NathanRxl / bnp-cardif-challenge

After deadline Kaggle Competition, Master's Data Science, 2017
0 stars 0 forks source link

Make the grid_search_script operational #6

Open NathanRxl opened 7 years ago

NathanRxl commented 7 years ago

As a beginning, a verbose grid search script must be able to run with the command:

& python3-conda grid_search_script.py

The arguments have to be defined by the assignee, but the grid search has to create a file containing the well formatted results (which model has been optimized, on which grid, using which loss, ...), in order to avoid doing multiple grid search we have already done.

languiol commented 7 years ago

Grid Search DONE

NathanRxl commented 7 years ago

Grid search script has not been done. It is too late now, we decide to go ahead without it.

languiol commented 7 years ago

On fait au plus rapide...car on va passer 3 minutes dessus. Si tu le souhaite, je te transmets un stacking faite sur "étagère": en prenant le code d'Alexandre et un staked generalization déjà implémenté. Ca fonctionne sans me casser la tête...

Le Mercredi 22 février 2017 16h41, NRxl <notifications@github.com> a écrit :

Grid search script has not been done. It is too late now, we decide to go ahead without it.— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub, or mute the thread.

NathanRxl commented 7 years ago

No worries. I just indicated that we won't fix this issue, that is it ! :)

I opened an issue for the stacking (#13), and I am assigned on it with Benjamin. I work on it now !