-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathtrain.py
executable file
·196 lines (151 loc) · 7.29 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
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
from scipy.linalg import sqrtm
import torch.nn.functional as F
from torch.autograd import Variable
import numpy as np
import torch
import torch.optim
import os
import pdb
# import configs
from network import NSAE_model
import configs
from io_utils import model_dict, parse_args
from datasets import miniImageNet_few_shot
import torch.nn as nn
from sklearn.neighbors import NearestNeighbors
def train(base_loader, model, optimization, start_epoch, stop_epoch, params):
out_pre = '%s_%s' % (params.model, params.method)
model.train()
for epoch in range(start_epoch, stop_epoch):
model.train_autoencoder(epoch, base_loader, optimizer)
if (epoch % params.save_freq == 0) or (epoch == stop_epoch - 1):
outfile = os.path.join(params.checkpoint_dir, '%s_%d.tar' % (out_pre, epoch))
torch.save({'epoch': epoch, 'state': model.state_dict()}, outfile)
outfile_final = os.path.join(params.checkpoint_dir, '%s.tar' % out_pre)
torch.save({'epoch': stop_epoch - 1, 'state': model.state_dict()}, outfile_final)
return model
def finetune(model, optimizer, novel_loader, finetune_epochs_recon, finetune_epochs, n_way, n_support, n_query):
model.train()
acc_all_ori = []
acc_all_lp = []
# model.finetune(novel_loader, n_way, n_support, n_query)
for i, (x, y) in enumerate(novel_loader):
# n_query = x.size(1) - n_support
x = x.cuda()
x_var = Variable(x)
batch_size = 4
support_size = n_way * n_support
y_a_i = Variable(torch.from_numpy(np.repeat(range(n_way), n_support))).cuda()
x_b_i = x_var[:, n_support:, :, :, :].contiguous().view(n_way * n_query, *x.size()[2:])
x_a_i = x_var[:, :n_support, :, :, :].contiguous().view(n_way * n_support, *x.size()[2:])
for epoch in range(finetune_epochs_recon):
rand_id = np.random.permutation(support_size)
for j in range(0, support_size, batch_size):
selected_id = torch.from_numpy( rand_id[j: min(j+batch_size, support_size)]).cuda()
z_batch = x_a_i[selected_id]
y_batch = y_a_i[selected_id]
scores, x_hat, x_hat_scores, _ = model(z_batch)
loss = torch.mean((z_batch.cuda() - x_hat) ** 2)
print('finetune recon loss:', loss.item())
loss.backward()
optimizer.step()
optimizer = torch.optim.SGD(list(model.parameters()), lr=0.01, momentum=0.9, weight_decay=1e-3)
cross_ent_loss = nn.CrossEntropyLoss()
for epoch in range(finetune_epochs):
rand_id = np.random.permutation(support_size)
for j in range(0, support_size, batch_size):
selected_id = torch.from_numpy( rand_id[j: min(j+batch_size, support_size)]).cuda()
z_batch = x_a_i[selected_id]
y_batch = y_a_i[selected_id]
scores, x_hat, x_hat_scores, _ = model(z_batch)
# loss = torch.mean((z_batch.cuda() - x_hat) ** 2)
loss = cross_ent_loss(scores, y_batch)
print('finetune loss:', loss.item())
loss.backward()
optimizer.step()
model.eval()
with torch.no_grad():
scores, x_hat, x_hat_scores, output = model(x_b_i)
# output = pretrained_model(x_b_i)
# scores = classifier(output)
x_lp = output.cpu().numpy()
y_lp = F.softmax(scores, 1).cpu().numpy()
topk_scores, topk_labels = scores.data.topk(1, 1, True, True)
topk_ind_ori = topk_labels.cpu().numpy()
