Fig. 8From: Real-time earthquake magnitude estimation via a deep learning network based on waveform and text mixed modalSpatial distribution of the magnitude prediction errors of the proposed model in the study region with time (\({t}_{1}\)) at 1 (a), 3 (b), 5 (c), 6 (d), 7 (e), and 8 (f) s when the earthquake’s triggered seismometers exceed one. The squares and circles represent the earthquakes located in the inland and offshore, respectively. Different colors indicate different error levelsBack to article page