search

cunjunyu / STAR

[ECCV 2020] Code for "Spatio-Temporal Graph Transformer Networks for Pedestrian Trajectory Prediction"
MIT License
338 stars 78 forks source link

skip=6 #7

Closed ycl204432 closed 3 years ago

ycl204432 commented 3 years ago

Hello, the author,

Thank you very much for sharing the code of this project. There is a problem I don't understand. I'd like to ask you. As for why in eth scenario, when testing the performance of ETH scenario, skip = 6. In this way, there is no way to use the mode of 8-step predict 12 step to realize the experiment of 3.2s predict 4.8s . It should take 20 steps to predict the position of 4.8s. I'd like to hear your opinion。

cunjunyu commented 3 years ago

We follow the data preprocessing from sr-lstm and the reason is specified in this paper

Thank you

ycl204432 commented 3 years ago

Thank you very much for your reply. I've learned about it in my thesis.

Thank you very much

cunjunyu commented 3 years ago

I will close the issue for now, please feel free to reopen it if you have any further question.