Update scalars_and_keras.ipynb

[rmothukuru]

Added a statement to visualize Metrics, in Tensorboard Scalars Documentation

• Motivation for features / changes : Point raised in this issue

• Technical description of changes : Explanation to Visualizing Metrics was missing in the documentation.

• Screenshots of UI changes

• Detailed steps to verify changes work correctly (as executed by you) : Verified it in my Branch

• Alternate designs / implementations considered : This is just a Documentation change

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[wchargin]

Instead of a loud disclaimer, could we just add metrics=['accuracy']
to our sample model above? It seems likely that users will already be
tracking metrics that they’re interested in, so no need to draw their
attention to an Important Note; we can just demonstrate the right
workflow.

[rmothukuru]

Instead of a loud disclaimer, could we just add metrics=['accuracy']
to our sample model above? It seems likely that users will already be
tracking metrics that they’re interested in, so no need to draw their
attention to an Important Note; we can just demonstrate the right
workflow.

@wchargin,
We can do that but then the Tensorboard Graph, which shows only Loss, has to be updated, and even the Heading, training_the_model_and_logging_loss needs to be changed.

Please let me know how to proceed. Thanks!

[psybuzz]

To chime in, I think it is reasonable to add metrics=['accuracy'] and update the docs accordingly, with updated headings, instructions, and screenshots as needed.