnx_save, num_class):
checkpoint = torch.load(model_path, map_location='cpu')
checkpoint['state_dict'] = proc_node_module(checkpoint, 'state_dict')
model = senet.se_resnet50()
model.load_state_dict(checkpoint['state_dict'])
model.eval()
input_names = ["actual_input_1"]
output_names = ["output1"]
dummy_input = torch.randn(1, 3, 224, 224)
if len(onnx_save) > 0:
save_path = os.path.join(onnx_save, "se_resnet50_2_npu_16.onnx")
else:
save_path = "se_resnet50_2_npu_16.onnx"
print(save_path)
torch.onnx.export(model, dummy_input, save_path
, input_names=input_names, output_names=output_names
, opset_version=11)
def profiling(data_loader, model, criterion, optimizer, args):
model.train()
def update(model, images, target, optimizer):
output = model(images)
loss = criterion(output, target)
if args.amp:
with amp.scale_loss(loss, optimizer) as scaled_loss:
scaled_loss.backward()
else:
loss.backward()
optimizer.zero_grad()
optimizer.step()
for step, (images, target) in enumerate(data_loader):
if args.device == 'npu':
loc = CALCULATE_DEVICE
images = images.to(loc, non_blocking=True).to(torch.float)
target = target.to(torch.int32).to(loc, non_blocking=True)
else:
images = images.cuda(args.gpu, non_blocking=True)
target = target.cuda(args.gpu, non_blocking=True)
if step < 5:
update(model, images, target, optimizer)
else:
if args.device == 'npu':
with torch.autograd.profiler.profile(use_npu=True) as prof:
update(model, images, target, optimizer)
else:
with torch.autograd.profiler.profile(use_cuda=True) as prof:
update(model, images, target, optimizer)
break
prof.export_chrome_trace("output.prof")
def train(train_loader, model, criterion, optimizer, epoch, args, ngpus_per_node):
batch_time = AverageMeter('Time', ':6.3f')
data_time = AverageMeter('Data', ':6.3f')
losses = AverageMeter('Loss', ':.4e')
top1 = AverageMeter('Acc@1', ':6.2f')
top5 = AverageMeter('Acc@5', ':6.2f')
progress = ProgressMeter(
len(train_loader),
[batch_time, data_time, losses, top1, top5],
prefix="Epoch: [{}]".format(epoch))
model.train()
end = time.time()
for i, (images, target) in enumerate(train_loader):
data_time.update(time.time() - end)
if args.gpu is not None:
images = images.cuda(args.gpu, non_blocking=True)
if 'npu' in CALCULATE_DEVICE:
target = target.to(torch.int32)
images, target = images.to(CALCULATE_DEVICE, non_blocking=True), target.to(CALCULATE_DEVICE, non_blocking=True)
output = model(images)
loss = criterion(output, target)
acc1, acc5 = accuracy(output, target, topk=(1, 5))
losses.update(loss.item(), images.size(0))
top1.update(acc1[0], images.size(0))
top5.update(acc5[0], images.size(0))
optimizer.zero_grad()
if args.amp:
with amp.scale_loss(loss, optimizer) as scaled_loss:
scaled_loss.backward()
else:
loss.backward()
optimizer.step()
batch_time.update(time.time() - end)
end = time.time()
if i % args.print_freq == 0:
if not args.multiprocessing_distributed or (args.multiprocessing_distributed
and args.rank % ngpus_per_node == 0):
progress.display(i)
if not args.multiprocessing_distributed or (args.multiprocessing_distributed
and args.rank % ngpus_per_node == 0):
if batch_time.avg:
print("[npu id:", CALCULATE_DEVICE, "]", "batch_size:", args.world_size * args.batch_size,
'Time: {:.3f}'.format(batch_time.avg), '* FPS@all {:.3f}'.format(
args.batch_size * args.world_size / batch_time.avg))
def validate(val_loader, model, criterion, args):
batch_time = AverageMeter('Time', ':6.3f', start_count_index= 5)
losses = AverageMeter('Loss', ':.4e')
top1 = AverageMeter('Acc@1', ':6.2f')
top5 = AverageMeter('Acc@5', ':6.2f')
progress = ProgressMeter(
len(val_loader),
