fix facenet.get_learning_rate_from_file()

[BaeMinCheon]

I expected when learning rate becomes -1, the train is closed
but get_learning_rate_from_file() returns nonetype whenever learning rate becomes -1
( other users have this problem, too : #853 )
it is the problem, and I modified the function so that solve the problem

images below show the different results

before:

after:

[hsm4703]

回溯（最近一次調用最後一次）：
文件“src / train_softmax.py”，第580行，在
main（parse_arguments（sys.argv [1：]））
文件“src / train_softmax.py”，第234行，在主要的
prelogits中， prelogits_center_loss，args.random_rotate，args.random_crop，args.random_flip，prelogits_norm，args.prelogits_hist_max，args.use_fixed_image_standardization）
文件“SRC / train_softmax.py”，線306，在列車
LR = facenet.get_learning_rate_from_file（learning_rate_schedule_file，歷元）
文件“ C：\ facenet-master \ src \ facenet.py“，第295行，在get_learning_rate_from_file中
e = int（par [0]）
ValueError：對於帶有基數為10的int（）的無效文字：''
e = int(par[0]) this way how can i fix
i change code and use your code but i have this problem can u fix it ?

[BaeMinCheon]

can i get the error message written with english ? i can hardly know about the error

[hsm4703]

Backtracking (last call last time):
The file "src / train_softmax.py", line 580, at
Master (parse_arguments(sys.argv [1:]))
File "src / train_softmax.py", line 234, in the main
In prelogits, prelogits_center_loss, args.random_rotate, args.random_crop, args.random_flip, prelogits_norm, args.prelogits_hist_max, args.use_fixed_image_standardization)
File "SRC / train_softmax.py", line 306, on the train
LR = facenet.get_learning_rate_from_file(learning_rate_schedule_file , epoch)
The file "C:\facenet-master \ src \ facenet.py", line 295, in get_learning_rate_from_file
e = int(par [0])
ValueError: Invalid text for int() with a base of 10: ''
e = int(par [0]) so how can I fix it?
I changed the code and used your code but can I fix it with this problem?
i am taiwanese sry

[BaeMinCheon]

you should check out a value that the par[0] has when the error occurs. use print() or watch for debugging it.

it seems that the error occurs when the par[0] is not in the format of integer (ex: '' , '123.456' ...etc)

if the par[0] contains the value not in the format of integer, it would be good to check out the learning rate file. in any case, your learning rate file may have contained unnecessary lines or characters.

p.s. you do not have to be sorry about the language. but it would be better sharing codes in english especially on github for saving a conversation cost.

[hsm4703]

you should check out a value that the par[0] has when the error occurs. use print() or watch for debugging it.
e = print(int(par[0])) like it or how do i change code?

[hsm4703]

Learning rate schedule

Maps an epoch number to a learning rate 0: 0.05 60: 0.005 80: 0.0005 91: -1

my learning rate classifier_casia file may have contained unnecessary lines or characters?

[BaeMinCheon]

create a repository with your code on github and let me know the url of the repository. then, I can check out what is the matter.

[hsm4703]

create a repository with your code on github and let me know the url of the repository. then, I can check out what is the matter.

https://github.com/hsm4703/facenet/blob/master/softmax.py

[BaeMinCheon]

I have reproduced your error. the error seems to happen because of the learning rate file. look at the screenshots below.

#1 : if it does not contain any unnecessary line

#2 : if it contains one unnecessary line

#3 : if it contains two unnecessary lines

therefore, you should check out whether your learning rate file contains more than one unnecessary line. and I made PR about it, so you can do this for ease, too.

[hsm4703]

我已經復制了你的錯誤。由於學習率文件，似乎發生了錯誤。看下面的截圖。

＃1：如果它不包含任何不必要的行

＃2：如果它包含一條不必要的行

＃3：如果它包含兩個不必要的行

因此，您應該檢查您的學習率文件是否包含多個不必要的行。我做了關於它的PR，所以你也可以輕鬆做到這一點。

thank you begin i use note so i don't see contains more than one unnecessary line. but i use sublime or your notepad++ i find have three unnecessary line