This is a limited subset of the data used for training in arXiv:2005.05334. The network architectures and instructions to generate more data are available at here. Electromagnetic calorimeter for the ILD consists of 30 active silicon layers in a tungsten absorber stack with 20 layers of 2.1 mm followed by 10 layers of 4.2 mm thickness respectively. We project the sensors onto a rectangular grid of 30×30×30 cells. Each cell in this grid corresponds to exactly one sensor, resulting in total of 27k channels. The file has the following structure: Group named 30x30 energy : Dataset {1000, 1} layers : Dataset {1000, 30, 30, 30} The energy specifies the true energy of the incoming photons in units of GeV, where layers represent the energy deposited (MeV) in 30 layers of the calorimeter in an image data format. This file contains approximately 24.000 showers.

{"references": ["Getting High: High Fidelity Simulation of High Granularity Calorimeters with High Speed [arxiv:2005.05334]"]}

You may want to generate this data yourself: https://github.com/FLC-QU-hep/getting_high