y_query = np.repeat(range(n_way), n_query)
neigh = NearestNeighbors(params.k_lp)
neigh.fit(x_lp)
d_lp, idx_lp = neigh.kneighbors(x_lp)
d_lp = np.power(d_lp, 2)
sigma2_lp = np.mean(d_lp)
n_lp = len(y_query)
del_n = int(n_lp * (1.0 - params.delta))
for i in range(n_way):
yi = y_lp[:, i]
top_del_idx = np.argsort(yi)[0:del_n]
y_lp[top_del_idx, i] = 0
w_lp = np.zeros((n_lp, n_lp))
for i in range(n_lp):
for j in range(params.k_lp):
xj = idx_lp[i, j]
w_lp[i, xj] = np.exp(-d_lp[i, j] / (2 * sigma2_lp))
w_lp[xj, i] = np.exp(-d_lp[i, j] / (2 * sigma2_lp))
q_lp = np.diag(np.sum(w_lp, axis=1))
q2_lp = sqrtm(q_lp)
q2_lp = np.linalg.inv(q2_lp)
L_lp = np.matmul(np.matmul(q2_lp, w_lp), q2_lp)
a_lp = np.eye(n_lp) - params.alpha * L_lp
a_lp = np.linalg.inv(a_lp)
ynew_lp = np.matmul(a_lp, y_lp)
count_this = len(y_query)
top1_correct_ori = np.sum(topk_ind_ori[:, 0] == y_query)
correct_ori = float(top1_correct_ori)
print('BSR: %f' % (correct_ori / count_this * 100))
acc_all_ori.append((correct_ori / count_this * 100))
topk_ind_lp = np.argmax(ynew_lp, 1)
top1_correct_lp = np.sum(topk_ind_lp == y_query)
correct_lp = float(top1_correct_lp)
print('BSR+LP: %f' % (correct_lp / count_this * 100))
acc_all_lp.append((correct_lp / count_this * 100))
acc_all_ori = np.asarray(acc_all_ori)
acc_mean_ori = np.mean(acc_all_ori)
acc_std_ori = np.std(acc_all_ori)
print('BSR: %d Test Acc = %4.2f%% +- %4.2f%%' %
(finetune_epochs, acc_mean_ori, 1.96 * acc_std_ori / np.sqrt(finetune_epochs)))
acc_all_lp = np.asarray(acc_all_lp)
acc_mean_lp = np.mean(acc_all_lp)
acc_std_lp = np.std(acc_all_lp)
print('BSR+LP: %d Test Acc = %4.2f%% +- %4.2f%%' %
(finetune_epochs, acc_mean_lp, 1.96 * acc_std_lp / np.sqrt(finetune_epochs)))
if __name__=='__main__':
np.random.seed(10)
params = parse_args('train')
finetune_params = parse_args('finetune')
finetune_epochs_recon = 1 #30
finetune_epochs = 1 #200
model = NSAE_model(model_dict[params.model], params.num_classes, lamda1=params.lamda1, lamda2=params.lamda2)
model = model.cuda()
image_size = 224
optimization = 'SGD'
if optimization == 'Adam':
optimizer = torch.optim.Adam(model.parameters())
elif optimization == 'SGD':
optimizer = torch.optim.SGD(list(model.parameters()), lr=0.001, momentum=0.9, weight_decay=0.0005)
else:
raise ValueError('Unknown optimization, please define by yourself')
datamgr = miniImageNet_few_shot.SimpleDataManager(image_size, batch_size=16)
base_loader = datamgr.get_data_loader(aug=params.train_aug)
few_shot_params = dict(n_way=finetune_params.test_n_way, n_support=finetune_params.n_shot, n_query=15)
datamgr = miniImageNet_few_shot.SetDataManager(image_size, n_eposide=1, **few_shot_params)
novel_loader = datamgr.get_data_loader(aug=False)
save_dir = configs.save_dir
params.method = 'bsr'
params.checkpoint_dir = '%s/checkpoints/%s_%s' % (save_dir, params.model, params.method)
if params.train_aug:
params.checkpoint_dir += '_aug'
if not os.path.isdir(params.checkpoint_dir):
os.makedirs(params.checkpoint_dir)
start_epoch = params.start_epoch
stop_epoch = params.stop_epoch
model = train(base_loader, model, optimizer, start_epoch, stop_epoch, params)
model = finetune(model, optimizer, novel_loader, finetune_epochs_recon, finetune_epochs, **few_shot_params)