-
Notifications
You must be signed in to change notification settings - Fork 191
/
Copy pathaverage_log.py
138 lines (128 loc) · 6.22 KB
/
average_log.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
# Copyright (c) Microsoft Corporation.
# Licensed under the MIT License.
import os
import re
import operator
import numpy as np
save_path = r'../saved_models/'
static_dict = {}
def get_static(file_name):
re_bestAcc = r'BEST_EVAL_ACC: (([0-9]|\.)*)' # .group(1)
re_bestIt = r'at ([0-9]*)' # .group(1)
re_top1Acc = r"eval\/top-1-acc': (([0-9]|\.)*)"
re_top5Acc = r"eval\/top-5-acc': (([0-9]|\.)*)"
stat = {"bestAcc": 0,
"bestIt": 0,
"Top1Acc": [],
"Top5Acc": [],
}
with open(file_name, 'r', encoding='utf-8') as f:
lines = f.readlines()
continue_flag = False
for line in lines:
if '1048000 iteration' in line:
continue_flag = True
if continue_flag == False:
return {'Top1_1': [],
'Top1_20': [],
'Top1_50': [],
'Top5_1': [],
'Top5_20': [],
'Top5_50': [],
'BestAcc': 0,
'BestIt': 0,
'Finish': False}
with open(file_name, 'r', encoding='utf-8') as f:
lines = f.readlines()
for line in lines:
if line.endswith('iters\n'):
stat['bestAcc'] = re.search(re_bestAcc,line).group(1)
stat['bestIt'] = re.search(re_bestIt,line).group(1)
stat['Top1Acc'].append(re.search(re_top1Acc,line).group(1))
stat['Top5Acc'].append(re.search(re_top5Acc,line).group(1))
for i in range(len(stat['Top1Acc'])):
stat['Top1Acc'][i] = float(stat['Top1Acc'][i])
for i in range(len(stat['Top5Acc'])):
stat['Top5Acc'][i] = float(stat['Top5Acc'][i])
stat['bestAcc'] = float(stat['bestAcc'])
avg_1_1acc = stat['Top1Acc'][-1]
avg_20_1acc = sum(stat['Top1Acc'][-20:])/20
avg_50_1acc = sum(stat['Top1Acc'][-50:])/50
avg_1_5acc = stat['Top5Acc'][-1]
avg_20_5acc = sum(stat['Top5Acc'][-20:])/ 20
avg_50_5acc = sum(stat['Top5Acc'][-50:])/ 50
return {'Top1_1': avg_1_1acc,
'Top1_20': avg_20_1acc,
'Top1_50': avg_50_1acc,
'Top5_1': avg_1_5acc,
'Top5_20': avg_20_1acc,
'Top5_50': avg_50_1acc,
'BestAcc': stat['bestAcc'],
'BestIt': stat['bestIt'],
'Finish': True}
# str = r"[2021-04-13 15:57:33,078 INFO] 228000 iteration, USE_EMA: True, {'train/sup_loss': tensor(0.0311, device='cuda:0'), 'train/unsup_loss': tensor(0.2391, device='cuda:0'), 'train/total_loss': tensor(0.3913, device='cuda:0'), 'train/mask_ratio': tensor(0.5246, device='cuda:0'), 'lr': 0.028670201217471786, 'train/prefetch_time': 0.0050832958221435545,'train/run_time': 0.315829833984375, 'eval/loss': tensor(1.0763, device='cuda:0'), 'eval/top-1-acc': 0.6306},BEST_EVAL_ACC: 0.9348, at 173000 iters"
statics = {}
for name in os.listdir(save_path):
cur_path = save_path + name
if os.path.isdir(cur_path):
cur_name = name
for n in os.listdir(cur_path):
if n == 'log.txt':
#try:
# statics[cur_name] = get_static(cur_path + '/' + n)
#except:
