TypeError: fit() got an unexpected keyword argument 'show_accuracy' ---- Graph().fit

[fluency03]

I am training the Graph model with using fit with parameter show_accuracy. However, I got the error:

TypeError: fit() got an unexpected keyword argument 'show_accuracy'

When I check the source code on github here, it shows that Graph's fit has show_accuracy:

def fit(self, data, batch_size=128, nb_epoch=100, verbose=1, callbacks=[],
        validation_split=0., validation_data=None, shuffle=True,
        show_accuracy=False,
        class_weight={}, sample_weight={}):

But when I check the release version 0.3.2, it is like:

def fit(self, data, batch_size=128, nb_epoch=100, verbose=1, callbacks=[],
        validation_split=0., validation_data=None, shuffle=True,
        class_weight={}, sample_weight={}):

I am wondering, which one is the newest version?

I used this pip install git+git://github.com/fchollet/keras.git to install Keras. Even this, the Graph's fit I got has no show_accuracy parameter.

[jisungk]

Hi @fluency03,

The Github version is the newer one -- it has some modules which the PyPI version does not have. I checked my version (which was downloaded from Github) and I see the following code:

def fit(self, data, batch_size=128, nb_epoch=100, verbose=1, callbacks=[],
            validation_split=0., validation_data=None, shuffle=True,
            show_accuracy=False,
            class_weight={}, sample_weight={}):

Did you previously install Keras using the command, sudo pip install keras?

If so, maybe you should try upgrading your installation with the following command:
sudo pip install --upgrade git+git://github.com/fchollet/keras.git

[fluency03]

@jisungk I have done this sudo pip install git+git://github.com/fchollet/keras.git. Is it necessary to add --upgrade?

[jisungk]

@fluency03 if you have any previous installation of Keras, then calling sudo pip install.... will not make any updates. You have to specify that you want to upgrade your existing installation with the --upgrade flag.

Hope this helps!

[fluency03]

@jisungk It's solved. thanks

[KingfaLuis]

model.fit(X_train, Y_train,
batch_size=128, nb_epoch=4,
show_accuracy=True, verbose=1,
validation_data=(X_test, Y_test))

get the error
TypeError Traceback (most recent call last)
in ()
2 batch_size=128, nb_epoch=4,
3 show_accuracy=True, verbose=1,
----> 4 validation_data=(X_test, Y_test))

/usr/local/lib/python3.6/dist-packages/keras/models.py in fit(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, **kwargs)
983 epochs = kwargs.pop('nb_epoch')
984 if kwargs:
--> 985 raise TypeError('Unrecognized keyword arguments: ' + str(kwargs))
986
987 if not self.built:

TypeError: Unrecognized keyword arguments: {'show_accuracy': True}

[tsinghuazl22]

what about the error fit_generator() got an unexpected keyword argument 'samples_per_epoch'?My keras version is 2.4.1.