Sie befinden Sich nicht im Netzwerk der Universität Paderborn. Der Zugriff auf elektronische Ressourcen ist gegebenenfalls nur via VPN oder Shibboleth (DFN-AAI) möglich. mehr Informationen...
Traditional machine translation based on RNN has two major defects: (1)it can only process word-by-word and result in slow training speed; (2)when the sentence is too long, gradient disappearance and explosion reduce the accuracy of translation. To solve this problem, this paper designs a Transformer based machine translation implemented by PyTorch. Compared with traditional machine translation, Transformer uses the attention mechanism to redesign the defects of RNN and effectively solves the problems of efficiency and forgetting. The French-English machine translation based on Transformer designed in this paper achieves a translation accuracy rate of 80% from French to English after training and practical application.