search

hulianyuyy / CorrNet

Continuous Sign Language Recognition with Correlation Network (CVPR 2023)
84 stars 14 forks source link

Fixed Frames #24

Closed atonyo11 closed 7 months ago

atonyo11 commented 7 months ago

Sorry if I misunderstood. Usually, we need to set a constant for number of frame in a video when fitting through network. I can't find how you set the number of frames for a video to training (some thing like MAX_SEQ_LENGTH or Fixed_Frames). I can see the class TemporalRescale(object) in utils, but I think it will create clips of different lengths. Can you tell me how did you do?

Thanks!

hulianyuyy commented 7 months ago

In CorrNet, we don't set a constant number for frames in a video, but send all frames into the model. Videos in a batch will be padded into the maximum length of videos in the same batch.

atonyo11 commented 7 months ago

I got it. Thank you very much!