1 coral_random_loss

rri_metrics.coral.coral_random_loss(num_classes)

Expected CORAL loss for a random classifier.

For random logits, each threshold probability is 0.5, yielding a binary cross-entropy of log(2) per threshold.

1.1 Parameters

Name	Type	Description	Default
num_classes	int	Number of ordinal classes `K`.	required

1.2 Returns

Name	Type	Description
	float	Expected CORAL loss for uniform random predictions.

# coral_random_loss { #aria_nbv.rri_metrics.coral.coral_random_loss }

```python
rri_metrics.coral.coral_random_loss(num_classes)
```

Expected CORAL loss for a random classifier.

For random logits, each threshold probability is 0.5, yielding a
binary cross-entropy of log(2) per threshold.

## Parameters {.doc-section .doc-section-parameters}

| Name        | Type   | Description                      | Default    |
|-------------|--------|----------------------------------|------------|
| num_classes | int    | Number of ordinal classes ``K``. | _required_ |

## Returns {.doc-section .doc-section-returns}

| Name   | Type   | Description                                         |
|--------|--------|-----------------------------------------------------|
|        | float  | Expected CORAL loss for uniform random predictions. |