Step 5: Visualizing Datasets
Visualizing and inspecting your datasets.
One of Hub's core features is to enable users to visualize and interpret large amounts of data. Let's load the COCO dataset, which is one of the most popular datasets in computer vision.
1
import hub
2
3
ds = hub.load('hub://activeloop/coco-train')
Copied!
The tensor layout for this dataset can be inspected using:
1
ds.summary()
Copied!
The dataset can be visualized in Activeloop Platform, or using an iframe in a jupyter notebook:
1
ds.visualize()
Copied!
Visualizing datasets in Activeloop Platform will unlock more features and faster performance compared to visualization in Jupyter notebooks.

Visualizing your own datasets

Any hub dataset can be visualized using the methods above as long as it follows the conventions necessary for the visualization engine to interpret and parse the data. These conventions are explained in the link below: