chronax.models.mlp.mlp

`Scaler`

mlp_scaler.Scaler

(No summary provided.)

`stats(self, x: jnp.ndarray, axis: int = 1) -> tuple[jnp.ndarray, jnp.ndarray]`

(No summary provided.)

Parameter	Type	Default	Description
`x`	`jnp.ndarray`	-	(undocumented)
`axis`	`int`	`1`	(undocumented)

`transform(self, x: jnp.ndarray, shift: jnp.ndarray, scale: jnp.ndarray) -> jnp.ndarray`

(No summary provided.)

Parameter	Type	Default	Description
`x`	`jnp.ndarray`	-	(undocumented)
`shift`	`jnp.ndarray`	-	(undocumented)
`scale`	`jnp.ndarray`	-	(undocumented)

`inverse(self, z: jnp.ndarray, shift: jnp.ndarray, scale: jnp.ndarray) -> jnp.ndarray`

(No summary provided.)

Parameter	Type	Default	Description
`z`	`jnp.ndarray`	-	(undocumented)
`shift`	`jnp.ndarray`	-	(undocumented)
`scale`	`jnp.ndarray`	-	(undocumented)

`IdentityScaler`

mlp_scaler.IdentityScaler

No-op scaler (shift=0, scale=1). Matches neuralforecast's scaler_type='identity'.

`stats(self, x: jnp.ndarray, axis: int = 1) -> tuple[jnp.ndarray, jnp.ndarray]`

(No summary provided.)

Parameter	Type	Default	Description
`x`	`jnp.ndarray`	-	(undocumented)
`axis`	`int`	`1`	(undocumented)

`transform(self, x: jnp.ndarray, shift: jnp.ndarray, scale: jnp.ndarray) -> jnp.ndarray`

(No summary provided.)

Parameter	Type	Default	Description
`x`	`jnp.ndarray`	-	(undocumented)
`shift`	`jnp.ndarray`	-	(undocumented)
`scale`	`jnp.ndarray`	-	(undocumented)

`inverse(self, z: jnp.ndarray, shift: jnp.ndarray, scale: jnp.ndarray) -> jnp.ndarray`

(No summary provided.)

Parameter	Type	Default	Description
`z`	`jnp.ndarray`	-	(undocumented)
`shift`	`jnp.ndarray`	-	(undocumented)
`scale`	`jnp.ndarray`	-	(undocumented)

`RobustScaler`

mlp_scaler.RobustScaler

Median + MAD scaler with a 0.6745*std fallback when MAD is zero.

The shift is the window median and the scale is the median absolute deviation median(|x - median|). Where MAD is zero the scale falls back to 0.6745*std; exact zeros are then pinned to 1.0 and eps is added. Mirrors neuralforecast's robust_statistics.

`stats(self, x: jnp.ndarray, axis: int = 1) -> tuple[jnp.ndarray, jnp.ndarray]`

(No summary provided.)

Parameter	Type	Default	Description
`x`	`jnp.ndarray`	-	(undocumented)
`axis`	`int`	`1`	(undocumented)

`transform(self, x: jnp.ndarray, shift: jnp.ndarray, scale: jnp.ndarray) -> jnp.ndarray`

(No summary provided.)

Parameter	Type	Default	Description
`x`	`jnp.ndarray`	-	(undocumented)
`shift`	`jnp.ndarray`	-	(undocumented)
`scale`	`jnp.ndarray`	-	(undocumented)

`inverse(self, z: jnp.ndarray, shift: jnp.ndarray, scale: jnp.ndarray) -> jnp.ndarray`

(No summary provided.)

Parameter	Type	Default	Description
`z`	`jnp.ndarray`	-	(undocumented)
`shift`	`jnp.ndarray`	-	(undocumented)
`scale`	`jnp.ndarray`	-	(undocumented)

`resolve_scaler`

mlp_scaler.resolve_scaler

Resolve a scaler from a name ('identity'/'robust') or a Scaler instance.

Parameter	Type	Default	Description
`scaler`	`str \\| Scaler`	-	(undocumented)

Returns: Scaler (undocumented). Raises: ValueError.