search

GLambard / SMILES-X

Autonomous characterization of molecular compounds from small datasets without descriptors
MIT License
43 stars 13 forks source link

Model is bigger than in the example #7

Closed pnavaro closed 4 years ago

pnavaro commented 4 years ago

Dear Guillaume,

Thank you very much for sharing this. I try to run the notebook and I have some issues with the size of the problem. It is much bigger than yours and I use the same parameters. do you have an idea of what happened ?

My executed notebook https://github.com/pnavaro/SMILES-X/blob/master/SMILESX_Prediction_github.ipynb

***Bayesian Optimization of the SMILESX's architecture.***

Random initialization:

Model: [[512.  512.   32.  128.    3.9]]

Best regards Pierre

GLambard commented 4 years ago

Dear Pierre,

Thank you for your interest in the SMILES-X.

Also, what do you mean by the size of the problem?

I have checked your notebook, and I don't see any particular issue apart from the one possibly coming from a versioning problem that can be ignored so far.

Best.

pnavaro commented 4 years ago

Thank you for your response.

It was difficult to install tensorflow v1 with last versions of cuda toolkit and cuddn so I will try to upgrade to tensorflow v2. Before that I want to be sure that my computation is good. Thank you for reassuring me.