-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathtrain_denoise.py
223 lines (174 loc) · 9.51 KB
/
train_denoise.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
import torch
import numpy as np
import torch.optim as optim
from torch.cuda import amp
from tqdm import tqdm
import time
from dataset import dataloader_semseg as dataloader
from tools import dir_utils
from configs.load_yaml import load_yaml
from tools.lr_warmup_scheduler import GradualWarmupScheduler
def main(yaml_file,test_mode=False):
device = torch.device('cuda')
opt = load_yaml(yaml_file,saveYaml2output=True)
epoch = opt.OPTIM.NUM_EPOCHS
model_dir = opt.SAVE_DIR+'models/'
dir_utils.mkdir_with_del(model_dir)
######### dataset ###########
train_dataset = dataloader.Noise_Dataset(opt.DATASET.TRAIN_CSV, leads=opt.DATASET_CUSTOME.LEADS,
date_len=opt.DATASET_CUSTOME.INPUT_LENGTH,
n_max_cls=opt.DATASET_CUSTOME.OUT_C,
random_crop=True,
transform = dataloader.get_transform(train=True)
)
val_dataset = dataloader.Noise_Dataset(opt.DATASET.VAL_CSV, leads=opt.DATASET_CUSTOME.LEADS,
date_len=opt.DATASET_CUSTOME.INPUT_LENGTH,
n_max_cls=opt.DATASET_CUSTOME.OUT_C,
random_crop=False,
transform = dataloader.get_transform(train=False)
)
train_dataloader = torch.utils.data.DataLoader(train_dataset, batch_size=opt.OPTIM.BATCH_SIZE,
shuffle=True, num_workers=4,
prefetch_factor=3,
persistent_workers=False, #maintain woker alive even consumed
)
val_dataloader = torch.utils.data.DataLoader(val_dataset, batch_size=opt.OPTIM.BATCH_SIZE,
shuffle=False, num_workers=4,
prefetch_factor=3,
persistent_workers=False, #maintain woker alive even consumed
# drop_last=True,
)
dataset_sizes = {'train':len(train_dataset),
'val':len(val_dataset)}
print('===> Loading datasets done')
######### model ###########
from models.model import Model
model = Model(in_c=1,
out_c=opt.DATASET_CUSTOME.OUT_C, \
img_size=opt.DATASET_CUSTOME.INPUT_LENGTH, \
embed_dim=opt.MODEL.EMBED_DIM, \
patch_size=opt.MODEL.PATCH_SIZE, \
window_size=opt.MODEL.WINDOW_SIZE, \
depths=opt.MODEL.DEPTHS, \
num_heads=opt.MODEL.N_HEADS, \
# denoise_mode=opt.MODEL.Denoise_Mode, \
).to(device)
######### optim ###########
new_lr = opt.OPTIM.LR_INITIAL
optimizer = optim.Adam(model.parameters(), lr=new_lr, betas=(0.9, 0.999),eps=1e-8)
warmup_epochs = 3
scheduler_cosine = optim.lr_scheduler.CosineAnnealingLR(optimizer, opt.OPTIM.NUM_EPOCHS-warmup_epochs, eta_min=opt.OPTIM.LR_MIN)
scheduler = GradualWarmupScheduler(optimizer, multiplier=1, total_epoch=warmup_epochs, after_scheduler=scheduler_cosine)
scheduler.step()
criterion = torch.nn.CrossEntropyLoss(weight=torch.FloatTensor([1,1,1]).to(device))
# criterion = torch.nn.HuberLoss(delta=1.0)
grad_scaler = amp.GradScaler()
start_epoch = 0
for epoch in range(start_epoch, opt.OPTIM.NUM_EPOCHS + 1):
# for epoch in range(start_epoch, 2 + 1):
epoch_start_time = time.time()
epoch_train_loss = 0
train_outputs_all,train_labels_all,train_inputs_all = [],[],[]
#### train ####
model.train()
for i, data in enumerate(tqdm(train_dataloader), 0):
# for i, data in enumerate(train_dataloader):
inputs = data['input'].to(device)
labels = data['label'].to(device)
optimizer.zero_grad()
# with torch.set_grad_enabled(True):
torch.set_grad_enabled(True)
with amp.autocast():
outputs = model(inputs)
loss = criterion(outputs, labels)
#val for SNR
train_outputs = outputs.cpu().detach().numpy().flatten().tolist()
train_labels = labels.cpu().numpy().flatten().tolist()
train_inputs = inputs.cpu().numpy().flatten().tolist()
train_outputs_all += train_outputs
train_labels_all += train_labels
train_inputs_all += train_inputs
grad_scaler.scale(loss).backward()
grad_scaler.step(optimizer)
grad_scaler.update()
epoch_train_loss += loss.item() * inputs.size(0)
train_loss_mean = epoch_train_loss / dataset_sizes['train']
