-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtrain_ERA5_Land.py
319 lines (245 loc) · 13 KB
/
train_ERA5_Land.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
# ------------------------------------------------------------------
"""
Script for training and validating on ERA5-Land dataset
Contact Person: Mohamad Hakam Shams Eddin <[email protected]>
Computer Vision Group - Institute of Computer Science III - University of Bonn
"""
# ------------------------------------------------------------------
import torch
import numpy as np
from tqdm import tqdm
from models.losses import BCE_loss, Anomaly_L1_loss
import utils.utils_train as utils
from models.build import VQ_model
import time
import os
from torch.utils.tensorboard import SummaryWriter
from dataset.ERA5_Land_dataset import ERA5_Land_Dataset
import config as config_file
#import matplotlib.pyplot as plt
np.set_printoptions(suppress=True)
torch.set_printoptions(sci_mode=False)
np.seterr(divide='ignore', invalid='ignore')
# torch.autograd.set_detect_anomaly(True)
# ------------------------------------------------------------------
def train(config_file):
# read config arguments
config = config_file.read_arguments(train=True)
# get logger
logger = utils.get_logger(config)
# get tensorboard writer
writer = SummaryWriter(os.path.join(config.dir_log, config.name))
# fix random seed
utils.fix_seed(config.seed)
# dataloader
utils.log_string(logger, "loading training dataset ...")
train_dataset = ERA5_Land_Dataset(
root_ERA5_Land=config.root_ERA5_Land,
root_NOAA=config.root_NOAA,
nan_fill=config.nan_fill,
is_aug=config.is_aug,
is_shuffle=config.is_shuffle,
is_norm=config.is_norm,
is_clima_scale=config.is_clima_scale,
variables=config.variables,
years=config.years_train,
threshold=config.threshold,
alpha=config.alpha,
x_min=config.x_min,
x_max=config.x_max,
y_min=config.y_min,
y_max=config.y_max,
delta_t=config.delta_t,
window_size=config.window_size
)
utils.log_string(logger, "loading validation dataset ...")
val_dataset = ERA5_Land_Dataset(
root_ERA5_Land=config.root_ERA5_Land,
root_NOAA=config.root_NOAA,
nan_fill=config.nan_fill,
is_aug=False,
is_shuffle=False,
is_norm=config.is_norm,
is_clima_scale=config.is_clima_scale,
variables=config.variables,
years=config.years_val,
threshold=config.threshold,
alpha=config.alpha,
x_min=config.x_min,
x_max=config.x_max,
y_min=config.y_min,
y_max=config.y_max,
delta_t=config.delta_t,
window_size=config.window_size
)
train_dataloader = torch.utils.data.DataLoader(train_dataset,
batch_size=config.batch_size,
shuffle=True,
pin_memory=config.pin_memory,
num_workers=config.n_workers)
val_dataloader = torch.utils.data.DataLoader(val_dataset,
batch_size=config.batch_size,
drop_last=False,
shuffle=True,
pin_memory=config.pin_memory,
num_workers=config.n_workers)
utils.log_string(logger, "# training samples: %d" % len(train_dataset))
utils.log_string(logger, "# evaluation samples: %d" % len(val_dataset))
# get models
utils.log_string(logger, "\nloading the model ...")
if config.gpu_id != "-1":
os.environ["CUDA_DEVICE_ORDER"] = "PCI_BUS_ID"
os.environ["CUDA_VISIBLE_DEVICES"] = str(config.gpu_id)
device = 'cuda'
else:
device = 'cpu'
model = VQ_model(config)
utils.log_string(logger, "model parameters ...")
utils.log_string(logger, "encoder parameters: %d" % utils.count_parameters(model.encoder))
utils.log_string(logger, "classifier parameters: %d" % utils.count_parameters(model.cls))
utils.log_string(logger, "vq parameters: %d" % utils.count_parameters(model.vq))
utils.log_string(logger, "all parameters: %d\n" % utils.count_parameters(model))
# get losses
utils.log_string(logger, "get criterion ...")
criterion = BCE_loss().to(device)
criterion_anomaly = Anomaly_L1_loss(n_dynamic=config.in_channels_dynamic,
delta_t=config.delta_t,
dim=config.en_embed_dim[-1]
).to(device)
# get optimizer
utils.log_string(logger, "get optimizer and learning rate scheduler ...")