[batch_time, losses, top1, top5],
prefix='Test: ')
model.eval()
with torch.no_grad():
end = time.time()
for i, (images, target) in enumerate(val_loader):
if args.device == 'npu':
loc = CALCULATE_DEVICE
images = images.to(loc).to(torch.float)
if args.device == 'npu':
loc = CALCULATE_DEVICE
target = target.to(torch.int32).to(loc, non_blocking=True)
output = model(images)
loss = criterion(output, target)
acc1, acc5 = accuracy(output, target, topk=(1, 5))
losses.update(loss.item(), images.size(0))
top1.update(acc1[0], images.size(0))
top5.update(acc5[0], images.size(0))
batch_time.update(time.time() - end)
end = time.time()
if i % args.print_freq == 0:
progress.display(i)
print(' * Acc@1 {top1.avg:.3f} Acc@5 {top5.avg:.3f}'
.format(top1=top1, top5=top5))
return top1.avg
def save_checkpoint(state, is_best, filename='checkpoint.pth.tar'):
args = parser.parse_args()
if args.train_url:
os.makedirs(CACHE_TRAINING_URL, 0o755, exist_ok=True)
filename = os.path.join(CACHE_TRAINING_URL, filename)
torch.save(state, filename)
convert(filename, CACHE_TRAINING_URL, args.num_classes)
path_best = os.path.join(CACHE_TRAINING_URL, 'model_best.pth.tar')
if is_best:
shutil.copyfile(filename, path_best)
else:
filename = os.path.join(args.save_path, filename)
torch.save(state, filename)
path_best = os.path.join(args.save_path, 'model_best.pth.tar')
if is_best:
shutil.copyfile(filename, path_best)
class AverageMeter(object):
"""Computes and stores the average and current value"""
def __init__(self, name, fmt=':f', start_count_index=2):
self.name = name
self.fmt = fmt
self.reset()
self.start_count_index = start_count_index
def reset(self):
self.val = 0
self.avg = 0
self.sum = 0
self.count = 0
def update(self, val, n=1):
if self.count == 0:
self.N = n
self.val = val
self.count += n
if self.count > (self.start_count_index * self.N):
self.sum += val * n
self.avg = self.sum / (self.count - self.start_count_index * self.N)
def __str__(self):
fmtstr = '{name} {val' + self.fmt + '} ({avg' + self.fmt + '})'
return fmtstr.format(**self.__dict__)
class ProgressMeter(object):
def __init__(self, num_batches, meters, prefix=""):
self.batch_fmtstr = self._get_batch_fmtstr(num_batches)
self.meters = meters
self.prefix = prefix
def display(self, batch):
entries = [self.prefix + self.batch_fmtstr.format(batch)]
entries += [str(meter) for meter in self.meters]
print('\t'.join(entries))
def _get_batch_fmtstr(self, num_batches):
num_digits = len(str(num_batches // 1))
fmt = '{:' + str(num_digits) + 'd}'
return '[' + fmt + '/' + fmt.format(num_batches) + ']'
def adjust_learning_rate(optimizer, epoch, args):
"""Sets the learning rate to the initial LR decayed by cosine method"""
if args.warm_up_epochs > 0 and epoch < args.warm_up_epochs:
lr = args.lr * ((epoch + 1) / (args.warm_up_epochs + 1))
else:
alpha = 0
cosine_decay = 0.5 * (
1 + np.cos(np.pi * (epoch - args.warm_up_epochs) / (args.epochs - args.warm_up_epochs)))
decayed = (1 - alpha) * cosine_decay + alpha
lr = args.lr * decayed
print("=> Epoch[%d] Setting lr: %.4f" % (epoch, lr))
for param_group in optimizer.param_groups:
param_group['lr'] = lr
def accuracy(output, target, topk=(1,)):
"""Computes the accuracy over the k top predictions for the specified values of k"""
with torch.no_grad():
maxk = max(topk)
batch_size = target.size(0)
_, pred = output.topk(maxk, 1, True, True)
pred = pred.t()
correct = pred.eq(target.view(1, -1).expand_as(pred))
res = []
for k in topk:
correct_k = correct[:k].reshape(-1).float().sum(0, keepdim=True)
res.append(correct_k.mul_(100.0 / batch_size))
return res
if __name__ == '__main__':
main()