What's the meaning of 'python run_eval.py --hypos output_file --refs reference_file'?

[SongQiongcheng]

How to run it on the right output_file and reference_file?Can you give a example?
I try to replace output_file to my candidatereference,and replace output_file to my reference,but it doesn't work.
Thank you.

[GrittyChen]

How to run it on the right output_file and reference_file?Can you give a example?
I try to replace output_file to my candidatereference,and replace output_file to my reference,but it doesn't work.
Thank you.

@SongQiongcheng I'm sorry for the late reply. The output_file is the file that contains all outputs of your system, and its format is one example per line. The referece_file is the file that contains the candidate reference, and its format is the same as the output_file. Note that if you have more than one candidate reference for each example, you need to split these candidate references into multiple reference_file(s). For example, if you have two candidate references for each example, you will split them into two reference_file(s), the reference_file1 contains all the first candidate references and the reference_file2 contains all the second candidate references. We give an example of output_file and reference_file as follows:
output_file:
我 喜欢 苹果 。
你 早上 吃了 一个 香蕉 。
reference_file:
I like apples .
You ate a banana in the morning .

[SongQiongcheng]

How to run it on the right output_file and reference_file?Can you give a example?
I try to replace output_file to my candidatereference,and replace output_file to my reference,but it doesn't work.
Thank you.

@SongQiongcheng I'm sorry for the late reply. The output_file is the file that contains all outputs of your system, and its format is one example per line. The referece_file is the file that contains the candidate reference, and its format is the same as the output_file. Note that if you have more than one candidate reference for each example, you need to split these candidate references into multiple reference_file(s). For example, if you have two candidate references for each example, you will split them into two reference_file(s), the reference_file1 contains all the first candidate references and the reference_file2 contains all the second candidate references. We give an example of output_file and reference_file as follows:
output_file:
我 喜欢 苹果 。
你 早上 吃了 一个 香蕉 。
reference_file:
I like apples .
You ate a banana in the morning .

Thank you so much.I got it .