Keras AttributeError: 'Tensor' object has no attribute 'log'

Question

I am having an error - 'Tensor' object has no attribute 'log' that I code in Keras to build a network while apply custom loss function to Keras. I think somehow I need to get rid of np.log but not sure how.

import Numpy 

import numpy as np

Custom Function

def rmsle(y_pred,y_test):
   return np.sqrt(np.mean((np.log(1+y_pred) - np.log(1+y_test))**2))

My network

def base_model():
   model = Sequential()
   model.add(Dense(50, input_dim=X_train.shape[1], init='normal',     activation='sigmoid'))
   model.add(Dropout(0.5))

   model.add(Dense(1, init='normal'))
   sgd = SGD(lr=0.01, momentum=0.8, decay=0.1, nesterov=False)
   model.compile(loss=rmsle, optimizer = sgd)# )'adam') #
   return model

keras = KerasRegressor(build_fn=base_model, nb_epoch=80, batch_size=1,verbose=1)
keras.fit(X_train ,y_train)

When i check the error msg in detail, it shows that

424         """
425         # score_array has ndim >= 2
--> 426         score_array = fn(y_true, y_pred)
427         if mask is not None:
428             # Cast the mask to floatX to avoid float64 upcasting in theano
2     #return np.sqrt(np.mean(np.square( np.log( (np.exp(a)) + 1 ) - np.log((np.exp(b))+1) )))
----> 4     return np.sqrt(np.mean((np.log(1+y_pred) - np.log(1+y_test))**2))
2     #return np.sqrt(np.mean(np.square( np.log( (np.exp(a)) + 1 ) - np.log((np.exp(b))+1) )))

Please show how you import numpy, ideally show us a Minimal, Complete, and Verifiable example of your problem. My first suspicion would be that you re-assign np to something other than numpy, so that instead of calling numpy.log, you are trying to call whatever Keras object you assigned to np. — IonicSolutions
– IonicSolutions, Commented Jul 11, 2018 at 5:34
Can you show the whole code? Your snippet is at least missing the import statements. For debugging, you can put a print(type(np), np.__file__) into your loss function to check that np really is the numpy module. — IonicSolutions
– IonicSolutions, Commented Jul 11, 2018 at 5:56
Also, can you please show the error message and the stack trace? — IonicSolutions
– IonicSolutions, Commented Jul 11, 2018 at 5:57

rvinas · Accepted Answer · 2018-07-11 06:48:24Z

1

You must use valid tensor operations from your backend (i.e. from keras.backend) in order to define a custom loss function. For example, your loss function could be defined as follows:

import keras.backend as K

def rmsle(y_test, y_pred):
    return K.sqrt(K.mean(K.square(K.log(1 + y_pred) - K.log(1 + y_test))))

NOTE: Keras expects the first argument to be y_test (alias the ground truth).

answered Jul 11, 2018 at 6:48

rvinas

11.9k38 silver badges58 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Milind Deore · Accepted Answer · 2018-07-11 06:21:30Z

0

Lambda layers in Keras help you to implement functionality that is not prebuilt and which do not require trainable weights. So you get this liberty to implement your own logic as in this case 'Log'

This can also be done using keras Lambda layer as below:

from keras.layers import Lambda
import keras.backend as K

Define your function here:

def logFun(x):
   return keras.backend.log(x)

And later create a lambda layer:

model.add(Lambda(logFun, ...))

edited Jul 11, 2018 at 6:21

answered Jul 11, 2018 at 5:58

Milind Deore

3,0615 gold badges27 silver badges45 bronze badges

1 Comment

foy Over a year ago

thx for helping man, it gives me idea how to solve it. thx again

Collectives™ on Stack Overflow

Keras AttributeError: 'Tensor' object has no attribute 'log'

2 Answers 2

Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related