MichiganCOG / Video-Grounding-from-Text

Source code for "Weakly-Supervised Video Object Grounding from Text by Loss Weighting and Object Interaction"
Other
44 stars 9 forks source link

Could you release the sentence anotation of testing set? #5

Closed Xun-Yang closed 5 years ago

Xun-Yang commented 5 years ago

Hi I know the bbx annotations of testing set is only for online evaluation. However, the sentence annotations of testing set do not affect the online evaluation. I didn't find them in the released testing split.

Could you release the sentence annotations of testing set? They are useful in our model.

Thanks.

LuoweiZhou commented 5 years ago

@Xun-Yang we do not plan to release the testing sentences/segments for dense video captioning evaluation purposes. We'd suggest you report the grounding results on the validation set for now and email me (luozhou@umich.edu) and Nathan (natlouis@umich.edu) with your model/docker/instructions on how to run your code so we can run it on our test set.