Skip to main content

Embedding Visualizer Using Tensorboard

Project description

TensorBoard Embedding Visualizer


As the name suggests, this tool is made for visualizing high dimensional embeddings into 3D space using TensorBoard Projector tool. It performs PCA and also has an option for T-SNE for doing dimensionality reduction. All you have to do is provide you embeddings data and its corresponding labels and optionally images(paths) in pickle format as a dictionary. The structure of this dictionary is explained below.

Getting Started


  • Python3
  • Tensorflow > 1.4


Install this tool using pip with the following command.
pip install tbev
Upgrade to the latest version by using the --upgrade flag
This will be installed as a command line tool and you can simply run tbev in you command line to get started.


    tbev demo 
    tbev <pickle_file> [--logdir=<path>]

    -h --h    Show help screen
    --logdir=<path>  Location to store log files [default: ./logs/]

tbev <pickle_file> [--logdir=<path>] You will need to pass your embeddings in form of a pickle file. The pickle file should contain a dictionary in the following format.

    "embedding":2D matrix of shape [m embeddings, n dimensions],
        "label1": List of shape (m,), A label for each embedding,
        "label2": List of shape (m,), A label for each embedding,
        You can put as many labels as you want.
    "sprite_paths":(Optional) List of shape (m,), A path to image to be shown in tensorboard.

Save this dictionary into a pickle file and use the command tbev <pickle_file> [--logdir=<path>]
This will create a ./logs folder by default where it will store the checkpoints file. You can optionally mention your own name for storing the logs using the --logdir option. If everything works out well. It will start the tensorboard server for you.


After the Tensorboard is started, it will show you the local URL to view the tensorboard. By default it is localhost:6006.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for tbev, version 0.1.2
Filename, size File type Python version Upload date Hashes
Filename, size tbev-0.1.2-py3-none-any.whl (56.9 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size tbev-0.1.2.tar.gz (52.8 kB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring DigiCert DigiCert EV certificate Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page