Skip to content

Experiment pipeline

  • Fix fetchers in data_loader.py.
  • Put the dataset normalization in the dataloading part, and load the data used in bayesian hyperparameters research using that.
  • Compute the bayesian hyperparameters research over k seeds.
  • [IN PROGRESS. Remainings: lfw_people, covtype, rcv1] Compute and upload new best param files using dataset normalization).
  • Organize better results storing (cause right now it's hard to compute easily the plots specified in compute_results.py from the results dir).
  • Prepare stage 1 code.
  • Prepare stage 2 code.
  • Prepare stage 3 code.
  • Run a script for stages 1, 2 and 3 on all remaining datasets (all except california_housing).
  • Prepare stage 4 code.
  • Prepare stage 5 code.
  • Check that omp multiclasses classifier is working as expected.
  • Integrate baseline drafts into main code.
  • Run stage 1 experiments.
  • Run stage 2 experiments.
  • Run stage 3 experiments.
  • Run stage 4 experiments.
  • Run stage 5 experiments.
Edited by Charly Lamothe