How Tensorboard Works

What is Tensorboard

The Tensorboard interface is used to visualize data using graphs and other tools to understand, debug, and optimize machine learning models. Tensorboard helps track metrics like loss and accuracy, visualize model graphs, and project embedding to lower-dimensional spaces, among other things.

For machine learning experiments, TensorBoard provides tools such as:

• Tracking and visualizing metrics such as loss and accuracy
• Visualizing the model graph (ops and layers)
• Viewing histograms of weights, biases, or other tensors as they change over time
• Projecting embeddings to a lower dimensional space
• Displaying images, text, and audio data
• Profiling TensorFlow programs
• And much more

The graphic below shows around the Tensorboard dashboard

The picture below shows the TensorBoard graph visualization panel. There are different tabs in the panel depending on the amount of information you add to the model.

• Scalars: Provide different training information
• Graphs: Display the model
• Histogram: Show weights as a histogram
• Distribution: Display the weight distribution
• Projector: Display the Principal Component Analysis and the T-SNE algorithm (Dimensionality reduction technique)