Hyperparameter tuning in the model for better performance by introduction of callback function.

**Describe the bug**
Finetuning the model of Lymphoma Cancer Classification and introducing a callback fuction to get the best epoch.
**To Reproduce**
Steps to reproduce the behavior:
1. Go to Breadcrumbs[tfjs-models](https://github.com/camicroscope/tfjs-models/tree/master)/[Classification Sample Models](https://github.com/camicroscope/tfjs-models/tree/master/Classification%20Sample%20Models)/[lymphoma-cancer-classification](https://github.com/camicroscope/tfjs-models/tree/master/Classification%20Sample%20Models/lymphoma-cancer-classification)
/model.ipynb

**Expected behavior**
Better learning rate with a callback function to get the best epoch and thus best result.

Thank you

Open Classification Sample Models/lymphoma-cancer-classification/model.ipynb and review the current model training flow and learning-rate settings. Determine how the requested callback should identify the best epoch, then verify that training produces improved learning-rate behavior and the best reported result.

Written by the indexing model from the issue text.