Deep Learning architectures for time series require a large number of training samples, however traditional sample size estimation for sufficient model performance is not applicable for machine learning, especially in the field of electrocardiograms (ECGs). This presentation outlines a sample size estimation strategy for binary classification problems on ECGs using different deep learning architectures and the large publicly available PTB-XL dataset. The post-hoc sample size estimations are based on a benchmark across different architectures (XResNet, Inception-, XceptionTime, fully convolutional network), and different classification targets (Myocardial Infarction, Conduction Disturbance, ST/T Change, and Sex). The results indicate trends for required sample sizes for given tasks and architectures, which can be used as orientation for future ECG studies or feasibility aspects.