mae
chronax.tcn_losses.mae
Mean absolute error.
| Parameter | Type | Default | Description |
|---|---|---|---|
pred |
jnp.ndarray |
- | (undocumented) |
target |
jnp.ndarray |
- | (undocumented) |
Returns: jnp.ndarray (The mean absolute error loss.)
mse
chronax.tcn_losses.mse
Mean squared error.
| Parameter | Type | Default | Description |
|---|---|---|---|
pred |
jnp.ndarray |
- | (undocumented) |
target |
jnp.ndarray |
- | (undocumented) |
Returns: jnp.ndarray (The mean squared error loss.)
huber
chronax.tcn_losses.huber
Huber loss, delta = 1.0.
| Parameter | Type | Default | Description |
|---|---|---|---|
pred |
jnp.ndarray |
- | (undocumented) |
target |
jnp.ndarray |
- | (undocumented) |
Returns: jnp.ndarray (The Huber loss.)
MultiQuantileLoss
chronax.tcn_losses.MultiQuantileLoss
Multi-quantile (pinball) loss. __call__(pred[...,h,Q], target[...,h]).
QL(y, y_hat, q) = q*(y-y_hat)+ + (1-q)*(y_hat-y)+, averaged over quantiles and all elements. Quantiles are sorted, must lie in (0, 1), and must include 0.5 (the median / "mean" head). Picklable (holds a plain tuple).
__init__(self, quantiles=(0.1, 0.5, 0.9))
Initializes the MultiQuantileLoss instance.
| Parameter | Type | Default | Description |
|---|---|---|---|
quantiles |
Sequence[float] |
(0.1, 0.5, 0.9) |
Quantiles to use for the loss calculation. Quantiles must be strictly between 0 and 1 and must include 0.5 (the median head). |
__call__(self, pred, target)
Calculates the multi-quantile pinball loss.
| Parameter | Type | Default | Description |
|---|---|---|---|
pred |
jnp.ndarray |
- | Prediction array, expected shape [..., h, Q], where Q is the number of quantiles. |
target |
jnp.ndarray |
- | Target array, expected shape [..., h]. |
Returns: jnp.ndarray (The scalar mean pinball loss.)
outputsize_multiplier
chronax.tcn_losses.outputsize_multiplier
Output-head width implied by loss (1 for point losses).
| Parameter | Type | Default | Description |
|---|---|---|---|
loss |
- | - | (undocumented) |
Returns: int (The implied output head width.)
resolve
chronax.tcn_losses.resolve
Return a callable loss from a registry string, a callable, or an MQ instance.
| Parameter | Type | Default | Description |
|---|---|---|---|
loss |
str \| LossFn \| MultiQuantileLoss |
- | (undocumented) |
Returns: (A callable loss function.)
Raises:
| Exception | Description |
|---|---|
ValueError |
If the provided string loss name is unknown. |