search

LisaAnne / LocalizingMoments

Github for my ICCV 2017 paper: "Localizing Moments in Video with Natural Language"
188 stars 44 forks source link

about the time points #2

Closed mengliu1991 closed 6 years ago

mengliu1991 commented 6 years ago

Hi, In the .json file, each video has more than one time points. For example, "video": "26292851@N04_4253489686_265c3c8051.m4v" has 7 time points: "times": [[4, 4], [4, 4], [0, 0], [4, 4], [0, 0], [0, 0], [4, 4]]. I wanna to know which one you choose as ground truth when training your model? And how to evaluate the testing result with these time points?

LisaAnne commented 6 years ago

Hi there -

There is eval code in the repo: https://github.com/LisaAnne/LocalizingMoments/blob/master/eval.py.

I randomly chose a time point of the provided time points. Usually videos have four time points; this has seven because it was put on AMT twice because my batches need to be divisible by 4 (there are a few examples like that).