Skip to content

Releases: vita-epfl/bounding-box-prediction

Multi-Task PV-LSTM Trained

17 Nov 07:09
Compare
Choose a tag to compare

Multi-Task PV-LSTM trained on JAAD dataset using the following setting:

  • Input sequence length = 16 frames (0.53 sec)
  • Output sequence length = 16 frames (0.53 sec)
  • Batch size = 100
  • number of epochs = 100