search

HarshayuGirase / Human-Path-Prediction

State-of-the-art methods for human trajectory forecasting. Contains code for papers published at ECCV 2020 and ICCV 2021.
MIT License
341 stars 81 forks source link

ETH-UCY result without TTST #26

Closed Gutianpei closed 2 years ago

Gutianpei commented 2 years ago

Hello,

Could you provide the ETH-UCY result without TTST trick? I guess the result reported in your YNet paper is with TTST, and I wonder how TTST works against ETH-UCY dataset. Thank you very much!

karttikeya commented 2 years ago

Hi, unfortunately we don’t have results without TTST for ETH/UCY since our focus was on ablations on long-term settings on SDD/InD.

If we run this setting in future, we will make sure to share the results here.

Gutianpei commented 2 years ago

Hello,

Thanks for the reply! I've checked the code and it looks like testing without TTST can be simply performed by setting use_ttst parameter to false. YNet has superior performance on ETH/UCY dataset, and I wonder how much does the trick part contributes. Or is that possible to release all the train/test data you used or the pretrain models on ETH/UCY?

Gutianpei commented 2 years ago

Or maybe you have the testing result without TTST on SDD dataset with np=8, nf=12 setting?