TFRecord loader

Overview

This is an example data-loader plugin that lets you view a TFRecord of Events (i.e., Tensorboard log files). It uses the external data loader mechanism to add this capability to the Rerun viewer without modifying the viewer itself.

This example is written in Python, and uses TensorFlow to read the files. The events are then logged to Rerun.

Note: Not all events are supported yet. Scalars, images, text, and tensors should work. Unsupported events are skipped.

Installing the plug-in

The repository has detailed installation instruction. In a nutshell, the easiest is to use pipx:

pipx install git+https://github.com/rerun-io/rerun-loader-python-example-tfrecord.git pipx ensurepath

Try it out

To try the plug-in, first download an example xxx.tfevents.xxx file:

curl -OL https://github.com/rerun-io/rerun-loader-python-example-tfrecord/raw/main/events.tfevents.example

Then you can open the viewer and open the file using drag-and-drop or the open dialog, or you can open it directly from the terminal:

rerun events.tfevents.example