-
Notifications
You must be signed in to change notification settings - Fork 14
/
Copy pathmain.py
160 lines (138 loc) · 6.99 KB
/
main.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
import torch
import torch.optim as optim
import torch.nn as nn
from torch.utils.data import Dataset, DataLoader, random_split
from torchvision import transforms
from torchvision.utils import make_grid
from tensorboardX import SummaryWriter
import argparse
import datetime
import os
import random
from datasets import FivekDataset
from models import CAN, SandOCAN, UNet
from torch_utils import JoinedDataLoader, load_model
from loss import ColorSSIM, NimaLoss
parser = argparse.ArgumentParser()
parser.add_argument('--batch_size', type=int, default=8, help='input batch size')
parser.add_argument('--epochs', type=int, default=52, help='number of epochs to train for')
parser.add_argument('--lr', type=float, default=2e-4, help='learning rate')
parser.add_argument('--cuda', action='store_true', help='enables cuda')
parser.add_argument('--cuda_idx', type=int, default=1, help='cuda device id')
parser.add_argument('--manual_seed', type=int, help='manual seed')
parser.add_argument('--logdir', default='log', help='logdir for tensorboard')
parser.add_argument('--run_tag', default='', help='tags for the current run')
parser.add_argument('--checkpoint_every', default=10, help='number of epochs after which saving checkpoints')
parser.add_argument('--checkpoint_dir', default="checkpoints", help='directory for the checkpoints')
parser.add_argument('--final_dir', default="final_models", help='directory for the final_models')
parser.add_argument('--model_type', default='can32', choices=['can32', 'sandocan32','unet'], help='type of model to use')
parser.add_argument('--load_model', action='store_true', help='enables load from latest checkpoint')
parser.add_argument('--loss', default='mse', choices=['mse','mae','l1nima','l2nima','l1ssim','colorssim'], help='loss to be used')
parser.add_argument('--gamma', default=0.001, type=float, help='gamma to be used only in case of Nima Loss')
parser.add_argument('--data_path', default='/home/iacv3_1/fivek', help='path of the base directory of the dataset')
opt = parser.parse_args()
#Create writer for tensorboard
date = datetime.datetime.now().strftime("%d-%m-%y_%H:%M")
run_name = "{}_{}".format(opt.run_tag,date) if opt.run_tag != '' else date
log_dir_name = os.path.join(opt.logdir, run_name)
writer = SummaryWriter(log_dir_name)
writer.add_text('Options', str(opt), 0)
print(opt)
if opt.manual_seed is None:
opt.manual_seed = random.randint(1, 10000)
print("Random Seed: ", opt.manual_seed)
random.seed(opt.manual_seed)
torch.manual_seed(opt.manual_seed)
start_epoch = 0
os.makedirs(opt.checkpoint_dir, exist_ok=True)
if torch.cuda.is_available() and not opt.cuda:
print("You should run with CUDA.")
device = torch.device("cuda:"+str(opt.cuda_idx) if opt.cuda else "cpu")
landscape_transform = transforms.Compose([
transforms.Resize((332, 500)),
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)) #normalize in [-1,1]
])
portrait_transform = transforms.Compose([
transforms.Resize((500, 332)),
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)) #normalize in [-1,1]
])
landscape_dataset = FivekDataset(opt.data_path, expert_idx=2, transform=landscape_transform, filter_ratio="landscape")
portrait_dataset = FivekDataset(opt.data_path, expert_idx=2, transform=portrait_transform, filter_ratio="portrait")
train_size = int(0.8 * len(landscape_dataset))
test_size = len(landscape_dataset) - train_size
train_landscape_dataset, test_landscape_dataset = random_split(landscape_dataset, [train_size, test_size])
train_size = int(0.8 * len(portrait_dataset))
test_size = len(portrait_dataset) - train_size
train_portrait_dataset, test_portrait_dataset = random_split(portrait_dataset, [train_size, test_size])
train_landscape_loader = DataLoader(train_landscape_dataset, batch_size=opt.batch_size, shuffle=True, num_workers=2)
train_portrait_loader = DataLoader(train_portrait_dataset, batch_size=opt.batch_size, shuffle=True, num_workers=2)
train_loader = JoinedDataLoader(train_landscape_loader, train_portrait_loader)
test_landscape_loader = DataLoader(test_landscape_dataset, batch_size=opt.batch_size, shuffle=True, num_workers=2)
test_portrait_loader = DataLoader(test_portrait_dataset, batch_size=opt.batch_size, shuffle=True, num_workers=2)
test_loader = JoinedDataLoader(test_landscape_loader, test_portrait_loader)
if opt.model_type == 'can32':
model = CAN(n_channels=32)
if opt.model_type == 'sandocan32':
model = SandOCAN()
if opt.model_type == 'unet':
model = UNet()
assert model
if opt.load_model:
model, start_epoch = load_model(model, opt.checkpoint_dir, opt.run_tag)
model = model.to(device)
if opt.loss == "mse":
criterion = nn.MSELoss()
if opt.loss == "mae":
criterion = nn.L1Loss()
if opt.loss == "l1nima":
criterion = NimaLoss(device,opt.gamma,nn.L1Loss())
if opt.loss == "l2nima":
criterion = NimaLoss(device,opt.gamma,nn.MSELoss())
if opt.loss == "l1ssim":
criterion = ColorSSIM(device,'l1')
if opt.loss == "colorssim":
criterion = ColorSSIM(device)
assert criterion
criterion = criterion.to(device)
optimizer = optim.Adam(model.parameters(), lr=opt.lr)
#Select random idxs for displaying
test_idxs = random.sample(range(len(test_landscape_dataset)), 3)
for epoch in range(start_epoch, opt.epochs):
model.train()
cumulative_loss = 0.0
for i, (im_o, im_t) in enumerate(train_loader):
im_o, im_t = im_o.to(device), im_t.to(device)
optimizer.zero_grad()
output = model(im_o)
loss = criterion(output, im_t)
loss.backward()
optimizer.step()
cumulative_loss += loss.item()
print('[Epoch %d, Batch %2d] loss: %.3f' %
(epoch + 1, i + 1, cumulative_loss / (i+1)), end="\r")
#Evaluate
writer.add_scalar('Train Error', cumulative_loss / len(train_loader), epoch)
#Checkpointing
if (epoch+1) % opt.checkpoint_every == 0:
torch.save(model.state_dict(), os.path.join(opt.checkpoint_dir, "{}_epoch{}.pt".format(opt.run_tag, epoch+1)))
#Model evaluation
model.eval()
test_loss = []
for i, (im_o, im_t) in enumerate(test_loader):
im_o, im_t = im_o.to(device), im_t.to(device)
with torch.no_grad():
output = model(im_o)
test_loss.append(criterion(output, im_t).item())
avg_loss = sum(test_loss)/len(test_loss)
writer.add_scalar('Test Error', avg_loss, epoch)
for idx in test_idxs:
original, actual = test_landscape_dataset[idx]
original, actual = original.unsqueeze(0).to(device), actual.unsqueeze(0).to(device)
estimated = model(original)
images = torch.cat((original, estimated, actual))
grid = make_grid(images, nrow=1, normalize=True, range=(-1,1))
writer.add_image('{}:Original|Estimated|Actual'.format(idx), grid, epoch)
print("Training completed succesfully")
torch.save(model.state_dict(), os.path.join(opt.final_dir, "{}_final.pt".format(opt.run_tag)))