Programatically load CVAT bounding box datasets into Gradio using the CVAT SDK.

Install the required packages using the following command:

pip install 'cvat_sdk[pytorch]' gradio

Next, load your dataset into CVAT and annotate it. Optionally, you can load publicly available datasets.

In my setup, I loaded the Aquarium dataset from Roboflow. The dataset consists of bounding boxe annotations for the following classes:

• fish
• jellyfish
• penguins
• sharks
• puffins
• stingrays
• starfish

We can use the CVAT SDK to pull the dataset from CVAT into a PyTorch Dataset object.

In the most basic form:

from cvat_sdk import make_client
from cvat_sdk.pytorch import TaskVisionDataset

client = make_client(host, port=port, credentials=(user, password))
dataset = TaskVisionDataset(client, task_id)

## Visualize in Gradio and share

Read more here

Run the Gradio demo app using the following code snippet:

python run_gradio.py