[1]
“Loss Spike in Training Neural Networks”, JCM, Jan. 2025, doi: 10.4208/jcm.2412-m2024-0083.