-
Notifications
You must be signed in to change notification settings - Fork 2
/
train_eval_gc_l2_no_contrast_baseline.py
executable file
·522 lines (435 loc) · 19.6 KB
/
train_eval_gc_l2_no_contrast_baseline.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
## Code to train and evaluate a ResNet model with our CGC method on ImageNet/CUB-200/Cars-196/FGVC-Aircraft/VGG Flowers-102 datasets.
## This code is adapated from https://github.com/pytorch/examples/blob/master/imagenet/main.py
import argparse
import os
import random
import shutil
import time
import warnings
import torch
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
from datasets.imagefolder_cgc_ssl import ImageFolder
import models.resnet_multigpu_cgc as resnet
import logging
def get_logger(logpath, filepath, package_files=[], displaying=True, saving=True, debug=False):
logger = logging.getLogger()
if debug:
level = logging.DEBUG
else:
level = logging.INFO
logger.setLevel(level)
if saving:
info_file_handler = logging.FileHandler(logpath, mode="a")
info_file_handler.setLevel(level)
logger.addHandler(info_file_handler)
if displaying:
console_handler = logging.StreamHandler()
console_handler.setLevel(level)
logger.addHandler(console_handler)
logger.info(filepath)
with open(filepath, "r") as f:
logger.info(f.read())
for f in package_files:
logger.info(f)
with open(f, "r") as package_f:
logger.info(package_f.read())
return logger
model_names = ['resnet18' , 'resnet50']
parser = argparse.ArgumentParser(description='PyTorch ImageNet Training')
parser.add_argument('data', metavar='DIR',
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', 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('--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('--save_dir', default='checkpoint', type=str, metavar='SV_PATH',
help='path to save checkpoints (default: none)')
parser.add_argument('--log_dir', default='logs', type=str, metavar='LG_PATH',
help='path to write logs (default: logs)')
parser.add_argument('--dataset', type=str, default='imagenet',
help='dataset to use: [imagenet, tiny_imagenet]')
parser.add_argument('-t', type=float, default=0.01)
parser.add_argument('--lambda', default=1000, type=float,
metavar='LAM', help='lambda hyperparameter for GCAM loss', dest='lambda_val')
best_acc1 = 0
def main():
args = parser.parse_args()
os.makedirs(args.save_dir, exist_ok=True)
os.makedirs(args.log_dir, exist_ok=True)
logger = get_logger(logpath=os.path.join(args.log_dir, 'logs'), filepath=os.path.abspath(__file__))
logger.info(args)
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:
# Since we have ngpus_per_node processes per node, the total world_size
# needs to be adjusted accordingly
args.world_size = ngpus_per_node * args.world_size
# Use torch.multiprocessing.spawn to launch distributed processes: the
# main_worker process function
mp.spawn(main_worker, nprocs=ngpus_per_node, args=(ngpus_per_node, args))
else:
# Simply call main_worker function
main_worker(args.gpu, ngpus_per_node, args, logger)
class NCESoftmaxLoss(nn.Module):
"""Softmax cross-entropy loss (a.k.a., info-NCE loss in CPC paper)"""
def __init__(self , T=0.01):
super(NCESoftmaxLoss, self).__init__()
self.T = T
self.criterion = nn.CrossEntropyLoss()
def forward(self, x):
bsz = x.shape[0]
x = x.squeeze()
x = torch.div(x, self.T)
label = torch.zeros([bsz]).cuda().long()
loss = self.criterion(x, label)
return loss
def main_worker(gpu, ngpus_per_node, args, logger):
global best_acc1
args.gpu = gpu
if args.gpu is not None:
logger.info("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:
# For multiprocessing distributed training, rank needs to be the
# global rank among all the processes
args.rank = args.rank * ngpus_per_node + gpu
dist.init_process_group(backend=args.dist_backend, init_method=args.dist_url,
world_size=args.world_size, rank=args.rank)
kwargs = {}
num_classes = 1000
val_dir_name = 'val'
if args.dataset == 'tiny_imagenet':
kwargs = {'num_classes': 200}
num_classes = 200
elif args.dataset == 'cub':
kwargs = {'num_classes': 200}
num_classes = 200
val_dir_name = 'test'
elif args.dataset == 'dogs':
kwargs = {'num_classes': 120}
num_classes = 120
elif args.dataset == 'fgvc':
kwargs = {'num_classes': 100}
num_classes = 100
elif args.dataset == 'flowers':
kwargs = {'num_classes': 102}
num_classes = 102
elif args.dataset == 'cars':
kwargs = {'num_classes': 196}
num_classes = 196
elif args.dataset == 'in9l':
kwargs = {'num_classes': 9}
num_classes = 9
# create model
if args.pretrained:
if args.arch == 'resnet18':
logger.info("=> using pre-trained model 'resnet18'")
model = resnet.resnet18(pretrained=True)
elif args.arch == 'resnet50':
logger.info("=> using pre-trained model 'resnet50'")
model = resnet.resnet50(pretrained=True)
else:
print('Arch not supported!!')
