The script to reproduce the results

[SiyuanLee]

Hi, thank you for open-sourcing the code. There is a problem about the script to reproduce the code. The link in the README file has expired. Could you please upload it again? Thx.

[qxcv]

Thanks for alerting us to this. I've been cleaning up the code, and I think I broke the link by moving some files around. I'll fix the instructions and reply here when finished (most likely tomorrow or the day after).

[qxcv]

I've uploaded a new demo file (same actual trajectories but compressed to be 10x smaller) and updated everything to use Docker. The docs are updated too:

https://il-representations.readthedocs.io/en/latest/reproduction_scripts.html

(you may need to ctrl+F5 to load a fresh copy of the page)

I tried building and running the docker file on a fresh git clone & it worked okay, but I didn't try running all the experiment commands to completion with this setup (we use Kubernetes rather than running in Docker directly). Let me know if something is broken!

[SiyuanLee]

Thx, I will have a try :)

[SiyuanLee]

I guess that the newly uploaded dataset is incomplete. When I unzip the file, there is an error "gzip: stdin: unexpected end of file". I have re-downloaded the file several times, but the error still exists. Could you please have a check? Thanks :)

[qxcv]

Could you please try re-downloading again? It looks like the original file on Google Drive was a bit smaller than it should have been, so I re-uploaded it. I also uploaded a checksum: you can check whether eirli-demos.tar.gz is complete by running md5sum eirli-demos.tar.gz and checking whether the hash matches the one in eirli-demos-md5sum.txt (e5a55a4753eb24c21c0a7dc392f25840).