# print(cur_path,'failed')
statics[cur_name] = get_static(cur_path + '/' + n)
statics = sorted(statics.items())
final_res = {}
for s in statics:
exp_name = '_'.join(s[0].split('_')[:-1])
if s[1]['Finish'] == False:
print(s[0], 'is not finished')
continue
if exp_name not in final_res:
tmp = {'Top1_1': [s[1]['Top1_1']*100],
'Top1_20': [s[1]['Top1_20']*100],
'Top1_50': [s[1]['Top1_50']*100],
'Top5_1': [s[1]['Top5_1']*100],
'Top5_20': [s[1]['Top5_20']*100],
'Top5_50': [s[1]['Top5_50']*100],
'BestAcc': [s[1]['BestAcc']*100]}
final_res[exp_name] = tmp
else:
final_res[exp_name]['Top1_1'].append(s[1]['Top1_1']*100)
final_res[exp_name]['Top1_20'].append(s[1]['Top1_20']*100)
final_res[exp_name]['Top1_50'].append(s[1]['Top1_50']*100)
final_res[exp_name]['Top5_1'].append(s[1]['Top5_1']*100)
final_res[exp_name]['Top5_20'].append(s[1]['Top5_20']*100)
final_res[exp_name]['Top5_50'].append(s[1]['Top5_50']*100)
final_res[exp_name]['BestAcc'].append(s[1]['BestAcc']*100)
#for k,v in final_res.items():
#print(k,'Last 01 epoch, Top1 acc, mean: ',np.mean(v['Top1_1']), ', std: ', np.std(v['Top1_1']))
#print(k,'Last 20 epoch, Top1 acc, mean: ',np.mean(v['Top1_20']), ', std: ', np.std(v['Top1_20']))
#print(k,'Last 50 epoch, Top1 acc, mean: ',np.mean(v['Top1_50']), ', std: ', np.std(v['Top1_50']))
#print(k,'Last 01 epoch, Top5 acc, mean: ',np.mean(v['Top5_1']), ', std: ', np.std(v['Top5_1']))
#print(k,'Last 20 epoch, Top5 acc, mean: ',np.mean(v['Top5_20']), ', std: ', np.std(v['Top5_20']))
#print(k,'Last 50 epoch, Top5 acc, mean: ',np.mean(v['Top5_50']), ', std: ', np.std(v['Top5_50']))
#print(k,'Best epoch, Top1 acc, mean: ',np.mean(v['BestAcc']), ', std: ', np.std(v['BestAcc']))
import xlwt
data_setting=['cifar10_40','cifar10_250','cifar10_4000','cifar100_400','cifar100_2500','cifar100_10000','stl10_40','stl10_250','stl10_1000','svhn_40','svhn_250','svhn_1000']
show_acc = ['BestAcc','Top1_1','Top1_20','Top1_50','Top5_1','Top5_20','Top5_50']
algs = ['pimodel','pseudolabel','pseudolabel_flex','meanteacher','vat','mixmatch','remixmatch','uda','uda_flex','fixmatch','flexmatch','fullysupervised']
workbook = xlwt.Workbook()
for i in range(len(show_acc)):
worksheet = workbook.add_sheet(show_acc[i],cell_overwrite_ok=True)
for j in range(len(algs)):
worksheet.write(j+1,0,algs[j])
for j in range(len(data_setting)):
for alg in algs:
worksheet.write(algs.index(alg)+1,j+1,'None')
for j in range(len(data_setting)):
d = data_setting[j]
worksheet.write(0,j+1,d)
for alg in algs:
for k,v in final_res.items():
if alg+'_'+d == k and show_acc[i] in v:
#if k.endswith('_'+d) and k.startswith(alg+'_') and show_acc[i] in v:
worksheet.write(algs.index(alg)+1,j+1,str(round(np.mean(v[show_acc[i]]),2))+u"\u00B1"+str(round(np.std(v[show_acc[i]]),2)))
workbook.save('../saved_models/final_res.xls')