exit()
else:
if args.arch == 'resnet18' :
logger.info("=> creating model 'resnet18'")
model = resnet.resnet18()
elif args.arch == 'resnet50':
logger.info("=> creating model 'resnet50'")
model = resnet.resnet50()
else:
print('Arch not supported!!')
exit()
model = torch.nn.DataParallel(model)
# optionally resume from a checkpoint
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:
# Map model to be loaded to specified single gpu.
loc = 'cuda:{}'.format(args.gpu)
checkpoint = torch.load(args.resume, map_location=loc)
best_acc1 = checkpoint['best_acc1']
if args.gpu is not None:
# best_acc1 may be from a checkpoint from a different GPU
best_acc1 = best_acc1.to(args.gpu)
model.load_state_dict(checkpoint['state_dict'])
print("=> loaded checkpoint '{}' (epoch {})"
.format(args.resume, checkpoint['epoch']))
else:
print("=> no checkpoint found at '{}'".format(args.resume))
if args.arch == 'resnet18':
model.module.fc = nn.Linear(512, num_classes)
elif args.arch == 'resnet50':
model.module.fc = nn.Linear(2048, num_classes)
model = model.cuda()
logger.info(model)
# define loss function (criterion) and optimizer
xent_criterion = nn.CrossEntropyLoss().cuda(args.gpu)
contrastive_criterion = NCESoftmaxLoss(args.t).cuda(args.gpu)
cudnn.benchmark = True
optimizer = torch.optim.SGD(model.parameters(), args.lr,
momentum=args.momentum,
weight_decay=args.weight_decay)
# Data loading code
traindir = os.path.join(args.data, 'train')
valdir = os.path.join(args.data, val_dir_name)
normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
train_dataset = ImageFolder(traindir) # transforms are handled within the implementation
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=True, sampler=train_sampler)
val_batch_size = args.batch_size
val_loader = torch.utils.data.DataLoader(
datasets.ImageFolder(valdir, transforms.Compose([
transforms.Resize(256),
transforms.CenterCrop(224),
transforms.ToTensor(),
normalize,
])),
batch_size=val_batch_size, shuffle=False,
num_workers=args.workers, pin_memory=True)
if args.evaluate:
validate(val_loader, model, contrastive_criterion, xent_criterion, args, logger)
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 for one epoch
train(train_loader, model, contrastive_criterion , xent_criterion, optimizer, epoch, args, logger)
# evaluate on validation set
acc1 = validate(val_loader, model, contrastive_criterion, xent_criterion, args, logger)
# remember best acc@1 and save checkpoint
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, args.save_dir)
def train(train_loader, model, contrastive_criterion, xent_criterion, optimizer, epoch, args, logger):
batch_time = AverageMeter('Time', ':6.3f')
data_time = AverageMeter('Data', ':6.3f')
losses = AverageMeter('Loss', ':.4e')
xe_losses = AverageMeter('XE Loss', ':.4e')
gc_losses = AverageMeter('GC L2 Loss', ':.4e')
top1 = AverageMeter('Acc@1', ':6.2f')
top5 = AverageMeter('Acc@5', ':6.2f')
progress = ProgressMeter(
len(train_loader),
[batch_time, data_time, xe_losses, gc_losses, losses, top1, top5],
logger,
prefix="Epoch: [{}]".format(epoch))
# switch to train mode
model.train()
train_len = len(train_loader)
train_iter = iter(train_loader)
end = time.time()
for i in range(train_len):
xe_images , images, aug_images, transforms_i, transforms_j, transforms_h, transforms_w, hor_flip, targets = train_iter.__next__()
# measure data loading time
data_time.update(time.time() - end)
