update README to reference download of dataset over BitTorrent

[carandraug]

As discussed over email. Please review.

Note that I do include the checksums in the same table since if a BitTorrent client already checks them automatically.

Since the torrent also includes the files currently in http://www.robots.ox.ac.uk/~ankush/data.tar.gz it should be possible to modify the script to take the path for the downloaded dataset instead of downloading things (but note that data.tar.gz has sample.h5 while in the torrent it was renamed sample.h5).

[ankush-me]

As discussed over email. Please review.

Note that I do include the checksums in the same table since if a BitTorrent client already checks them automatically.

Since the torrent also includes the files currently in http://www.robots.ox.ac.uk/~ankush/data.tar.gz it should be possible to modify the script to take the path for the downloaded dataset instead of downloading things (but note that data.tar.gz has sample.h5 while in the torrent it was renamed sample.h5).

Is it possible re-tar-ball the data on the https server to have consistent names for the .h5 files? This would be most convenient downstream. I will edit the script to optionally take in the path to the downloaded renderer-data as well.

[ankush-me]

Suggested minor changes for typos etc.

[ankush-me]

Thank you very much @carandraug for maintaining the dataset, and this PR!

[carandraug]

Is it possible re-tar-ball the data on the https server to have consistent names for the .h5 files? This would be most convenient downstream. I will edit the script to optionally take in the path to the downloaded renderer-data as well.

[...]
Since the torrent also includes the files currently in http://www.robots.ox.ac.uk/~ankush/data.tar.gz it should be possible to modify the script to take the path for the downloaded dataset instead of downloading things (but note that data.tar.gz has sample.h5 while in the torrent it was renamed sample.h5).

Is it possible re-tar-ball the data on the https server to have consistent names for the .h5 files? This would be most convenient downstream. I will edit the script to optionally take in the path to the downloaded renderer-data as well.

It's under your account and I do not have control over it. But I can ask. Also, since the data.tar.gz is not actually a gzipped, maybe also just rename it data.tar or gzip it?

[ankush-me]

Is it possible re-tar-ball the data on the https server to have consistent names for the .h5 files? This would be most convenient downstream. I will edit the script to optionally take in the path to the downloaded renderer-data as well.

[...]
Since the torrent also includes the files currently in http://www.robots.ox.ac.uk/~ankush/data.tar.gz it should be possible to modify the script to take the path for the downloaded dataset instead of downloading things (but note that data.tar.gz has sample.h5 while in the torrent it was renamed sample.h5).

Is it possible re-tar-ball the data on the https server to have consistent names for the .h5 files? This would be most convenient downstream. I will edit the script to optionally take in the path to the downloaded renderer-data as well.

It's under your account and I do not have control over it. But I can ask. Also, since the data.tar.gz is not actually a gzipped, maybe also just rename it data.tar or gzip it?

Thanks -- yes I have updated it from my account, and also zipped it.

[ankush-me]

All done, thanks again @carandraug !