MLReady-Bubble / README.md
mpfbench99's picture
Updating ReadMe
bf561a6 verified
metadata
license: cc-by-nc-sa-4.0

This dataset contains six families of datasets as npz files that can be read directly as follows:

import numpy as np
ds = np.read('/path/to/dataset/*.npz')["Y"]

Within each family we have the train and test datasets already splitted up for easy use,

train_X, train_Y, test_X, test_Y 

e.g. Use d1X_train_pad_flat.npz, d1Y_train_pad_flat.npz, d1X_test_pad_flat.npz and d1Y_test_pad_flat.npz as inputs to your SciML models. Our sample code is available in our github repo.

#Detailed Description of Each Family of datasets

  1. d1 : One input time step and one output time step
  2. d2 : Three input time steps and one output time step
  3. d3 : Five input time steps and one output time step
  4. d4 : Three input time steps and three output time steps
  5. d5 : Five input time steps and three output time steps
  6. d6 : Eight input time steps and three output time steps

For full description of these datasets, please refer to our website : https://lobster-app-cbyg9.ondigitalocean.app/