Overview

Overview

Dataloop pipelines provides a user-friendly interface for building, managing, and monitoring end-to-end machine learning workflows. Dataloop's pipeline can be made up of different nodes, each with a different role in the pipeline, such as storing data, executing functions, training models, or sending data to annotation or QA tasks.

Node Library

The Node Library in Dataloop's Pipeline offers a comprehensive list of available nodes, including both default and installed options. You can easily search for specific nodes using the search bar to help build your pipeline. Additionally, the Node Library allows you to add new model nodes and application nodes, from the Marketplace by clicking on the Plus (+) icon.

By default, the following node categories are available on the Dataloop's Node Library:

• Data Nodes
• Workflow Nodes
• Automation Nodes
• Model Nodes
• Image Utils Nodes

• The Search nodes field allows you to search for nodes based on their category and name.
• Each node elements will serve a unique support to the pipeline to accomplish its task. Once you place a node on the canvas and clicking it will bring up the node’s Config in the right-side panel.