A simulated dataset of silicon pixel clusters produced by charged particles (pions), where the particle kinematics are taken from fitted tracks in CMS Run 2 data. The interaction of the charged particles with a 100 micron thick silicon pixel detector is simulated using a time-sliced version of PixelAV. The pixel detector is taken to be a 21x13 array of pixels with 50x12.5 micron pitch, situated 30mm from the particle's origin point. The detector is immersed in a 3.8T magnetic field parallel to the sensor x coordinate. The simulation covers a $p_T$ range from 0.2-0.4 GeV. Positively and negatively charged particles are saved in the directories positive-charge and negative-charge, respectively. Truth properties of each particle are saved in the labels files, which include columns x-entry, y-entry, z-entry, n_x, n_y, n_z, number_eh_pairs, y-local, pt, cotAlpha, cotBeta, y-midplane, x-midplane. The particle impact point on the sensor surface is described by (x-entry, y-entry), and the impact point at the mid-plane of the sensor is (x-midplane, y-midplane). These impact points are in units of microns. The track direction is given by (n_x, n_y, n_z). The number of electron-hole pairs created in the silicon is number_eh_pairs. The y distance in mm between the center of the 21x13 pixel array and the center of a flat module is denoted by y-local (see included diagram). The particle transverse momentum ($p_T$) in GeV is stored in the pt variable, with the sign indicating the sign of the particle charge. The angle of incidence in the x-z plane is described by cotAlpha, and the angle of incidence in the y-z plane (the bending plane of the magnetic field) is described by cotBeta. The deposited charge per pixel per time slice is saved in the recon3D files. Reshaping each line as (20,13,21) gives the three dimensional cluster in (time, y, x). Each slice in time corresponds to a window of 200 ps. Changes with respect to the first version of this dataset: File format is changed to parquet for better compression Files contain simulation of both negatively and positively charged particles (before, the charges were saved separately) 20 200ps time slices are included, rather than only 8 Small bug fix in simulation impacting charge at low $p_T$