gwastro / ml-mock-data-challenge-1

Code and documentation for the first machine learning focused mock data challenge hosted by the Albert-Einstein-Institut Hannover and the Friedrich-Schiller Universität Jena.
Apache License 2.0
39 stars 16 forks source link

Same noise generated by generate_data.py for both detectors #5

Closed vivinousi closed 2 years ago

vivinousi commented 2 years ago

The noise generated by generated_data.py for dataset 1 seems to be the same for both detectors, as checked by

np.allclose(rawdata[0], rawdata[1], rtol=1e-45, atol=1e-36)

in example_torch.py#L93

MarlinSchaefer commented 2 years ago

Thank you for reporting this issue. This is caused by the seed in noise generation being the same for all detectors. It mainly affects data generation for data set 1. However it also has an influence on data sets 2 and 3, because the basic noise is the same. Just the coloring between the two is different.

MarlinSchaefer commented 2 years ago

Solved by #6