# Squad

> Computes the SQuAD metric using torchmetrics, given predictions and ground truth. Use when evaluating question-answering outputs with exact match and F1 scores.

- Skill: `qhjqhj00/squad` (Agent Skill)
- Install (CLI): `npx skillmds add qhjqhj00/squad`
- Raw SKILL.md: https://api.skillmd.com/api/skills/qhjqhj00/squad/raw
- Safety review: PASS (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML, Data & Analytics, Model Training & Fine-tuning
- Tags: Evaluation, Exact Match, F1, Question Answering, Squad, Torchmetrics
- Author: qhjqhj00 (https://skillmd.com/u/qhjqhj00)
- Updated: 2026-08-22
- Page: https://skillmd.com/skills/qhjqhj00/squad

---


# squad

> Metric `SQuAD` from `torchmetrics` (torchmetrics.SQuAD)

## When to invoke this skill

The user has predictions + ground truth and asks to evaluate with SQuAD, or
mentions `torchmetrics.SQuAD` directly, or wants the standard torchmetrics implementation.

## Reference signature

```python
from torchmetrics import SQuAD

# _SQuAD(**kwargs: Any) -> None
```

## Library docstring

```
Wrapper for deprecated import.

>>> preds = [{"prediction_text": "1976", "id": "56e10a3be3433e1400422b22"}]
>>> target = [{"answers": {"answer_start": [97], "text": ["1976"]}, "id": "56e10a3be3433e1400422b22"}]
>>> squad = _SQuAD()
>>> squad(preds, target)
{'exact_match': tensor(100.), 'f1': tensor(100.)}
```

## Quick recipe

```python
import torchmetrics as _m
score = _m.SQuAD(y_true, y_pred)
```

## Don'ts

- Don't reimplement when the library version handles edge cases (NaN, ties, empty inputs) better than a hand-rolled formula.
- Always check the library version's argument order — sklearn is `(y_true, y_pred)` while torchmetrics is `(preds, target)`.