Calculating the CER

[mayang1]

hello, in order to Calculating the CER , i run ''python3 pytasas_words.py 1000 si'',the result as follow :

can you help me to solve the problem

thank you

[leitro]

Hi! Could you provide some samples in your ground truth file and the prediction file? You need to replace the urls of the ground truth and prediction with your own files in pytasas_words.py. Have a try!

[mayang1]

yes , it is right,but the result is above

[mayang1]

when i run python3 pytasas_words.py 40 no
i have the error:
htrsh.inc.sh: error: required run_parallel.inc.sh not found in path
htrsh.inc.sh: error: required run_parallel.inc.sh not found in path
./tasas_cer.sh: line 3: htrsh_prep_tasas: command not found
./tasas_cer.sh: line 3: htrsh_prep_tasas: command not found

can you help me , thank you very much

[mauvilsa]

I have modified htrsh.inc.sh so that it also works if run_parallel.inc.sh is in the same directory as htrsh.inc.sh but not in the path. @mayang1 please pull the changes and try again.

[mayang1]

I just tested the new code and I have successfully completed the WER calculation. Thank you for your sharing and help. Thank you again.

[mayang1]

valid_predict_serq.8.log

e04-068-00-03,171 incuuded
f01-053-06-06,162 wocabuetry
b03-109-05-07,191 -
b06-053-01-06,184 practice
c06-043-00-03,177 an
e04-062-07-09,166 one
r06-126-03-05,185 helped
c06-020-03-06,182 untertien

ground_truth.thresh

n02-000-00-00,177 DARKNESS
n02-000-00-01,177 had
n02-000-00-02,177 descended
n02-000-00-03,177 like
n02-000-00-04,177 a
I want to know if I have to consider the inconsistency of the data order when calculating WER.

[mayang1]

cer_valid .log

0.56144434 0.3129064 0.21466828 0.16654824999999998 0.13386255 0.11146773 0.09625288 0.0844089 0.07435356 0.06895859 0.060859079999999996 0.05581309 0.051558719999999995 0.0470168 0.04370005

but in you artical " Convolve, Attend and Spell: An Attention-based Sequence-to-Sequence Model for Handwritten Word Recognition." the best result Valid_CER is 5.01