-
-
Notifications
You must be signed in to change notification settings - Fork 36
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add section on hyperparameter tuning #437
Conversation
❌ DANGER ❌This pull request has modified workflows that created output. Close this now. |
…-intro into hyperparameter-tuning
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good to me. I added a few minor remarks.
This PR is unrelated to the changes here, but perhaps a good occasion to merge in: #451
episodes/4-advanced-layer-types.Rmd
Outdated
The goal is to show that hyperparameter tuning can be done easily with `keras_tuner`, not to memorize all the exact syntax of how to do it. This will probably save you half an hour of participants typing over code that they already know from before. In addition, on really slow machines running the grid search could possibly take more than 10 minutes. | ||
::: | ||
|
||
In general if you are varying hyperparameters (such as the dropout rate) to find the combination of hyperparameters with the best model performance this is called hyperparameter tuning. A naive way to do this is to write a for-loop and train a slightly different model in every cycle. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think a sentence with a general reminder to define what hyperparameters are would be useful.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Fixed in c38b110
Co-authored-by: Carsten Schnober <[email protected]>
Co-authored-by: Carsten Schnober <[email protected]>
Co-authored-by: Carsten Schnober <[email protected]>
Fix path to image: remove leading slash.
…-intro into hyperparameter-tuning Conflicts: episodes/4-advanced-layer-types.Rmd
I made the results reflect the new dollarstreet dataset, since I didn't change any content since @carschno' review, I will merge this now. |
Auto-generated via {sandpaper} Source : 9da1b65 Branch : main Author : Sven van der Burg <[email protected]> Time : 2024-05-20 11:47:51 +0000 Message : Add section on hyperparameter tuning (#437) * Add section on hyperparameter tuning * Fix path to image: remove leading slash. * Update episodes/4-advanced-layer-types.Rmd Co-authored-by: Carsten Schnober <[email protected]> * Update episodes/4-advanced-layer-types.Rmd Co-authored-by: Carsten Schnober <[email protected]> * Update episodes/4-advanced-layer-types.Rmd Co-authored-by: Carsten Schnober <[email protected]> * Improve explanations of hyperparameter tuning * Update hyperparamter tuning section for dollar street dataset --------- Co-authored-by: Carsten Schnober <[email protected]> Co-authored-by: Carsten Schnober <[email protected]>
Auto-generated via {sandpaper} Source : 176e0c8 Branch : md-outputs Author : GitHub Actions <[email protected]> Time : 2024-05-20 11:48:40 +0000 Message : markdown source builds Auto-generated via {sandpaper} Source : 9da1b65 Branch : main Author : Sven van der Burg <[email protected]> Time : 2024-05-20 11:47:51 +0000 Message : Add section on hyperparameter tuning (#437) * Add section on hyperparameter tuning * Fix path to image: remove leading slash. * Update episodes/4-advanced-layer-types.Rmd Co-authored-by: Carsten Schnober <[email protected]> * Update episodes/4-advanced-layer-types.Rmd Co-authored-by: Carsten Schnober <[email protected]> * Update episodes/4-advanced-layer-types.Rmd Co-authored-by: Carsten Schnober <[email protected]> * Improve explanations of hyperparameter tuning * Update hyperparamter tuning section for dollar street dataset --------- Co-authored-by: Carsten Schnober <[email protected]> Co-authored-by: Carsten Schnober <[email protected]>
Auto-generated via {sandpaper} Source : 9da1b65 Branch : main Author : Sven van der Burg <[email protected]> Time : 2024-05-20 11:47:51 +0000 Message : Add section on hyperparameter tuning (#437) * Add section on hyperparameter tuning * Fix path to image: remove leading slash. * Update episodes/4-advanced-layer-types.Rmd Co-authored-by: Carsten Schnober <[email protected]> * Update episodes/4-advanced-layer-types.Rmd Co-authored-by: Carsten Schnober <[email protected]> * Update episodes/4-advanced-layer-types.Rmd Co-authored-by: Carsten Schnober <[email protected]> * Improve explanations of hyperparameter tuning * Update hyperparamter tuning section for dollar street dataset --------- Co-authored-by: Carsten Schnober <[email protected]> Co-authored-by: Carsten Schnober <[email protected]>
Auto-generated via {sandpaper} Source : 1535d4b Branch : md-outputs Author : GitHub Actions <[email protected]> Time : 2024-05-21 00:53:10 +0000 Message : markdown source builds Auto-generated via {sandpaper} Source : 9da1b65 Branch : main Author : Sven van der Burg <[email protected]> Time : 2024-05-20 11:47:51 +0000 Message : Add section on hyperparameter tuning (#437) * Add section on hyperparameter tuning * Fix path to image: remove leading slash. * Update episodes/4-advanced-layer-types.Rmd Co-authored-by: Carsten Schnober <[email protected]> * Update episodes/4-advanced-layer-types.Rmd Co-authored-by: Carsten Schnober <[email protected]> * Update episodes/4-advanced-layer-types.Rmd Co-authored-by: Carsten Schnober <[email protected]> * Improve explanations of hyperparameter tuning * Update hyperparamter tuning section for dollar street dataset --------- Co-authored-by: Carsten Schnober <[email protected]> Co-authored-by: Carsten Schnober <[email protected]>
Fixes #351
@Reviewer: This is now in episode 4, I was thinking to maybe put it together with transfer learning in the new episode 5 (See #349). Episode 5 could then be called 'advanced deep learning concepts'.