images = images.cuda(args.gpu, non_blocking=True)
aug_images = aug_images.cuda(args.gpu, non_blocking=True)
targets = targets.cuda(args.gpu, non_blocking=True)
aug_params_dict = {'transforms_i': transforms_i, 'transforms_j': transforms_j, 'transforms_h': transforms_h,
'transforms_w': transforms_w, 'hor_flip': hor_flip}
aug_output, xe_loss, sim_loss = model(images, contrastive_criterion, xe_images=xe_images, aug_images=aug_images,
aug_params_dict=aug_params_dict, targets=targets, xent_criterion=xent_criterion, disable_contrast=True)
xe_loss = xe_loss.mean()
sim_loss = sim_loss.mean()
loss = xe_loss + args.lambda_val * sim_loss
# measure accuracy and record loss
acc1, acc5 = accuracy(aug_output, targets, topk=(1, 5))
losses.update(loss.item(), images.size(0))
xe_losses.update(xe_loss.item(), images.size(0))
gc_losses.update(sim_loss.item(), images.size(0))
top1.update(acc1[0], images.size(0))
top5.update(acc5[0], images.size(0))
# compute gradient and do SGD step
optimizer.zero_grad()
loss.backward()
optimizer.step()
# measure elapsed time
batch_time.update(time.time() - end)
end = time.time()
if i % args.print_freq == 0:
progress.display(i)
def validate(val_loader, model, contrastive_criterion, criterion, args, logger):
batch_time = AverageMeter('Time', ':6.3f')
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],
logger,
prefix='Test: ')
# switch to evaluate mode
model.eval()
with torch.no_grad():
end = time.time()
for i, (images, targets) in enumerate(val_loader):
if args.gpu is not None:
images = images.cuda(args.gpu, non_blocking=True)
targets = targets.cuda(args.gpu, non_blocking=True)
# compute output
output = model(images, contrastive_criterion, vanilla=True)
loss = criterion(output, targets)
# measure accuracy and record loss
acc1, acc5 = accuracy(output, targets, topk=(1, 5))
losses.update(loss.item(), images.size(0))
top1.update(acc1[0], images.size(0))
top5.update(acc5[0], images.size(0))
# measure elapsed time
batch_time.update(time.time() - end)
end = time.time()
if i % args.print_freq == 0:
progress.display(i)
logger.info(' * Acc@1 {top1.avg:.3f} Acc@5 {top5.avg:.3f}'
.format(top1=top1, top5=top5))
return top1.avg
def save_checkpoint(state, is_best, save_dir):
epoch = state['epoch']
filename = 'checkpoint_' + str(epoch).zfill(3) + '.pth.tar'
save_path = os.path.join(save_dir, filename)
torch.save(state, save_path)
if is_best:
best_filename = 'model_best.pth.tar'
best_save_path = os.path.join(save_dir, best_filename)
shutil.copyfile(save_path, best_save_path)
class AverageMeter(object):
"""Computes and stores the average and current value"""
def __init__(self, name, fmt=':f'):
self.name = name
self.fmt = fmt
self.reset()
def reset(self):
self.val = 0
self.avg = 0
self.sum = 0
self.count = 0
def update(self, val, n=1):
self.val = val
self.sum += val * n
self.count += n
self.avg = self.sum / self.count
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,logger, prefix="" ):
self.batch_fmtstr = self._get_batch_fmtstr(num_batches)
self.meters = meters
self.prefix = prefix
self.logger = logger
def display(self, batch):
entries = [self.prefix + self.batch_fmtstr.format(batch)]
entries += [str(meter) for meter in self.meters]
self.logger.info('\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 10 every 30 epochs"""
lr = args.lr * (0.1 ** (epoch // 30))
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()