Error : Segmentation fault (core dumped)

[BlankSama]

Hi,

while running the generation script (python script/sample.py -g ./test/generation_config/tpsa.yaml -s "c1ccccc1" --num_samples 100 --logp 6 -o ./result_sample/logp\=6.smi) it is showing Segmentation fault (core dumped)

Iam attaching a screenshot of the same, please help

Thanks, and regards.

[SeonghwanSeo]

Oh there is a mistake on README.md. I will modify that bug as soon as possible. Thank you for pointing out the error.

The error seems to be cause by mismatch between generation configure file and argument. Try modifying ./test/generation_config/tpsa.yaml to ./test/generation_config/logp.yaml

python script/sample.py -g ./test/generation_config/logp.yaml -s "c1ccccc1" --num_samples 100 --logp 6 -o ./result_sample/logp=6.smi

[BlankSama]

Hi @SeonghwanSeo ,

Thanks for replying, I tried the updated command but still get the same error and a new error Iam attaching the screenshot of the same.

Thanks, and regards!

[SeonghwanSeo]

I haven't found any issues on CentOS7 and MacOS (Apple Silicon M1). Can you let me know what your working environment is?

[BlankSama]

Hi,

Iam working on ubuntu 20.04 with cuda 11.6 Let me know if you need more details

Get Outlook for Androidhttps://aka.ms/AAb9ysg

---

From: Seonghwan Seo @.> Sent: Sunday, June 11, 2023 5:46:33 PM To: SeonghwanSeo/BBAR @.> Cc: neeraj mehra @.>; Author @.> Subject: Re: [SeonghwanSeo/BBAR] Error : Segmentation fault (core dumped) (Issue #3)

I haven't found any issues on CentOS7 and MacOS (Apple Silicon M1). Can you let me know what your working environment is?

— Reply to this email directly, view it on GitHubhttps://github.com/SeonghwanSeo/BBAR/issues/3#issuecomment-1586141666, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AJVAOPUM7YW5KPAQRPGXHVDXKWZKDANCNFSM6AAAAAAZAGMIGQ. You are receiving this because you authored the thread.Message ID: @.***>

[SeonghwanSeo]

As far as I know, there was importing issue for torch geometric. (I also encountered that issue in MacOS.) Could you confirm that torch geometric can be imported well in your environment?

$ python
>>> import torch
>>> import torch_geometric

[BlankSama]

Hi, @SeonghwanSeo

I also encountered the torch geometric issue but I fixed it by downloading through different channel.

iam attaching the screenshot of the downloaded version.

and I am now able to import torch_geometric without any errors.

please let me know if you are able to replicate the issue.

thanks and regards

[BlankSama]

Hi @SeonghwanSeo ,

The system is working now, I have one query to ask. is it possible to combine logp and molecular weight to optimize simultaneously. because in the example generation command --generator_config can point to only one prebuilt model.

please let me know if there is a possibility to combine mw and logp.

Thanks and regards Neeraj

[SeonghwanSeo]

I'm glad to see the system is working.

I will train the model for MW-logP and deliver the trained model until this week.

[BlankSama]

@SeonghwanSeo thank you so much for the help.

will wait for the trained model.

[SeonghwanSeo]

The mw-logp model is uploaded. You can download the model weight with following script.

gdown 'https://drive.google.com/uc?id=10UXskmoQQW6KZAeNSVqF4LliMZB9j5Iq'

Also, required config file (./test/generation_config/mw_logp.yaml) is uploaded on this repository.

[Healther886]

Hi, I'm curious about the source of your prebuilt data. I understand your data comes from the Zinc Database, which has various data types available for download, such as fragments and drug-like compounds. I'd like to know which specific data type you use and why you perform docking tasks in this program.

Your prebuilt model consists of a PDP 7l13 file and a Zinc file. The data in these files seem nearly identical, with the only difference being that the data in the Zinc file is slightly larger. Could you explain why there is this difference?

The molecular weight would be significant and complex if you downloaded drug-like data. I'm interested in understanding the rationale behind using the module Brics of Rdikt to break down the molecules into smaller ones based on specific criteria.

Furthermore, I'm curious about the type of molecules you downloaded from the Zinc database. Are you explicitly searching for the target protein 3cl pro in the Zinc database?