Fix: Copy trained weights to MC Dropout model for correct inference

[ShindeShivam]

Chapter 11
Cell: Exercise [130]

Fix :

This small fix ensures that the MC Dropout model uses the same trained weights as the original model by adding:

mc_model.set_weights(model.get_weights())

Without this, the MC model uses random weights, leading to incorrect predictions and low accuracy.
Now the example works as intended, with dropout active during inference.

Also, I’ve opened PRs recently that are still pending review — just wanted to kindly mention in case they were missed.🙂

#196
#202
#205
#206

Thanks again for the fantastic book!

— Shivam

[ageron]

Thanks for your PR!

I don't think adding this line makes any difference since we're already reusing the exact same layers, including their weights.