gpmol dependency

[DavidWalz]

@beangoben First of all, thanks for providing model and data in form of a well formatted notebook. It's exciting to see that this good practice is spreading more and more across different fields!

However, I'm currently missing the gpmol dependency, which seems to contain the most integral parts of the analysis. Is it a group-internal python package? If so, would you consider open sourcing it or removing this dependency? With gpmol missing the analysis is neither reproducible, nor is gpHSP open-source as claimed in the paper.

[beangoben]

Hi @DavidWalz, We wrapping work on gpmol and should have open-sourced soon (end of June), the delay has been mostly life-stuff...we definitely want it to be shared, improved and reproduced.

[DavidWalz]

Great! Looking at the paper, it does shed a bad light on the reviewers who apparently didn't even bother checking out the code. I do appreciate your effort , and am looking to testing out your model.

[gbrault]

Thanks for releasing this work open source, and of course waiting for the gpmol dependency!

[beangoben]

definitely, will have a usable dependency in 1 week approx, along with an improved model (multioutput gp, deep gp). Thanks for the patience.

[napoleon9000]

@beangoben Thanks for sharing all the data. Just wondering if you still have the plan to share the gpHSP source code?

[rodrija]

Please inform when the gpmol code would be released. It is unfortunate that nothing has happened since June last year. I found your paper very useful and would like to test your code for solvent screening.

[ignaczgerg]

Hey! Is there any news regarding to make gpmol open source?

[beangoben]

Thanks for your interest, I'll aim to put the code up next Friday (~21)!

On Thu, Feb 13, 2020 at 11:24 AM ignaczgerg [email protected] wrote:

Hey! Is there any news regarding to make gpmol open source?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/aspuru-guzik-group/gpHSP/issues/1?email_source=notifications&email_token=AADK656OQBM3XKHF3KHNQVLRCV65XA5CNFSM4HQAGQW2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOELV3QHQ#issuecomment-585873438, or unsubscribe https://github.com/notifications/unsubscribe-auth/AADK653VPDKUM6GZFYIXKDLRCV65XANCNFSM4HQAGQWQ .

[ignaczgerg]

Many thanks! We are really looking forward to it :)

[sptiwari]

Any news about the gpmol?

[sptiwari]

Hi beangoben, We are really looking forward to run the gpHSP. Could you please upload whatever you have about the gpmol. It need not be perfect. We will be more than happy with just a running version of gpHSP.

Thanks a lot.

[ignaczgerg]

Hi! Is there any news regarding the release of gpmol? Thanks.

[xuzhang5788]

@beangoben When will you release gpmol?

[thava]

@beangoben Any insights into usage of gpmol or alternatives would be highly appreciated.

[sptiwari]

@beangoben Any update? We have been waiting for so long!

[hududed]

This is frustrating when we can't follow-up on a publication due to technicality - I didn't know academia is not fond of sharing knowledge. Could it be that gpmol is being withheld for IP reasons ?

[madi20210904]

Is there any news regarding the release of gpmol? Thanks.

[jmohin]

Still waiting on release of GPmol -- coming up on 4 years now. Would love to reproduce these results and do some extensions. Some other group's papers are claiming it as a dependency, which seems to indicate to me this library is shareable.

[beangoben]

Hi, the most recent edits (march '22) (notebooks 0 to 3) and gphsp.py cover all code required to generate features/train models and predict without requiring gpmol.

The closest thing to gpmol is https://github.com/aspuru-guzik-group/dionysus/.

[jmohin]

Thanks! I'll play with the notebooks.

[jmohin]

The notebooks fail to import gpHSP due to a dependency on gpFlow within the gphsp.py file: ModuleNotFoundError: No module named 'gpflow'

[jmohin]

nevermind- was able to satisfy that dependency