Solve InvalidArgumentError in tensorflow

Question

Nothing much I can say either from that I'm getting an error which I can't seem to solve from the following piece of code.

import tensorflow as tf
import matplotlib.pyplot as plt

(x_train, y_train), (x_test, y_test) = tf.keras.datasets.cifar10.load_data()

inpTensor = tf.keras.layers.Input(x_train.shape[1:],)

hidden1Out = tf.keras.layers.Dense(units=128, activation=tf.nn.relu)(inpTensor)
hidden2Out = tf.keras.layers.Dense(units=128, activation=tf.nn.relu)(hidden1Out)  
finalOut = tf.keras.layers.Dense(units=10, activation=tf.nn.softmax)(hidden2Out)

model = tf.keras.Model(inputs=inpTensor, outputs=finalOut)

model.summary()

model.compile(optimizer='adam',
              loss='sparse_categorical_crossentropy',
              metrics=['accuracy'])

model.fit(x_train,y_train, epochs = 4)

I've tried changing the loss function to 'categorical_crossentropy', but seemed to not work either. I am running Python 3.7 and would really appriciate some help. I am kind of new to this as well.

Thanks in advance.

We cannot guess your error, please include the complete traceback. — Dr. Snoopy
– Dr. Snoopy, Commented Jul 5, 2020 at 14:49

Marco Cerliani · Accepted Answer · 2020-07-05 14:53:06Z

the problem is in the way you manage the dimensionality in your network... you receive 3D images and don't pass to 2D to obtain probabilities... this can simply be done using Flatten or global pooling operation. sparse_categorical_crossentropy is correct in your case. here an example

import tensorflow as tf
import matplotlib.pyplot as plt

(x_train, y_train), (x_test, y_test) = tf.keras.datasets.cifar10.load_data()

inpTensor = tf.keras.layers.Input(x_train.shape[1:],)

hidden1Out = tf.keras.layers.Dense(units=128, activation=tf.nn.relu)(inpTensor)
hidden2Out = tf.keras.layers.Dense(units=128, activation=tf.nn.relu)(hidden1Out)
pooling = tf.keras.layers.GlobalMaxPool2D()(hidden2Out) #<== also GlobalAvgPool2D or Flatten are ok
finalOut = tf.keras.layers.Dense(units=10, activation=tf.nn.softmax)(pooling)

model = tf.keras.Model(inputs=inpTensor, outputs=finalOut)

model.summary()

model.compile(optimizer='adam',
              loss='sparse_categorical_crossentropy',
              metrics=['accuracy'])

model.fit(x_train,y_train, epochs = 4)

Collectives™ on Stack Overflow

Solve InvalidArgumentError in tensorflow

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related