Ultrashort pulses classification in an Yb-doped fiber laser based on EfficientNet model

In this paper, we propose using the EfficientNet deep learning neural network to classify the ultrashort pulses in an Yb-doped mode-locked fiber laser. The results showed that the model achieved a classification accuracy of 99.8% for solitons, self-similar pulses, and amplifier similaritons, demonstrating its effectiveness in classifying ultrashort pulses.