categoricalentropy vs sparsecategoricalentropy

[Chikamalu]

Hi Everyone
Please i am at the point of building a multiclass model and i built one with the input as integers and loss as categorical entropy and my code ran. i was expecting an error.
below is the code.
could you look at this and tell me what you think.
thanks

`import tensorflow as tf
from tensorflow.keras.datasets import fashion_mnist

The data has already been sorted into training and test sets for us

(train_data, train_labels), (test_data, test_labels) = fashion_mnist.load_data()

Set random seed

tf.random.set_seed(42)

Build the model

model_12 = tf.keras.Sequential([
tf.keras.layers.Flatten(input_shape=(28, 28)), # input layer (we had to reshape 28x28 to 784, the Flatten layer does this for us)
tf.keras.layers.Dense(3, activation = "relu"),
tf.keras.layers.Dense(3, activation= "relu"),
tf.keras.layers.Dense(1, activation = "softmax")
])

Compile the model

model_12.compile(loss= "CategoricalCrossentropy",
optimizer = tf.keras.optimizers.Adam(),
metrics=["accuracy"])

Fit the model

non_norm_history1 = model_12.fit(train_data,
train_labels,
epochs=10,
validation_data=(test_data, test_labels), verbose=1)`

[ishandandekar]

Hey @Chikamalu , there is a slight difference in sparse_categorical_crossentropy and categorical_crossentropy. Sparse categorical cross-entropy expects your labels in the form of integers. Whereas, Categorical cross-entropy expects your labels to be one hot encoded. For example, if you have three classes, then the labels for sparse categorical cross-entropy should be in the form 0, 1, 2 and for categorical cross-entropy the labels should be in the form [1, 0, 0], [0, 1, 0] and [0, 0, 1]. The internal mathematical loss function is same for both, it is just the difference between the type of label formatting. Your fashion mnist data has the labels ready for sparse_categorical_crossentropy, so it is preferred to use this loss function. You can also one-hot encode these labels and use the desired Categorical Crossentropy. Hope this helps!