-
Notifications
You must be signed in to change notification settings - Fork 0
/
train.py
146 lines (109 loc) · 4.74 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
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
import pickle
import argparse
import os
from tqdm import tqdm
import random
import numpy as np
import torch
from torch.utils.data import DataLoader
from model import RNModel
from utils import load_data, splice_data, tensorize
def train(model, rel, norel, args):
for epoch in range(args.epochs):
model.train()
random.shuffle(rel)
random.shuffle(norel)
rel_train = splice_data(rel)
norel_train = splice_data(norel)
acc_rels = []
acc_norels = []
loss_rels = []
loss_norels = []
# Create progress bar
num_steps = len(rel_train[0]) // args.batch_size
pbar = tqdm(total=num_steps * 2, desc='Epoch {}'.format(epoch+1))
# Training
for batch_idx in range(num_steps):
# Train on Relational questions
images, questions, answers = tensorize(rel_train, batch_idx, args)
rel_acc, rel_loss = model.train_(images, questions, answers)
# Train on Non-Relational questions
images, questions, answers = tensorize(norel_train, batch_idx, args)
norel_acc, norel_loss = model.train_(images, questions, answers)
# Progress Bar Logging
pbar.update(2)
pbar.set_postfix({'Relation Accuracy': rel_acc,
'Non-relations Accuracy': norel_acc})
acc_rels.append(rel_acc.item())
acc_norels.append(norel_acc.item())
loss_rels.append(rel_loss.item())
loss_norels.append(norel_loss.item())
# Save checkpoint
model.save_model(epoch+1)
mean_rel_acc = np.array(acc_rels).mean()
mean_norel_acc = np.array(acc_norels).mean()
pbar.set_postfix({'Mean Relation Accuracy': mean_rel_acc,
'Mean Non-relations Accuracy': mean_norel_acc})
pbar.close()
def test(model, rel, norel, args):
model.eval()
random.shuffle(rel)
random.shuffle(norel)
rel_test = splice_data(rel)
norel_test = splice_data(norel)
acc_rels = []
acc_norels = []
loss_rels = []
loss_norels = []
# Create progress bar
num_steps = len(rel_test[0]) // args.batch_size
pbar = tqdm(total=num_steps * 2, desc='Evaluating...')
with torch.no_grad():
for batch_idx in range(num_steps):
# Train on Relational questions
images, questions, answers = tensorize(rel_test, batch_idx, args)
rel_acc, rel_loss = model.evaluate(images, questions, answers)
# Train on Non-Relational questions
images, questions, answers = tensorize(norel_test, batch_idx, args)
norel_acc, norel_loss = model.evaluate(images, questions, answers)
# Progress Bar Logging
pbar.update(2)
pbar.set_postfix({'Relation Accuracy': rel_acc,
'Non-relations Accuracy': norel_acc})
acc_rels.append(rel_acc.item())
acc_norels.append(norel_acc.item())
loss_rels.append(rel_loss.item())
loss_norels.append(norel_loss.item())
mean_rel_acc = np.array(acc_rels).mean()
mean_norel_acc = np.array(acc_norels).mean()
pbar.set_postfix({'Mean Test Relation Accuracy': mean_rel_acc,
'Mean Test Non-relations Accuracy': mean_norel_acc})
pbar.close()
if __name__ == '__main__':
parser = argparse.ArgumentParser(description='Relation Network Sort-of-CLEVR Training Script')
parser.add_argument('--epochs', type=int, default=20, help='Number of training epochs')
parser.add_argument('--batch-size', type=int, default=64, help='Batch size')
parser.add_argument('--lr', type=float, default=0.0001, help='Learning rate')
parser.add_argument('--no-cuda', action='store_true', default=False, help='disables CUDA training')
parser.add_argument('--data-dir', type=str, default='data', help='Pickle data path')
args = parser.parse_args()
# Check cuda
args.device = 'cuda' if torch.cuda.is_available() and not args.no_cuda else 'cpu'
SEED = 1
torch.manual_seed(SEED)
if args.device == 'cuda':
torch.cuda.manual_seed(SEED)
# Loading Data from generated pickle files
rel_train, rel_test, norel_train, norel_test = load_data(args.data_dir)
# Create model
model = RNModel(args)
model.to(args.device)
# Create './models' directory for saving weights
try:
os.mkdir('models')
except:
pass
# Train
train(model, rel_train, norel_train, args)
# Evaluate on test set
test(model, rel_test, norel_test, args)