TensorBoard - Tracking Metrics in Python
The following is a program used to demonstrate how to log different metrics in Tensorboard for visualization later. An example CNN network is used. The TensorBoard callback is defined to log the loss function and accuracy during training. Then, the extension is loaded in order to visualize these metrics. Some other related topics you might be interested in are TensorBoard - Confusion Metrics in Python, TensorBoard - Tuning Hyperparameters in Python, Converting Images into Arrays. The TensorBoard - Tracking Metrics in Python template is among the topics covered in the 365 Data Science Program.
Who is it for
This free .ipynb template is designed for any Learning Engineers and Machine Learning practitioners who want to monitor and track the performance of their machine learning models.
How it can help you
Machine Learning algorithms rarely preform well from the first attempt, as it is a iterative process. This is where the TensorBoard metric tracking capabilities and visualization come into play.