Menu
Is there a possibility to use UMAP with Tensorboard projector. Indeed, at the moment, we can only use PCA or T-SNE with it.
Alternatively, is there a Python library that would allow generating interactive plots such as Tensorboard's?
963
TensorBoard Projector allows to graphically represent low-dimensional embeddings. Here I show you how, instead of displaying a point, you can render the image to which the embedding refers.
The Embedding Projector offers three commonly used methods of data dimensionality reduction, which allow easier visualization of complex data: PCA, t-SNE and custom linear projections. PCA is often effective at exploring the internal structure of the embeddings, revealing the most influential dimensions in the data.
Just like t-SNE, UMAP is a dimensionality reduction specifically designed for visualizing complex data in low dimensions (2D or 3D).
TensorBoard is a tool for providing the measurements and visualizations needed during the machine learning workflow. It enables tracking experiment metrics like loss and accuracy, visualizing the model graph, projecting embeddings to a lower dimensional space, and much more.
the online version of the Tensorflow Projector does now include UMAP. I guess they have added that feature recently as I was using the projector quite some times lately and until last week they only provided PCA, TSNE or a custom visualization in the online version. But now it is there, give it a try.
Tensorflow Projector with UMAP
I didn't find a Python library to create such interactive visualizations as tensorboards but if you'd like to try a JavaScript Version you could start with this making-an-interactive-umap-visualization-of-the-mnist-data-set blog post. Here is the showcase: Online UMAP Explorer with three.js
93
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With
