Accessing predictions when using k_argmax() instead of predict_classes()

Question

I've created a CNN in RStudio using keras to predict MNIST digits. I am now trying to predict with this model, since predict_classes() was deprecated, I'm attempting to use k_argmax() with the following code:

cnn_pred <- cnn_model %>%
    predict(x_test) %>%
    k_argmax()

When I print cnn_pred, this is what I get:

tf.Tensor([8 7 7 ... 3 4 9], shape=(4252), dtype=int64)

How do I access the predicted values in order to examine them and then print a confusion matrix?

amatof · Accepted Answer · 2022-08-12 00:16:02Z

1

As per the answers on a previous post, I've found the following code to work:

cnn_pred <- cnn_model %>%
   predict(x_test) %>%
   k_argmax() %>%
   as.integer()

answered Aug 12, 2022 at 0:16

amatof

1952 silver badges15 bronze badges

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

Collectives™ on Stack Overflow

Accessing predictions when using k_argmax() instead of predict_classes()

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related