Loss functions for Chronax N-BEATS.
masked_mae
chronax.loss.masked_mae
Mean Absolute Error, optionally masked.
masked_mae(y, y_hat, mask=None)
| Parameter | Type | Default | Description |
|---|---|---|---|
y |
jnp.ndarray |
- | Target values [B, h] (or broadcastable). |
y_hat |
jnp.ndarray |
- | Predicted values, same shape as y. |
mask |
Optional[jnp.ndarray] |
None |
Float mask (1=include, 0=exclude); defaults to all-ones. |
Returns: jnp.ndarray (Scalar loss value.)
masked_mse
chronax.loss.masked_mse
Mean Squared Error, optionally masked.
masked_mse(y, y_hat, mask=None)
| Parameter | Type | Default | Description |
|---|---|---|---|
y |
jnp.ndarray |
- | (undocumented) |
y_hat |
jnp.ndarray |
- | (undocumented) |
mask |
Optional[jnp.ndarray] |
None |
(undocumented) |
Returns: jnp.ndarray
resolve
chronax.loss.resolve
Return a callable loss from either a registry string or a callable.
resolve(loss)
| Parameter | Type | Default | Description |
|---|---|---|---|
loss |
- | - | (undocumented) |
Returns: (undocumented)