-
Notifications
You must be signed in to change notification settings - Fork 20
/
main_ddp.py
238 lines (205 loc) · 8.41 KB
/
main_ddp.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
import os
import argparse
import utils
import shutil
import wandb
import random
import numpy as np
from tqdm import tqdm
import models
import torch
import torch.nn as nn
from torchvision import transforms as T
import torch.distributed as dist
from torch.nn.parallel import DistributedDataParallel as DDP
from torch.utils.data.distributed import DistributedSampler
import torchvision.datasets as datasets
def same_seeds(seed=18):
np.random.seed(seed) # Numpy module.
random.seed(seed) # Python random module.
torch.manual_seed(seed)
if torch.cuda.is_available():
torch.cuda.manual_seed(seed)
torch.cuda.manual_seed_all(seed) # if you are using multi-GPU.
torch.backends.cudnn.benchmark = False
torch.backends.cudnn.deterministic = True
def save_checkpoint(state, is_best, filename = 'checkpoint.pth.tar'):
torch.save(state, filename)
if is_best:
shutil.copyfile(filename, 'model_best.pth.tar')
def train_epoch(epoch, net, train_loader, val_loader , criterion, optimizer, scheduler, device):
"""
Training logic for an epoch
"""
if args.local_rank == 0:
global best_acc1
same_seeds(args.seed_num)
train_loss = utils.AverageMeter("Epoch losses", ":.4e")
train_acc1 = utils.AverageMeter("Train Acc@1", ":6.2f")
train_acc5 = utils.AverageMeter("Train Acc@5", ":6.2f")
progress_train = utils.ProgressMeter(
num_batches = len(train_loader),
meters = [train_loss, train_acc1, train_acc5],
prefix = 'Epoch: {} '.format(epoch + 1),
batch_info = " Iter"
)
net.train()
# Callbacks
# warm_up_cos = lambda epoch: epoch / args.warmup_epochs if epoch <= args.warmup_epochs else 0.5 * (math.cos((epoch - args.warmup_epochs) /(args.epochs - args.warmup_epochs) * math.pi) + 1)
# scheduler_wucos = torch.optim.lr_scheduler.LambdaLR(optimizer=optimizer, lr_lambda=warm_up_cos)
for it, (inputs, targets) in enumerate(tqdm(train_loader)):
inputs = inputs.to(device)
targets = targets.to(device)
# Forward pass
outputs = net(inputs)
loss = criterion(outputs, targets)
acc1, acc5 = utils.accuracy(outputs, targets, topk = (1, 5))
train_loss.update(loss.item(), inputs.size(0))
train_acc1.update(acc1.item(), inputs.size(0))
train_acc5.update(acc5.item(), inputs.size(0))
if it % args.print_freq == 0:
progress_train.display(it)
# Backward and optimize
optimizer.zero_grad()
loss.backward()
optimizer.step()
# Log on Wandb
if args.local_rank == 0:
wandb.log({
"Loss/train" : train_loss.avg,
"Acc@1/train" : train_acc1.avg,
"Acc@5/train" : train_acc5.avg,
})
# Validation model
val_loss = utils.AverageMeter("Val losses", ":.4e")
val_acc1 = utils.AverageMeter("Val Acc@1", ":6.2f")
val_acc5 = utils.AverageMeter("Val Acc@5", ":6.2f")
progress_val = utils.ProgressMeter(
num_batches = len(val_loader),
meters = [val_loss, val_acc1, val_acc5],
prefix = 'Epoch: {} '.format(epoch + 1),
batch_info = " Iter"
)
net.eval()
for it, (inputs, targets) in enumerate(val_loader):
inputs = inputs.to(device)
targets = targets.to(device)
# Forward pass
with torch.no_grad():
outputs = net(inputs)
loss = criterion(outputs, targets)
acc1, acc5 = utils.accuracy(outputs, targets, topk=(1, 5))
val_loss.update(loss.item(), inputs.size(0))
val_acc1.update(acc1.item(), inputs.size(0))
val_acc5.update(acc5.item(), inputs.size(0))
acc1 = val_acc1.avg
if it % args.print_freq == 0:
progress_val.display(it)
# Log on Wandb
if args.local_rank == 0:
wandb.log({
"Loss/val" : val_loss.avg,
"Acc@1/val" : val_acc1.avg,
"Acc@5/val" : val_acc5.avg
})
# Learning_rate callbacks
scheduler.step()
if args.local_rank == 0:
is_best = acc1 > best_acc1
