D Matrix Dmx Perplexity

---
name: d-matrix-dmx-perplexity
description: Compute d-matrix/dmx_perplexity via the HuggingFace `evaluate` library. Use when the user has predictions + references and wants the canonical implementation of d-matrix/dmx_perplexity.
metadata:
  skill_kind: metric
  source_lib: huggingface-evaluate
  hf_module: d-matrix/dmx_perplexity
  source: library_introspection
---

# d-matrix-dmx-perplexity

> Metric `d-matrix/dmx_perplexity` from the HuggingFace `evaluate` library.

## When to invoke

User asks to compute `d-matrix/dmx_perplexity` or wants HF evaluate's canonical version.

## Recipe

```python
import evaluate
metric = evaluate.load("d-matrix/dmx_perplexity")
result = metric.compute(predictions=preds, references=refs)
print(result)
```

## Don'ts

- Don't assume your in-house `d-matrix/dmx_perplexity` matches HF — version conventions vary.
- Many evaluate metrics have task-specific arguments (`average=`, `lang=`, `model_type=`); read the metric card before reporting numbers.