Hanjun-Dai / GLN

Implementation of Retrosynthesis Prediction with Conditional Graph Logic Network

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

USPTO-Full Csv file

vsomnath opened this issue · comments

Hi,

Would you have the processed csv file for the USPTO-full dataset, with duplicates and incorrectly atom-mapped examples removed?

It can be accessed via the dropbox link as well.

Hello, thank you for sharing you work!

Is there any chance you could post the cooked version of uspto-full? For some reason I am unable to cook it myself, the script on step 4 (step4_run_find_all_reactions.sh) shows an estimated time of over 200 hours (for 40 cores) and ends up draining RAM (had to kill it before it killed my server).

Sorry the cooked file is also too large for me to upload. I used cluster nodes to process (and I no longer have access to that since my graduation). Sorry about that.

AWS or google cloud has some free trail credits, maybe you can leverage that for computing.

That's unfortunate :(
Thanks for your answer anyway!