[ Keras 常見問題 ] How to calculate precision and recall in Keras

 Source From Here

Question
I am building a multi-class classifier with Keras 2.02 (with Tensorflow backend)，and I do not know how to calculate precision and recall in Keras. Please help me.

How-To
Python package keras-metrics could be useful for this:

view plaincopy to clipboardprint?

1. import keras  
2.   
3. model = models.Sequential()  
4. model.add(keras.layers.Dense(1, activation="sigmoid", input_dim=2))  
5. model.add(keras.layers.Dense(1, activation="softmax"))  
6.   
7. m_preccision = keras.metrics.Precision()  
8. m_recall = keras.metrics.Recall()  
9. model.compile(optimizer="sgd",  
10.               loss="binary_crossentropy",  
11.               metrics = [m_preccision, m_recall])