Visualization#

Visualization supports single model evaluation results and multi-model comparison, as well as visualization of mixed dataset evaluations.

Install Dependencies#

Install the dependencies required for visualization, including gradio, plotly, etc.

pip install 'evalscope[app]'

Note

Visualization requires evalscope version greater than or equal to 0.10.0. If the version is less than 0.10.0, please upgrade evalscope first.

Start Visualization Service#

Run the following command to start the visualization service.

evalscope app

You can access the visualization service in the browser if the following output appears.

* Running on local URL:  http://127.0.0.1:7861

To create a public link, set `share=True` in `launch()`.

Features Introduction#

  1. Configuration options on the left:

    • Root directory of evaluation reports

    • Selection of evaluation reports alt text

  2. Single Model Evaluation Results:

    • Evaluation Overview: Displays the composition of the evaluation dataset and the evaluation results alt text

    • Detailed evaluation of a single dataset, including model prediction results alt text

  3. Comparison of Multiple Model Evaluation Results:

    • Displayed using radar charts and comparison tables alt text

  4. Visualization of Mixed Dataset Evaluations:

    • Visual representation based on model capability dimensions alt text