--------------------modification 2022.11.12------------------
I found there are annotation files in https://scenario-oriented-object-navigation.github.io . Still, I can't find val_seen_instrs dataset. According to the paper, it seems that the instructions are same with the train split but have different initial locations.
It would be very helpful if you share the datasets and also your trained checkpoint (if possible!).
Hi, I'm trying to reproduce your paper's result but I cannot find any SOON dataset.
I found that there are val/test datasets in https://github.com/ZhuFengdaaa/SOON-EvalAI-Starters/tree/challenge/annotations , but where is the training dataset?
--------------------modification 2022.11.12------------------ I found there are annotation files in https://scenario-oriented-object-navigation.github.io . Still, I can't find val_seen_instrs dataset. According to the paper, it seems that the instructions are same with the train split but have different initial locations.
It would be very helpful if you share the datasets and also your trained checkpoint (if possible!).
Thanks, Minyoung.