This PR aims to add error handling to the evaluation process to prevent errors from affecting the results and also add a condition to avoid the error for Accuracy Metric. Additionally, it introduces a "failed count" metadata field into the summary trace, which will help us easily track the number of successful and failed traces.
This PR aims to add error handling to the evaluation process to prevent errors from affecting the results and also add a condition to avoid the error for Accuracy Metric. Additionally, it introduces a "failed count" metadata field into the summary trace, which will help us easily track the number of successful and failed traces.