TensorFlow Integration

Overview

Neptune provides an easy integration with TensorFlow library. You can see all charts, images and graphs available through TensorBoard in Neptune.

Prerequisites

Neptune supports integration with TensorFlow 0.11, 0.12.1, and 1.0+.

Supported Features

Integration

To integrate your Neptune job with TensorFlow all you need to do is:

from deepsense import neptune
context = neptune.Context()
context.integrate_with_tensorflow()

All summaries added to SummaryWriter objects from TensorFlow will be available in Neptune.

To see an example of TensorFlow code integrated with Neptune, read: Flower Species Prediction.