Entirely un-replicable.

[dmolitor]

Just dropping an FYI for the authors/slash any innocent soul that is potentially attempting to apply this methodology to new data. Not only is it impossible to replicate the authors' results as instructed in the README, it is even more impossible to generate results on new data, as the pipeline depends on files that ought to be created but are not, among numerous other bugs that I won't even attempt to document here. This seems like a very cool paper, so it's a bummer that it's currently in the state it is. Hope it gets renovated at some point in the future!

[ShuaiQi2025]

Dude, I think you're right

[MassimilianoLuca]

@ShuaiQi2025 I am afraid you closed your previous issue without listing a set of errors we can work on.

@ShuaiQi2025 @dmolitor any chance you can provide some details feedback? We are working on renewing the repo so any feedback is appreciated

[ShuaiQi2025]

I am sorry that I raised the wrong question due to my personal reasons. I will delete my question. I am sorry to you. Thank you for your contribution. What I want to tell you is that your torch1.7.1+python3.7 case will always cause DLL missing error. I switched to torch1.5.0+python3.8 to fix the problem. At the same time, I want you to explain that scikit-mobility 1.1.0 'no longer exists and cannot be followed, please update the README file. It is also best to state the python version. Too little explanation of the configuration environment can make it difficult for beginners to learn your code. Finally, thank you again for your contribution, I wish you have more and more achievements in the future, thank you. ------------------ 原始邮件 ------------------ 发件人: "scikit-mobility/DeepGravity" ***@***.***>; 发送时间: 2025年1月9日(星期四) 晚上8:21 ***@***.***>; ***@***.******@***.***>; 主题: Re: [scikit-mobility/DeepGravity] Entirely un-replicable. (Issue #7) @ShuaiQi2025 I am afraid you closed your previous issue without listing a set of errors we can work on. @ShuaiQi2025 @dmolitor any chance you can provide some details feedback? We are working on renewing the repo so any feedback is appreciated — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[MassimilianoLuca]

Thanks @ShuaiQi2025! No worries about that! All the comments are appreciated and very helpful to everyone!

We will work on the README and we will specify the Python version and dependencies.

I will keep you posted on progress here in this thread and please, drop us an email ([email protected]) if you need help or assistance to run or interpret the code!

Thanks again!