I have parsed all the bridge-train.tfrecord-xxx-of-01024 files from the bridge dataset, but I only obtained 25,460 trajectories. However, the bridge v2 mentions that there are a total of 60,000 trajectories. Can you please explain the reason for this significant difference in the number of trajectories?
I have parsed all the bridge-train.tfrecord-xxx-of-01024 files from the bridge dataset, but I only obtained 25,460 trajectories. However, the bridge v2 mentions that there are a total of 60,000 trajectories. Can you please explain the reason for this significant difference in the number of trajectories?