Training on Trajnet++

[Mirorrn]

Hi I am about to try your code on the Trajnet++.

I am trying to get the FDE score of 1.14 .

Did you train on the whole (with cff) training dataset?
How many epochs?
And with which parameters?

If I reach your performance I will delete the submission if desired.

Thanks in advance
Many greetings

[YuejiangLIU]

Thanks for your interest in trying our method.

To get the FDE score of 1.14, I simply followed the instructions of Social LSTM baseline.

• I did not use the cff dataset, which seems very noisy.
• I trained the baseline for 25 epochs.

After obtaining a competitive baseline model, I finetuned it using our S-NCE for 5 epochs.

I plan to release the code, once our submission to Trajnet++ is no longer ranked top.

[Mirorrn]

Hi!

Thank you for the quick response!