The diagnosis of cardiovascular diseases is a lengthy and expensive procedure that often requires visual inspection of ECG signals by experts. In order to improve patient management and reduce healthcare costs, automated detection of these pathologies is of utmost importance. In this study, we classify 12-lead ECGs into nine classes (AF, I-AVB, LBBB, Normal, RBBB, PAC, PVC, STD or STE) as part of the Physionet/Computing in Cardiology Challenge 2020. In the unofficial stage, I used a convolutional neural network that achieved good results in CinC2017 to experiment, Inspired by the neural network, I plan to create a novel convolutional neural network (based on state-of-the-art neural network backbone) at the official stage. In the unofficial stage, The convolutional neural network obtained an F_2 Score of 0.841 on the training set (5-fold cross-validation), and 0.807 on the hidden test set. In the next work, in addition to creating a novel framework, I optimize the existing model so that it will be more suitable for 12-lead ECGs.