calculate the evaluation index layer of CP/CR/CF1/OP/OR/OF1 of top3

[mymuli]

If I want to calculate the evaluation index layer of CP/CR/CF1/OP/OR/OF1 of top3,

multi-label-image-classification/MSCOCO/evaluation_metrics.py

Line 39 in 1d9ef29

def prf_cal(y_pred,y_true,k):

what is the value of the function parameter?
Y_pred is a real number between 0 and 1?
Y_true values between {0,1}? Or something else?

[hellowangqian]

Y_pred do not have to be in the range of [0,1], you only need the ranking of different labels to be correct regardless of the absolute values. Y_true is 0 ( negative) or 1 (positive) for nus and mscoco datasets, and -1 (negative), 0 (ambiguous), 1 (positive) for voc2007 dataset, that's why you need to calculate these metrics in a slightly different way.

…

________________________________ From: Rmuli <notifications@github.com> Sent: 27 September 2019 1:06 To: hellowangqian/multi-label-image-classification <multi-label-image-classification@noreply.github.com> Cc: Subscribed <subscribed@noreply.github.com> Subject: [hellowangqian/multi-label-image-classification] calculate the evaluation index layer of CP/CR/CF1/OP/OR/OF1 of top3 (#2) If I want to calculate the evaluation index layer of CP/CR/CF1/OP/OR/OF1 of top3, https://github.com/hellowangqian/multi-label-image-classification/blob/1d9ef29d798f71e3b8d4c84c026017b6c1ce2695/MSCOCO/evaluation_metrics.py#L39 what is the value of the function parameter? Y_pred is a real number between 0 and 1? Y_true values between {0,1}? Or something else? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<#2?email_source=notifications&email_token=ABX467SYSDTRELBKLYJAHQLQLVMCFA5CNFSM4I3ATIL2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HOA5LEA>, or mute the thread<https://github.com/notifications/unsubscribe-auth/ABX467RSAAHS2XBTEJYEEATQLVMCFANCNFSM4I3ATILQ>.