Could you share training code?

[WinstonDeng]

@os1a
Hi, thanks for your great work.
Could you share your training code or training guidance?

[os1a]

Hi @WinstonDeng ,

Thanks for your interest in our work.

Unfortunately, currently we cannot share the source code for training since it contains parts which cannot be shared to the public. However, we are working on a pyTorch implementation to enable people to train our framework. This will take few more months to finalize.

I will keep this issue open and will post an update as soon as we have the implementation ready.

[WinstonDeng]

@os1a
Thank you! Looking forward to your code.
Great respect for your work again!

[Senwang98]

@os1a
Amazing work! Looking forward to your code!
By the way, did you reproduct STED from the scratch?
Last year I tried to reproduct STED by Yolov4 and recently I found my pytorch-version of STED maybe wrong.
So, is it convenient to open source code of your version of STED?
Wishing for your reply!

[os1a]

@1187697147

The source code of STED is available at:
https://github.com/olly-styles/Multiple-Object-Forecasting

In our results, we implement the STED following their repository and run it on the reported datasets. We did not investigate a lot of time tuning the baseline since the method is deterministic where ours is multimodal.

[Senwang98]

@os1a
Thanks for your reply!

[centiLinda]

Hi @os1a , do you still plan to share your PyTorch implementation?