Question about test results file

[ivyvideo]

Hi, I test the caffemodel and the results are saved to log file. But I have no idea what is the matrix below?
I notice that in test.py cls_win and cls_score are stacked, however, there are many spaces between two floating points. Could you please explain it?
Thanks for your kindness!

activity: 12
[[ 43.073074 183.75291 0.9771269 ]
[242.433 376.31635 0.7799468 ]
[346.45883 455.46292 0.67514086]
[118.58941 263.16885 0.60216105]]

[huijuan88]

[start frame, end frame, score] in the buffer input. You need to use the log analysis code to change the coordinates back to the original video input. On Aug 30, 2018, at 04:19, ivyvideo <[email protected]<mailto:[email protected]>> wrote: Hi, I test the caffemodel and the results are saved to log file. But I have no idea what is the matrix below? I notice that in test.py cls_win and cls_score are stacked, however, there are many spaces between two floating points. Could you please explain it? Thanks for your kindness! activity: 12 [[ 43.073074 183.75291 0.9771269 ] [242.433 376.31635 0.7799468 ] [346.45883 455.46292 0.67514086] [118.58941 263.16885 0.60216105]] — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<#40>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AFOa_9ME77EWPJbUkoH4WTaW4xLyaHWAks5uV5_3gaJpZM4WS_s3>.

[ivyvideo]

Got it...And can you please tell me what do the different rows mean...?

[huijuan88]

They are multiple predictions sorted by scores. On Aug 30, 2018, at 18:32, ivyvideo <[email protected]<mailto:[email protected]>> wrote: Got it...And can you please tell me what do the different rows mean...? — You are receiving this because you commented. Reply to this email directly, view it on GitHub<#40 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AFOa_2aKMAzUW3ZG7izWjDc1x1XvcgG1ks5uWJJDgaJpZM4WS_s3>.

[ivyvideo]

So there ought to be only one row of every prediction because of usage of NMS, do multi-rows indicate the model is not trained very well??