How to calculate the accuracy of this Transfer learning bert model

I have found this Git Hub code for intent classification. Could someone please tell me how to check the model accuracy of this?

https://colab.research.google.com/github/ShawonAshraf/nlu-jointbert-dl2021/blob/main/notebooks/nlu_jointbert_dl21.ipynb#scrollTo=Uae9vd77VT84

I have tried to add the joint_model.predict()

But not get the answer