search

chao1224 / MoleculeSTM

Multi-modal Molecule Structure-text Model for Text-based Editing and Retrieval, Nat Mach Intell 2023 (https://www.nature.com/articles/s42256-023-00759-6)
https://chao1224.github.io/MoleculeSTM
Other
201 stars 18 forks source link

Missing downsload_MegaBolBART.sh #2

Closed ZwormZ closed 1 year ago

ZwormZ commented 1 year ago

Hello, thank you for sharing this fantastic work.

I noticed that during the pretraining stage, as outlined in the paper, the SMILES encoder uses MegaMolBART's encoder. However, I could not find the downsload_MegaBolBART.sh file in this repository. Additionally, I checked the paper you cited in MoleculesSTM, which is "Chemformer: a pre-trained transformer for computational chemistry," but it seems that they did not provide the model checkpoint.

Could you please advise me on how I can obtain the pretrained checkpoint for the SMILES encoder that uses MegaMolBART's encoder?

chao1224 commented 1 year ago

Hi @ZwormZ,

You may now refer to this script. There are five folders after unzipping, and megamolbart is the one we used here.

(I am checking how to add the scripts to this repo.)

ZwormZ commented 1 year ago

Great! I downloaded this checkpoint using the script you provided and it worked. Thank you!

chao1224 commented 1 year ago

Glad it helps.