Skip to content

Skyblueballykid/RIIAA-18

Repository files navigation

RIIAA Escuela18

My edits and code related to the Higgs Boson dataset Deep Learning competition of the 2018 International Artificial Intelligence meeting at UNAM

My final submission consisted of:

  • 7 hidden layers with 80 units each
  • A truncated Normal distribution kernel
  • ELU activation
  • Batch size of 28
  • Adam optimizer (outperformed Powersign for this configuration)

The model accuracy consistently topped out at around 170 epochs of training time.

The IAmigos_RIIAA_Reto_Machine_Learning.ipynb file was the winning submission from another team. This model used Sklearn's GBClassifier.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published