Implementation of Farsi speech components tagging with cnn_bilstm hybrid architecture Yasman Safari Machine learning lesson Professor: Dr. Salim Y Spring 00 Abstract Labeling speech components is one of the most important basic tasks in natural language processing. For this issue There are many classification, traditional and deep models. In this article, we tried to combine two architectures cnn and bilstm extract more important features from text to classify pos tags more accurately do it. Because by combining two architectures, features are extracted at different levels. Finally, the mentioned model with the model The basic models and their photo models were compared and the results were presented.