batched metric was not aggregated properly

[f14-bertolotti]

I’m not entirely sure this is the ideal approach, but here’s what I found:

batched_metrics returns a structure like:

{
  "metric_name": [1, 2, 3, 4]
}

Because of this additional dictionary layer, the aggregation step was throwing an IndexError. The issue can be resolved by extracting the metric name and its corresponding batch directly, rather than operating on the dictionary as-is.

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[Copilot]

Pull Request Overview

This PR attempts to fix an IndexError in batched metric aggregation by restructuring how metric outputs are indexed. The change modifies the loop to extract metric names and values from the batched outputs dictionary structure.

Key changes:

• Modified the batched metric aggregation loop to zip metrics with their outputs
• Changed indexing to extract values using metric name as dictionary key before accessing by index

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.