SimonBlanke / Hyperactive

An optimization and data collection toolbox for convenient and fast prototyping of computationally expensive models.
https://simonblanke.github.io/hyperactive-documentation
MIT License
503 stars 41 forks source link

[DOC] jupyter notebook tutorial #77

Open fkiraly opened 1 month ago

fkiraly commented 1 month ago

I would suggest to add a "standard" jupyter notebook tutorial to the entire package, in the examples directory, and also link ot from the docs.

In addition, it would be nice if users could read it on binder or otherwise in the cloud.

fkiraly commented 4 weeks ago

More specific suggestions:

Further, due to the popularity of AI, perhaps the content that is, technically, secondary (AI applications), might be worth to place as the "first" notebook? That is, lead with data sci and AI, and present the more general applications in a second/advanced notebook.

SimonBlanke commented 4 weeks ago

Am currently working on some corrections of this notebook (for general usage): https://github.com/SimonBlanke/hyperactive-tutorials/blob/main/notebooks/hyperactive_tutorial.ipynb

SimonBlanke commented 4 weeks ago

The Introduction notebook can be viewed (and opened in binder) here: https://nbviewer.org/github/SimonBlanke/hyperactive-tutorial/blob/main/notebooks/hyperactive_tutorial.ipynb

fkiraly commented 4 weeks ago

oh, I did not notice this other repo!

I am then wondering, why not add the nb to this repo, and link from docs and/or have a binder or colab link as well?

The notebooks do not need to go in the wheels, that keeps them small. Though I recognize that the git history and file size may be blown up having notebooks in this repo.

SimonBlanke commented 4 weeks ago

Hello @fkiraly,

I keep the notebooks in a separate repo, because:

I did not link the notebook to the readme/doc yet, because some changes were necessary.

SimonBlanke commented 4 weeks ago

I am then wondering, why not add the nb to this repo, and link from docs and/or have a binder or colab link as well?

I will see if this is possible in a separate repo