-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathengine_fer.py
160 lines (127 loc) · 6.36 KB
/
engine_fer.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
import os
import numpy as np
import scipy
import torch
import torch.nn.functional as F
from tqdm import tqdm
import utils.lr_sched as lrs
from utils.misc import compute_ACC, Averager, compute_MCACC
from torch.cuda.amp import autocast
def train(model, clip_model, args, optimizer, criterion, dataloader, logger, label_token, epoch, save_model=False):
logger.info("=======================TRAINING MODE, Epoch: {}/{}=======================".format(epoch, args.epochs))
print("=======================TRAINING MODE, Epoch: {}/{}=======================".format(epoch, args.epochs))
mean_dist_loss = 0
mean_rank_loss = 0
tl = Averager()
ta = Averager()
best_acc = 0.0
best_epoch = 0
for i, (train_inputs, train_labels) in enumerate(tqdm(dataloader)): # train_inputs: (bs,3,224,224), train_labels: (bs, 7)
lrs.adjust_learning_rate(optimizer, i / len(dataloader) + epoch, args)
optimizer.zero_grad()
train_inputs = train_inputs.cuda()
train_labels = train_labels.cuda()
label_token = label_token.cuda()
logits, _, dist_feat = model(train_inputs, label_token) # logits:(bs,7), _:(bs,196,512), dist_feat:(bs,512)
if args.loss_function == 'edl':
v_loss = criterion(logits, train_labels, n_class=args.classes, epoch=epoch, total_epoch=args.epochs) # EDL loss
cls_loss = v_loss['loss_cls'].mean()
if 'loss_kl' in v_loss:
v_loss_kl = v_loss['loss_kl'].mean()
rank_loss = cls_loss + v_loss_kl
elif 'loss_avu' in v_loss:
v_loss_avu = v_loss['loss_avu'].mean()
rank_loss = cls_loss + args.lamda_ruc * v_loss_avu
else:
rank_loss = cls_loss
else:
rank_loss = criterion(logits, train_labels) # CE loss
with torch.no_grad():
_, tea_dist_feat = clip_model.encode_image(train_inputs)
dist_loss = F.l1_loss(dist_feat, tea_dist_feat.float())
loss = rank_loss + args.lamda_kd * dist_loss
if args.loss_function == 'edl':
probs = model.evd_results(logits)['probs']
if args.dataset=='affectnet' or args.dataset=='affectnet_8':
acc = compute_MCACC(probs, train_labels)
else:
acc = compute_ACC(probs, train_labels)
else:
if args.dataset=='affectnet' or args.dataset=='affectnet_8':
acc = compute_MCACC(logits, train_labels)
else:
acc = compute_ACC(logits, train_labels)
tl.add(loss.item())
ta.add(acc)
mean_dist_loss += dist_loss.item()
mean_rank_loss += rank_loss.item()
loss.requires_grad_()
loss.backward()
optimizer.step()
mean_dist_loss /= len(dataloader)
mean_rank_loss /= len(dataloader)
learning_rate = optimizer.param_groups[-1]['lr']
tl = tl.item()
ta = ta.item()
logger.info("FINETUNING Epoch: {}/{} \tLoss: {:.4f} \tRankLoss: {:.4f} \tDistLoss: {:.4f} \tLearningRate {:.6f} \tTrain Acc: {:.4f} ".format(epoch, args.epochs, tl, mean_rank_loss, mean_dist_loss, learning_rate, ta))
print("FINETUNING Epoch: {}/{} \tLoss: {:.4f} \tRankLoss: {:.4f} \tDistLoss: {:.4f} \tLearningRate {:.6f} \tTrain Acc: {:.4f} ".format(epoch, args.epochs, tl, mean_rank_loss, mean_dist_loss, learning_rate, ta))
if save_model:
torch.save(model.state_dict(), os.path.join(args.record_path, "model_epoch_{}.pth".format(epoch)))
return ta, tl
########### TEST FUNC ###########
def test(model, args, criterion, dataloader, logger, label_token, epoch=-1,save_np=False):
logger.info("-----------------------EVALUATION MODE-----------------------")
print("-----------------------EVALUATION MODE-----------------------")
predRST = []
labelRET = []
va = Averager()
vl = Averager()
with torch.no_grad():
for i, (features, labels) in enumerate(tqdm(dataloader)):
pred_feat, dist_feat = model.encode_img(features.cuda())
label_emb = model.forward_text(label_token.cuda())
score1 = torch.topk(pred_feat @ label_emb.t(), k=model.topk, dim=1)[0].mean(dim=1)
score2 = dist_feat @ label_emb.t()
score1 = score1 / score1.norm(dim=-1, keepdim=True)
score2 = score2 / score2.norm(dim=-1, keepdim=True)
logits = args.alpha * score1 + (1-args.alpha) * score2
if args.loss_function == 'edl':
v_loss = criterion(logits, labels.cuda(), n_class=args.classes, epoch=epoch, total_epoch=args.epochs) # EDL loss
cls_loss = v_loss['loss_cls'].mean()
if 'loss_kl' in v_loss:
v_loss_kl = v_loss['loss_kl'].mean()
loss = cls_loss + v_loss_kl
elif 'loss_avu' in v_loss:
v_loss_avu = v_loss['loss_avu'].mean()
loss = cls_loss + v_loss_avu
else:
loss = cls_loss
probs = model.evd_results(logits)['probs']
if args.dataset=='affectnet' or args.dataset=='affectnet_8':
acc = compute_MCACC(probs, labels.cuda())
else:
acc = compute_ACC(probs, labels.cuda())
else:
loss = criterion(logits, labels.cuda())
if args.dataset=='affectnet' or args.dataset=='affectnet_8':
acc = compute_MCACC(logits, labels.cuda())
else:
acc = compute_ACC(logits, labels.cuda())
vl.add(loss.item())
va.add(acc)
preds_np = np.array(torch.argmax(logits, dim=1).cpu())
labels_np = np.array(labels)
predRST.append(preds_np)
labelRET.append(labels_np)
if save_np:
# 转换成一维数组
predRST = np.concatenate(predRST, axis=0)
labelRET = np.concatenate(labelRET, axis=0)
# 保存成mat文件, predRST为a列, labelRET为b列
scipy.io.savemat(os.path.join(args.record_path, "predRST_ep{:d}.mat".format(epoch)), {'a': predRST, 'b': labelRET})
logger.info("completed calculating predictions over all images")
vl = vl.item()
va = va.item()
logger.info("Test Loss: {:.4f} \t Test ACC: {:.4f}". format(vl, va))
print("Test Loss: {:.4f} \t Test ACC: {:.4f}". format(vl, va))
return va, vl