search

meaten / keepflow

MIT License
3 stars 0 forks source link

Pretrained Models for MID? #1

Open mh-kav-institute opened 3 months ago

mh-kav-institute commented 3 months ago

Hi guys, thank you for your great work. As one of the few, you seem to have been able to successfully train models for MID with the ETH dataset to perform your comparison studies. Could you share it because several people and myself have problems to retrain MID successfully with the available code and configuration. The authors promised to fix it but never did (9 months ago, nothing happened).

Thank you very much, keep up the great work you are doing.

meaten commented 3 months ago

Hi, I agree that all recent trajectory prediction methods should be reproducible. Some researchers raised an issue about the issue on MID data preprocessing. This should be the cause of failed reproduction. InhwanBae, who raised the issue on MID repo, provided the comparison studies including MID. You should use these MID results in your paper. This repo also includes MID training code without the issue above, thus you can try to train MID;)

mh-kav-institute commented 3 months ago

Hi, thanks for your response. Good to know that some has a copy of the working code, I will try it!