Skip to main content

Completing Your First Run

Create an Run​

In EvalsOne, you can create and complete an evaluation run in just a few simple steps.

Before you start, make sure you have the following prepared:

  • Evaluation samples ready for testing
  • Configured models for generation and evaluation
  • If performing automated evaluations, set up the evaluation metrics to be used

But don't worry, we've got you covered. Every new user receives a quick test sample set, shared models, preset evaluation metrics, and an initial credit of $3, making it easy to get started.

Creating a Run

Create a Fork Run​

Based on an existing run, you can quickly iterate evaluations using the Fork function by clicking the "Fork" button. During a Fork iteration, you can retain the existing configuration or switch individual settings such as template versions, generation models, and evaluation metrics. This is ideal for continuous improvement, comprehensive consideration, comparative analysis, and optimization.

Creating a Fork Run

By now, you have learned the basic way to create a run on EvalsOne. However, every AI application's evaluation needs are unique. In practice, you will need to prepare samples, models, and evaluation metrics according to your specific needs and scenarios.

Next, you can delve deeper into the specific features of samples, models, and evaluation metrics, and step by step become an expert in prompt evaluation on EvalsOne!