train_inputs_all = np.array(train_inputs_all)
train_outputs_all = np.array(train_outputs_all)
train_labels_all = np.array(train_labels_all)
# RMSE
RMSE_train_input = np.sqrt(np.mean((train_labels_all-train_inputs_all)**2))
RMSE_train_output = np.sqrt(np.mean((train_labels_all-train_outputs_all)**2))
train_RMSEde = RMSE_train_input-RMSE_train_output
train_RMSEde_perc = train_RMSEde/RMSE_train_input
train_average_input_noise = np.mean((train_labels_all-train_inputs_all)**2)
train_average_output_noise = np.mean((train_labels_all-train_outputs_all)**2)
train_average_signal = np.mean(train_labels_all**2)
SNR_train_input = np.log10(train_average_signal/train_average_input_noise)
SNR_train_output = np.log10(train_average_signal/train_average_output_noise)
SNR_train_diff = SNR_train_output-SNR_train_input
#### Evaluation ####
model.eval()
epoch_val_loss = 0
val_outputs_all,val_labels_all,val_inputs_all = [],[],[]
for data in val_dataloader:
inputs = data['input'].to(device)
labels = data['label'].to(device)
torch.set_grad_enabled(False)
outputs = model(inputs)
loss = criterion(outputs, labels)
outputs = np.argmax(outputs.cpu().detach().numpy(), axis=1)
epoch_val_loss += loss.item()* inputs.size(0)
# SNR
val_outputs = outputs.cpu().detach().numpy().flatten().tolist()
val_labels = labels.cpu().numpy().flatten().tolist()
val_inputs = inputs.cpu().numpy().flatten().tolist()
val_outputs_all += val_outputs
val_labels_all += val_labels
val_inputs_all += val_inputs
epoch_val_loss += loss.item()* inputs.size(0)
val_loss_mean = epoch_val_loss / dataset_sizes['val']
val_loss_mean = epoch_val_loss / dataset_sizes['val']
val_inputs_all = np.array(val_inputs_all)
val_outputs_all = np.array(val_outputs_all)
val_labels_all = np.array(val_labels_all)
# RMSE
RMSE_val_input = np.sqrt(np.mean((val_labels_all-val_inputs_all)**2))
RMSE_val_output = np.sqrt(np.mean((val_labels_all-val_outputs_all)**2))
val_RMSEde = RMSE_val_input-RMSE_val_output
val_RMSEde_perc = val_RMSEde/RMSE_val_input
# SNR
val_average_input_noise = np.mean((val_labels_all-val_inputs_all)**2)
val_average_output_noise = np.mean((val_labels_all-val_outputs_all)**2)
val_average_signal = np.mean(val_labels_all**2)
SNR_val_input = np.log10(val_average_signal/val_average_input_noise)
SNR_val_output = np.log10(val_average_signal/val_average_output_noise)
SNR_val_diff = SNR_val_output-SNR_val_input
val_inputs = inputs.cpu().numpy()
val_label = labels.cpu().numpy()
val_outputs = outputs.cpu().detach().numpy()
scheduler.step()
save_path = model_dir+'model_epoch_{}_val_{:.6f}.pth'.format(epoch,val_loss_mean)
torch.save(model, save_path)
print(save_path)
# assert 1>2
print("------------------------------------------------------------------")
print("Epoch: {}\tTime: {:.4f}s \t train Loss: {:.6f} \t val loss: {:.6f} LearningRate {:.8f}".format(
epoch, time.time()-epoch_start_time, train_loss_mean, val_loss_mean, scheduler.get_lr()[0]))
print('train_inputs_RMSE {:.6f}, train_output_RMSE: {:.6f} train_RMSEde {:.6} \
val_inputs_RMSE {:.6f}, val_output_RMSE: {:.6f} val_RMSEde {:.6f} '.format(
RMSE_train_input,RMSE_train_output,train_RMSEde,
RMSE_val_input,RMSE_val_output,val_RMSEde))
print('train_inputs_SNR {:.6f}, SNR_train_output: {:.6f} SNR_train_diff {:.6f} \
SNR_val_input {:.6f}, SNR_val_output: {:.6f} SNR_val_diff {:.6f} '.format(
SNR_train_input,SNR_train_output,SNR_train_diff,
SNR_val_input,SNR_val_output,SNR_val_diff))
print("------------------------------------------------------------------")
if __name__ == "__main__":
from argparse import ArgumentParser
parser = ArgumentParser(description="train")
parser.add_argument("-c", "--config", type=str,
default=None,
help="path to yaml file")
args = parser.parse_args()
main(args.config,test_mode=False)