Model Remediation is a library that provides solutions for machine learning practitioners working to create and train models in a way that reduces or eliminates user harm resulting from underlying performance biases.
The current implementation of TF Dataset to TF Examples List uses a lot of RAM and runs very slowly on large datasets.
In this colab, the CelebA dataset, as processed by the fairness indicators tutorial, can't be processed on a default hosted runtime. After 18-20 minutes, the runtime crashes because RAM requirements are exceeded.
Because of all the preprocessing required, it's not easy to create a smaller reproducible example.
The current implementation of TF Dataset to TF Examples List uses a lot of RAM and runs very slowly on large datasets.
In this colab, the CelebA dataset, as processed by the fairness indicators tutorial, can't be processed on a default hosted runtime. After 18-20 minutes, the runtime crashes because RAM requirements are exceeded.
Because of all the preprocessing required, it's not easy to create a smaller reproducible example.