best_acc1 = max(acc1, best_acc1)
save_checkpoint({
'epoch': epoch + 1,
'state_dict': net.state_dict(),
'best_acc1': best_acc1,
'optimizer' : optimizer.state_dict(),
}, is_best)
return val_loss.avg, val_acc1.avg, val_acc5.avg
if __name__ == "__main__":
best_acc1 = 0.0
parser = argparse.ArgumentParser(description = "Train classification of CMT model")
parser.add_argument('--data', metavar = 'DIR', default = '../imagenet_data',
help = 'path to dataset')
parser.add_argument("--gpu_device", type = int, default = 2,
help = "Select specific GPU to run the model")
parser.add_argument('--batch-size', type = int, default = 64, metavar = 'N',
help = 'Input batch size for training (default: 64)')
parser.add_argument('--epochs', type = int, default = 90, metavar = 'N',
help = 'Number of epochs to train (default: 90)')
parser.add_argument('-we', '--warmup_epochs', default=5, type=int, help='epochs for warmup')
parser.add_argument('--num-class', type = int, default = 1000, metavar = 'N',
help = 'Number of classes to classify (default: 10)')
parser.add_argument('--lr', type = float, default = 0.05, metavar='LR',
help = 'Learning rate (default: 6e-5)')
parser.add_argument('--weight-decay', type = float, default = 5e-5, metavar = 'WD',
help = 'Weight decay (default: 1e-5)')
parser.add_argument('-p', '--print-freq', default = 10, type = int, metavar='N',
help='print frequency (default: 10)')
parser.add_argument('--local_rank', type=int,
help='local rank for DistributedDataParallel')
parser.add_argument('-seed', '--seed_num', default=42, type=int,
help='number of random seed')
args = parser.parse_args()
# Multi GPU
print(f'Running DDP on rank: {args.local_rank}')
torch.cuda.set_device(args.local_rank)
dist.init_process_group(backend = 'nccl', init_method = 'env://')
# Create folder to save model
WEIGHTS_PATH = "./weights"
if not os.path.exists(WEIGHTS_PATH):
os.makedirs(WEIGHTS_PATH)
# Set device
# os.environ["CUDA_VISIBLE_DEVICES"] = "2,3"
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# Data loading
traindir = os.path.join(args.data, 'train')
valdir = os.path.join(args.data, 'val')
normalize = T.Normalize(mean = [0.485, 0.456, 0.406], std = [0.229, 0.224, 0.225])
train_dataset = datasets.ImageFolder(
traindir,
T.Compose([
T.RandomResizedCrop(256),
T.RandomHorizontalFlip(),
T.ToTensor(),
normalize,
]))
train_sampler = DistributedSampler(train_dataset)
val_dataset = datasets.ImageFolder(
valdir,
T.Compose([
T.Resize(256),
T.ToTensor(),
normalize,
]))
train_loader = torch.utils.data.DataLoader(
train_dataset, batch_size = args.batch_size, shuffle = (train_sampler is None),
num_workers = 8, pin_memory = True, sampler=train_sampler
)
val_loader = torch.utils.data.DataLoader(
val_dataset, batch_size = args.batch_size, shuffle = False,
num_workers = 8, pin_memory = True
)
# Create model
net = models.MobileViT_S()
# net.to(device)
# net = torch.nn.DataParallel(net).to(device)
net = net.to(args.local_rank)
net = DDP(net, device_ids=[args.local_rank], output_device=args.local_rank)
# Set loss function and optimizer
criterion = nn.CrossEntropyLoss().to(device)
optimizer = torch.optim.SGD(net.parameters(), args.lr,
momentum = 0.9,
weight_decay = args.weight_decay)
scheduler = torch.optim.lr_scheduler.CosineAnnealingLR(optimizer, args.epochs)
# Using wandb for logging
if args.local_rank == 0:
wandb.init()
wandb.config.update(args)
wandb.watch(net)
# Train the model
for epoch in tqdm(range(args.epochs)):
train_sampler.set_epoch(epoch)
loss, acc1, acc5 = train_epoch(epoch, net, train_loader,
val_loader, criterion, optimizer, scheduler, device
)
print(f"Epoch {epoch} -> Acc@1: {acc1}, Acc@5: {acc5}")
print("Training is done")