X-ray diffraction dataset for experimental noise filtering

<p>X-ray diffraction data set for the training of noise filtering algorithms. The data set contains groups of low- and high-counting statistics pairs. The sampling times are&nbsp;mostly 1 (20) seconds for low (high) counting data. Three files in HDF5 format are provided, corresponding to a training, validation and test data set. Each data group contains sequences of 41 consecutive frames, corresponding to a scan along the reciprocal h-direction. Next to the raw data, sampling times and monitor values are included. The test data set additionally contains denoised low-count frames obtained from a pre-trained neural network.</p> <p>Additionally, files containing the trained model weights are included for two different architectures described in the main article (10.1038/s42256-024-00790-1).</p> <p>The data has been recorded on a La_1.88Sr_0.12CuO₄&nbsp;single crystal at the beamline P21.1 at the PETRA III storage ring at DESY in Hamburg, Germany. The scattering intensities were recorded using Dectris Pilatus 100K CdTe detector. The diffractometer was operated with 100 keV photons and the sample was cooled to T ~ 30 K. The data contains different signals such as weak 2D charge density wave order, fundamental Bragg peaks, powder lines, spurions and dead pixels.</p>