optimizer = torch.optim.Adam(model.parameters(), lr=config.lr,
weight_decay=config.weight_decay,
betas=(config.beta1, config.beta2))
lr_scheduler = utils.get_learning_scheduler(optimizer, config)
# DataParallel fur multi-GPU
if torch.cuda.device_count() > 1:
model = torch.nn.DataParallel(model)
model.to(device)
# training loop
utils.log_string(logger, 'training on ERA5 Land dataset ...\n')
eval_train = utils.evaluator(logger, 'Training', config)
eval_val = utils.evaluator(logger, 'Validation', config)
time.sleep(1)
# initialize the best values
best_loss_train = np.inf
best_loss_val = np.inf
best_F1_train = 0
best_F1_val = 0
for epoch in range(config.n_epochs):
utils.log_string(logger, '################# Epoch (%s/%s) #################' % (epoch + 1, config.n_epochs))
model = model.train()
loss_train = 0
time.sleep(1)
for i, (data_d, data_drought, data_drought_loss, data_cold_surface, data_cold_surface_loss,
data_sea, data_no_vegetation, _) in tqdm(enumerate(train_dataloader),
total=len(train_dataloader),
smoothing=0.9,
postfix=" training"):
optimizer.zero_grad(set_to_none=True)
mask_valid = 1 - data_cold_surface # - data_no_vegetation
mask_valid[mask_valid < 0] = 0
pred, pred_y, _, z_q, loss_z_q = model(data_d.to(device))
# compute extreme loss
loss = criterion(pred[:, 0, ...], data_drought.to(device), mask_valid.to(device))
# compute anomaly loss
if torch.cuda.device_count() > 1:
loss_anomaly = criterion_anomaly(z_q,
data_drought_loss.to(device).float(),
data_cold_surface_loss.to(device),
model.module.vq.indices_to_codes(torch.Tensor([0]).long().to(device)).detach())
else:
loss_anomaly = criterion_anomaly(z_q,
data_drought_loss.to(device).float(),
data_cold_surface_loss.to(device),
model.vq.indices_to_codes(torch.Tensor([0]).long().to(device)).detach())
# compute multi-head extreme loss
loss_var = 0
for k in range(config.in_channels_dynamic):
loss_var += criterion(pred_y[k][:, 0, ...], data_drought.to(device).float(), mask_valid.to(device))
loss = loss + loss_anomaly * config.lambda_anomaly + loss_var + loss_z_q
loss.backward()
optimizer.step()
loss_train += loss.item()
pred = torch.sigmoid(pred.detach().cpu())
pred_c = pred.clone()
pred_c[pred > 0.35] = 1
pred_c[pred <= 0.35] = 0
eval_train(pred_c.numpy(), data_drought.cpu().numpy(), mask_valid.cpu().numpy())
mean_loss_train = loss_train / float(len(train_dataloader))
eval_train.get_results(mean_loss_train, best_loss_train)
if mean_loss_train <= best_loss_train:
best_loss_train = mean_loss_train
if eval_train.F1[1] >= best_F1_train:
best_F1_train = eval_train.F1[1]
#utils.save_model(model, optimizer, epoch, mean_loss_train, np.nan, logger, config, 'train')
# validation
with torch.no_grad():
model = model.eval()
loss_val = 0
time.sleep(1)
for i, (data_d, data_drought, data_drought_loss, data_cold_surface, data_cold_surface_loss,
data_sea, data_no_vegetation, _) in tqdm(enumerate(val_dataloader),
total=len(val_dataloader),
smoothing=0.9,
postfix=" validation"):
optimizer.zero_grad(set_to_none=True)
mask_valid = 1 - data_cold_surface - data_no_vegetation
mask_valid[mask_valid < 0] = 0
pred, pred_y, anomaly, z_q, loss_z_q = model(data_d.to(device))
anomaly = anomaly.float()
# compute extreme loss
loss = criterion(pred[:, 0, ...], data_drought.to(device), mask_valid.to(device))
# compute anomaly loss
if torch.cuda.device_count() > 1:
loss_anomaly = criterion_anomaly(z_q,
data_drought_loss.to(device).float(),
data_cold_surface_loss.to(device),
model.module.vq.indices_to_codes(torch.Tensor([0]).long().to(device)).detach())
else:
loss_anomaly = criterion_anomaly(z_q,
data_drought_loss.to(device).float(),
data_cold_surface_loss.to(device),
model.vq.indices_to_codes(torch.Tensor([0]).long().to(device)).detach())
# compute multi-head extreme loss
loss_var = 0
for k in range(config.in_channels_dynamic):
loss_var += criterion(pred_y[k][:, 0, ...], data_drought.to(device).float(), mask_valid.to(device))
loss = loss + loss_anomaly * config.lambda_anomaly + loss_var + loss_z_q
loss_val += loss.item()
pred = torch.sigmoid(pred.detach().cpu())
pred_c = pred.clone()
pred_c[pred > 0.35] = 1
pred_c[pred <= 0.35] = 0
eval_val(pred_c.numpy(), data_drought.cpu().numpy(), mask_valid.cpu().numpy())
if i == len(val_dataloader) - 1:
# plot results
im_pred, im_pred_c, im_target = utils.generate_images(pred[:, 0, :, :].cpu().numpy(),
pred_c[:, 0, :, :].numpy(),
data_drought.cpu().numpy(),
data_cold_surface.cpu().numpy(),
data_sea.numpy(),
data_no_vegetation.numpy(),
mask_valid.numpy())
im_anomaly = utils.generate_anomaly(anomaly.cpu().numpy())
mean_loss_val = loss_val / float(len(val_dataloader))
eval_val.get_results(mean_loss_val, best_loss_val)
if mean_loss_val <= best_loss_val:
best_loss_val = mean_loss_val
utils.save_model(model, optimizer, epoch, mean_loss_train, mean_loss_val, logger, config, 'loss')
if eval_val.F1[1] >= best_F1_val:
best_F1_val = eval_val.F1[1]
utils.save_model(model, optimizer, epoch, mean_loss_train, mean_loss_val, logger, config, 'F1')
writer.add_images('probability', im_pred, epoch + 1, dataformats='NHWC')
writer.add_images('prediction', im_pred_c, epoch + 1, dataformats='NHWC')
writer.add_images('target', im_target, epoch + 1, dataformats='NHWC')
for v, var in enumerate(val_dataset.variables_dynamic):
writer.add_images(var, im_anomaly[0, v, ...], epoch + 1, dataformats='HWC')
writer.add_scalars("Loss", {'train': mean_loss_train, 'val': mean_loss_val}, epoch + 1)
writer.add_scalars("IOU", {'train': eval_train.iou[1], 'val': eval_val.iou[1]}, epoch + 1)
writer.add_scalars("F1", {'train': eval_train.F1[1], 'val': eval_val.F1[1]}, epoch + 1)
eval_train.reset()
eval_val.reset()
lr_scheduler.step_update(epoch)
if __name__ == '__main__':
train(config_file)