CODEWITHSUNDEEP
pythonpython-3.xkerasscikit-learnmeasure
Minions
Minions

Reputation: 5477

TypeError: f1_score() got an unexpected keyword argument 'average'

I'm extracting the F1_macro value using f1_score sklearn function.

I doing that for my Keras network, this is the code:

Y_dev_pred = self.model.predict([self.dev[0], self.dev[1]], batch_size=self.BatchSize, verbose=0)
Y_dev_pred = np.argmax(Y_dev_pred, axis=1)
self.Y_dev = np.argmax(self.dev[2], axis=1)
print('####### ', self.Y_dev.shape, ' ', Y_dev_pred.shape)
print(self.Y_dev, ' ### ', Y_dev_pred)
print(f1_score(self.Y_dev, Y_dev_pred, average='macro'))

Output:

#######  (19647,)   (19647,)
[1 0 0 ... 2 0 2]  ###  [0 0 0 ... 2 1 0]

Traceback (most recent call last):
  File "main.py", line 185, in <module>
    deep.run_model(reTrain=True)
  File ".../models/DL.py", line 477, in run_model
    self.evaluate_on_test()
  File ".../models/DL.py", line 448, in evaluate_on_test
    print(f1_score(self.Y_dev, Y_dev_pred, average='macro'))
TypeError: f1_score() got an unexpected keyword argument 'average'

What is the problem, did I miss something?

Upvotes: 1

Views: 7823

Answers (3)

user443729
user443729

Reputation: 11

My scikit learn is Name: scikit-learn Version: 0.24.2 I'm still getting this issue. My problem is multi-class not binary. If i remove the average i got this error Target is multiclass but average='binary'. Please choose another average setting, one of [None, 'micro', 'macro', 'weighted'].

Upvotes: 0

Vitor Bento
Vitor Bento

Reputation: 404

I got this error when in my code i had another function called f1_score as well.

Upvotes: 1

JAbr
JAbr

Reputation: 332

Check if any of your y_true labels or y_predicted labels has only two classes? check the unique values in each. average macro works only for multiclass if any of the variables has only two classes it throws a error like this! and yes check version of your sklearn,

Upvotes: 0

Related Questions