 This paper proposes a novel fully convolutional neural network, sleepFCN, for sleep stage classification from single channel electroencephalogram, e.g., data. It uses multi-scale feature extraction and residual dilated causal convolutions to extract features and encode temporal sequences, respectively. Additionally, it incorporates a weight corresponding to the number of samples of each class in its loss function, which improves accuracy and reduces training time compared to other methods. Experimental results on the sleep EDF and SHHS data sets demonstrate that sleepFCN outperforms existing approaches in terms of classification accuracy and learning speed. This article was authored by Narayesh Ghashtazbi, Reza Bustani, and Saeed Saniyai.