🏃‍♀️

Jump right in by Creating an Application and Uploading the Data

Use Case Background

This demo illustrates how to assess and track the performance of a multi-step E-commerce summarization application using the Deepchecks platform. In this demo, an electronics retailer employs a language model-based application to produce two separate outputs:

1. Concise, factually correct, and appealing summaries to be displayed as previews before accessing the product page.
2. Key metadata to facilitate advanced filtering and grouping of products.

These two outputs are generated and evaluated independently, and then combined to form a final JSON file.

Interaction Input - Output Example

Structure of this Example

• Start by Uploading the Data with Deepchecks' SDK or UI.
• Then Add your own LLM property and define custom config for automatic annotation.
• Utilize the Deepchecks' platform to compare different versions and select the optimal one.
• Lastly, monitor the production data in both summarization and feature extraction.