-
Notifications
You must be signed in to change notification settings - Fork 41
/
train.py
122 lines (86 loc) · 3.62 KB
/
train.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
import os
os.environ["CUDA_VISIBLE_DEVICES"]="0"
#os.environ["WANDB_API_KEY"] = ""
from pathlib import Path
import time
from data.dataset import TextDataset, TextDatasetval
from models import create_model
import torch
import cv2
import os
import numpy as np
from itertools import cycle
from scipy import linalg
from models.model import TRGAN
from params import *
from torch import nn
import wandb
def main():
wandb.init(project="hwt-final", name = EXP_NAME)
init_project()
TextDatasetObj = TextDataset(num_examples = NUM_EXAMPLES)
dataset = torch.utils.data.DataLoader(
TextDatasetObj,
batch_size=batch_size,
shuffle=True,
num_workers=0,
pin_memory=True, drop_last=True,
collate_fn=TextDatasetObj.collate_fn)
TextDatasetObjval = TextDatasetval(num_examples = NUM_EXAMPLES)
datasetval = torch.utils.data.DataLoader(
TextDatasetObjval,
batch_size=batch_size,
shuffle=True,
num_workers=0,
pin_memory=True, drop_last=True,
collate_fn=TextDatasetObjval.collate_fn)
model = TRGAN()
os.makedirs('saved_models', exist_ok = True)
MODEL_PATH = os.path.join('saved_models', EXP_NAME)
if os.path.isdir(MODEL_PATH) and RESUME:
model.load_state_dict(torch.load(MODEL_PATH+'/model.pth'))
print (MODEL_PATH+' : Model loaded Successfully')
else:
if not os.path.isdir(MODEL_PATH): os.mkdir(MODEL_PATH)
for epoch in range(EPOCHS):
start_time = time.time()
for i,data in enumerate(dataset):
if (i % NUM_CRITIC_GOCR_TRAIN) == 0:
model._set_input(data)
model.optimize_G_only()
model.optimize_G_step()
if (i % NUM_CRITIC_DOCR_TRAIN) == 0:
model._set_input(data)
model.optimize_D_OCR()
model.optimize_D_OCR_step()
if (i % NUM_CRITIC_GWL_TRAIN) == 0:
model._set_input(data)
model.optimize_G_WL()
model.optimize_G_step()
if (i % NUM_CRITIC_DWL_TRAIN) == 0:
model._set_input(data)
model.optimize_D_WL()
model.optimize_D_WL_step()
end_time = time.time()
data_val = next(iter(datasetval))
losses = model.get_current_losses()
page = model._generate_page(model.sdata, model.input['swids'])
page_val = model._generate_page(data_val['simg'].to(DEVICE), data_val['swids'])
wandb.log({'loss-G': losses['G'],
'loss-D': losses['D'],
'loss-Dfake': losses['Dfake'],
'loss-Dreal': losses['Dreal'],
'loss-OCR_fake': losses['OCR_fake'],
'loss-OCR_real': losses['OCR_real'],
'loss-w_fake': losses['w_fake'],
'loss-w_real': losses['w_real'],
'epoch' : epoch,
'timeperepoch': end_time-start_time,
})
wandb.log({ "result":[wandb.Image(page, caption="page"),wandb.Image(page_val, caption="page_val")],
})
print ({'EPOCH':epoch, 'TIME':end_time-start_time, 'LOSSES': losses})
if epoch % SAVE_MODEL == 0: torch.save(model.state_dict(), MODEL_PATH+ '/model.pth')
if epoch % SAVE_MODEL_HISTORY == 0: torch.save(model.state_dict(), MODEL_PATH+ '/model'+str(epoch)+'.pth')
if __name__ == "__main__":
main()