cmhungsteve / TA3N

[ICCV 2019 (Oral)] Temporal Attentive Alignment for Large-Scale Video Domain Adaptation (PyTorch)
https://arxiv.org/abs/1907.12743
MIT License
259 stars 41 forks source link

Question about Frame Relation when use_target is set to "none" #52

Closed kerenartiaga closed 2 years ago

kerenartiaga commented 2 years ago

Hello Sir,

Sorry I forgot to ask this in my previous issue: When use_target is set to "none", does frame-relation still work?

Thank you

cmhungsteve commented 2 years ago

If you set "frame_aggregation=trn-m", it means that you are using "frame-relation". If you set "frame_aggregation=avgpool", it means that you're not using "frame-relation". "use_target" is not related.

kerenartiaga commented 2 years ago

I see. When using avgpool, do we align frames one by one instead of by relation?

cmhungsteve commented 2 years ago

No, since we are working on the classification task, we just evenly sample frames and average them to get the final output prediction. The frame indices between different videos are not aligned since that part is not critical for the classification task.