paperswithcode / galai

Model API for GALACTICA
Apache License 2.0
2.68k stars 276 forks source link

Preprocessing script for moleculenet / uniprot? #51

Open csinva opened 1 year ago

csinva commented 1 year ago

Hello,

I'd like to investigate / replicate the Galactica results on scientific domains. Is it possible to release the script used to preprocess the moleculenet/uniprot data? I'm unable to get Galactica to meaningfully answer queries about this data, likely due to my incorrect formatting of the datasets.

Thank you!

RJT1990 commented 1 year ago

thanks for raising! will prepare and get this to you

csinva commented 1 year ago

Hi again! Just checking if there were any updates?

chao1224 commented 1 year ago

Hi @RJT1990,

I have the same question here. We are interested in the text prompts you extracted from MoleculeNet.