/
vimak.py
62 lines (51 loc) · 3.54 KB
/
vimak.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
# 'newOneToOneOhem_lr-beg=1.3e-03_lr-sch=1e-03_shuffle=False',
# 'newOneToOneOhem_lr-beg=1.3e-03_lr-sch=None_shuffle=False',
from torch.utils.tensorboard import SummaryWriter
from toolbox.pickleOpers import loadup
ls=['5XOhemWithShuffleNoScheduler_epoch_36',
'new_1xOhem_shuffle_true_scheduler_e2_epoch_22',
'new_5xOhem_shuffle_true_scheduler_e2_epoch_10',
# 'newOneToOneOhem_lr-beg=1.3e-03_lr-sch=1e-03_shuffle=False_epoch_40',
# 'newOneToOneOhem_lr-beg=1.3e-03_lr-sch=None_shuffle=False_epoch_22',
'1XOhemWithShuffleNoScheduler_lr-beg=1.3e-04_lr-sch=None_shuffle=True_epoch_40',
'newOhemTrainSingleSampling32_lr-beg=1.3e-03_lr-sch=None_shuffle=True_epoch_1',
'newOhemTrainSingleSampling32_lr-beg=1.3e-04_lr-sch=None_shuffle=True_epoch_4']
# 'SingleSamplingOhemAdamLRe3_epoch_32',
# 'Resnet50_Final']
# ls=['5XOhemWithShuffleNoScheduler',
# 'new_1xOhem_shuffle_true_scheduler_e2',
# 'new_5xOhem_shuffle_true_scheduler_e2',
# '1XOhemWithShuffleNoScheduler_lr-beg=1.3e-04_lr-sch=None_shuffle=True',
# 'newOhemTrainSingleSampling32_lr-beg=1.3e-03_lr-sch=None_shuffle=True',
# 'newOhemTrainSingleSampling32_lr-beg=1.3e-04_lr-sch=None_shuffle=True','Renet50_Final','SingleSamplingOhemAdamLRe3_epoch_32']
writer=SummaryWriter("prData7")
for l in ls:
a=loadup("evalData/{}_inferConf=0.7/prData/prCurve_val_epoch_{}.pickle".format(l.split("_epoch_")[0],l.split("_epoch_")[1]))
for i in range(len(a)):
writer.add_scalars("curve/",{l.split("_epoch_")[0]:a[i][0]},a[i][1]*1000)
ls=['SingleSamplingOhemAdamLRe3_epoch_32',
'Resnet50_Final']
for l in ls:
a=loadup("evalData/{}_inferConf=0.7/prData/prCurve_val.pickle".format(l))
for i in range(len(a)):
writer.add_scalars("curve/",{l:a[i][0]},a[i][1]*1000)
writer.close()
# # # ls=[ 'newOhemTrainSingleSampling32_lr-beg=1.3e-03_lr-sch=None_shuffle=True_epoch_1', 'newOhemTrainSingleSampling32_lr-beg=1.3e-04_lr-sch=None_shuffle=True_epoch_4']
# ls=[ 'newOneToOneOhem_lr-beg=1.3e-03_lr-sch=1e-03_shuffle=False_epoch_40', 'newOneToOneOhem_lr-beg=1.3e-03_lr-sch=None_shuffle=False_epoch_22']
# import os
# # os.system(f'python evaluate.py --trained_model "SingleSamplingOhemAdamLRe3_epoch_32" --confidence_threshold_infer 0.7')
# # os.system(f'python evaluate.py --trained_model "Resnet50_Final" --confidence_threshold_infer 0.7')
# for l in ls:
# print(l)
# os.system(f'python evaluate.py --trained_model "{l.split("_epoch_")[0]}" --mode "series" --confidence_threshold_infer 0.7')
# for l in ls:
# print(l)
# # os.system(f'python detect.py --trained_model "{l}" --save_name "{l}" --vis_thres 0.055 --mode "images"')
# os.system(f'python detect.py --trained_model "{l}" --save_name "{l}" --vis_thres 0.7 --mode "images"')
# # os.system(f'python detect.py --trained_model "{l}" --save_name "{l}" --vis_thres 0.43 --mode "images"')
# # os.system(f'python detect.py --trained_model "{l}" --save_name "{l}" --vis_thres 0.055 --mode "videos" --fps 1')
# # os.system(f'python detect.py --trained_model "{l}" --save_name "{l}" --vis_thres 0.055 --mode "videos" --fps 5')
# # os.system(f'python detect.py --trained_model "{l}" --save_name "{l}" --vis_thres 0.43 --mode "videos" --fps 5')
# os.system(f'python detect.py --trained_model "{l}" --save_name "{l}" --vis_thres 0.7 --mode "videos" --fps 5')
# os.system(f'python detect.py --trained_model "{l}" --save_name "{l}" --vis_thres 0.7 --mode "videos" --fps 1')
# # os.system(f'python detect.py --trained_model "{l}" --save_name "{l}" --vis_thres 0.43 --mode "videos" --fps 1')