-
Notifications
You must be signed in to change notification settings - Fork 2
/
refine.py
163 lines (133 loc) · 5.72 KB
/
refine.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
import torch
import torch.optim as optim
from tensorboardX import SummaryWriter
import numpy as np
import os
import argparse
import time, datetime
import matplotlib; matplotlib.use('Agg')
from src import config, data
from src.checkpoints import CheckpointIO
from collections import defaultdict
import shutil
import torch.nn as nn ###
# Arguments
parser = argparse.ArgumentParser(
description='Train a 3D reconstruction model.'
)
parser.add_argument('config', type=str, help='Path to config file.')
parser.add_argument('--no-cuda', action='store_true', help='Do not use cuda.')
parser.add_argument('--exit-after', type=int, default=-1,
help='Checkpoint and exit after specified number of seconds'
'with exit code 2.')
args = parser.parse_args()
cfg = config.load_config(args.config, 'configs/default.yaml')
is_cuda = (torch.cuda.is_available() and not args.no_cuda)
device = torch.device("cuda" if is_cuda else "cpu") # zw
# Set t0
t0 = time.time()
# Shorthands
out_dir = cfg['training']['out_dir']
batch_size = cfg['training']['batch_size']
backup_every = cfg['training']['backup_every']
vis_n_outputs = cfg['generation']['vis_n_outputs']
exit_after = args.exit_after
model_selection_metric = cfg['training']['model_selection_metric']
if cfg['training']['model_selection_mode'] == 'maximize':
model_selection_sign = 1
elif cfg['training']['model_selection_mode'] == 'minimize':
model_selection_sign = -1
else:
raise ValueError('model_selection_mode must be '
'either maximize or minimize.')
# Output directory
if not os.path.exists(out_dir):
os.makedirs(out_dir)
shutil.copyfile(args.config, os.path.join(out_dir, 'config.yaml'))
# Dataset
train_dataset = config.get_dataset('train', cfg, return_idx=True)
train_boundary_dataset = config.get_dataset('train_boundary', cfg, return_idx=True)
val_dataset = config.get_dataset('val', cfg, return_idx=True)
train_loader = torch.utils.data.DataLoader(
train_dataset, batch_size=batch_size, num_workers=cfg['training']['n_workers'], shuffle=True,
collate_fn=data.collate_remove_none,
worker_init_fn=data.worker_init_fn)
train_boundary_loader = torch.utils.data.DataLoader(
train_boundary_dataset, batch_size=batch_size, num_workers=cfg['training']['n_workers'], shuffle=True,
collate_fn=data.collate_remove_none,
worker_init_fn=data.worker_init_fn)
val_loader = torch.utils.data.DataLoader(
val_dataset, batch_size=1, num_workers=cfg['training']['n_workers_val'], shuffle=False,
collate_fn=data.collate_remove_none,
worker_init_fn=data.worker_init_fn)
# For visualizations
vis_loader = torch.utils.data.DataLoader(
val_dataset, batch_size=1, shuffle=False,
collate_fn=data.collate_remove_none,
worker_init_fn=data.worker_init_fn)
model_counter = defaultdict(int)
data_vis_list = []
# Model
model = config.get_model(cfg, device=device, dataset=train_dataset)
### use 2 gpu
model.to(device)
model = nn.DataParallel(model, device_ids=[0, 1])
print('output path: ', cfg['training']['out_dir'])
# Generator
generator = config.get_generator(model, cfg, device=device)
learning_rate = 1e-6 # 5e-5
# Intialize training
optimizer = optim.Adam(model.parameters(), lr=learning_rate)
scheduler = optim.lr_scheduler.CosineAnnealingLR(optimizer, T_max=1500, eta_min=learning_rate * 0.01)
trainer = config.get_trainer(model, optimizer, scheduler, cfg, device=device)
checkpoint_io = CheckpointIO(out_dir, model=model, optimizer=optimizer, scheduler=scheduler)
try:
load_dict = checkpoint_io.load('model.pt')
except FileExistsError:
load_dict = dict()
epoch_it = load_dict.get('epoch_it', 0)
it = load_dict.get('it', 0)
metric_val_best = load_dict.get(
'loss_val_best', -model_selection_sign * np.inf)
if metric_val_best == np.inf or metric_val_best == -np.inf:
metric_val_best = -model_selection_sign * np.inf
print('Current best validation metric (%s): %.8f'
% (model_selection_metric, metric_val_best))
logger = SummaryWriter(os.path.join(out_dir, 'logs'))
# Shorthands
print_every = cfg['training']['print_every']
checkpoint_every = cfg['training']['checkpoint_every']
validate_every = cfg['training']['validate_every']
visualize_every = cfg['training']['visualize_every']
stage_1_epoch = cfg['training']['stage_1_epoch']
stage_2_epoch = cfg['training']['stage_2_epoch']
refine_epoch = 0
while refine_epoch <= stage_2_epoch:
refine_epoch += 1
epoch_it += 1
for batch in train_boundary_loader:
it += 1
loss = trainer.train_step(batch, m=2.0)
logger.add_scalar('train/loss', loss, epoch_it)
# Print output
if print_every > 0 and (it % print_every) == 0:
t = datetime.datetime.now()
print('[Epoch %02d] it=%03d, loss=%.4f, time: %.2fs, %02d:%02d'
% (epoch_it, it, loss, time.time() - t0, t.hour, t.minute))
checkpoint_io.save('model.pt', epoch_it=epoch_it, it=it,
loss_val_best=metric_val_best)
# run validation
if validate_every > 0 and (epoch_it % validate_every) == 0:
eval_dict = trainer.evaluate(val_loader)
metric_val = eval_dict[model_selection_metric]
print('Validation metric (%s): %.4f'
% (model_selection_metric, metric_val))
for k, v in eval_dict.items():
logger.add_scalar('val/%s' % k, v, it)
if model_selection_sign * (metric_val - metric_val_best) > 0:
metric_val_best = metric_val
print('New best model (loss %.4f)' % metric_val_best)
checkpoint_io.save('model_best.pt', epoch_it=epoch_it, it=it,
loss_val_best=metric_val_best)
checkpoint_io.save('model_%d.pt' % epoch_it, epoch_it=epoch_it, it=it,
loss_val_best=metric_val_best)