# YOLO26n performance review

## Measured revisions

| | Reference | Current |
| --- | --- | --- |
| Source revision | `<commit>` | `<commit>` |
| Model manifest SHA-256 | `<sha256>` | `<sha256>` |
| Runtime library SHA-256 | `<sha256>` | `<sha256>` |
| Dataset name | `<dataset>` | `<dataset>` |
| Dataset source-tree identity | `<identity>` | `<identity>` |
| Dataset configuration SHA-256 | `<sha256>` | `<sha256>` |
| Dataset images | `<count>` | `<count>` |
| Dataset instances | `<count>` | `<count>` |
| Ultralytics arguments | `<arguments>` | `<arguments>` |

## Ultralytics results

| Measurement | Reference | Current |
| --- | ---: | ---: |
| Precision | `<value>` | `<value>` |
| Recall | `<value>` | `<value>` |
| mAP50 | `<value>` | `<value>` |
| mAP50-95 | `<value>` | `<value>` |
| Preprocess | `<milliseconds>` | `<milliseconds>` |
| Inference | `<milliseconds>` | `<milliseconds>` |
| Postprocess | `<milliseconds>` | `<milliseconds>` |

- Reference evidence: `<path or link>`
- Current evidence: `<path or link>`

## Measurement context

Summarize conditions that materially help interpret the comparison, including
throttling, temperature, available memory, filesystem cache, major page faults,
and swap activity. Link to both `results.json` files for the complete record.

## Review

State whether the observed accuracy and performance are acceptable for this
change, including any qualifications.