chronax.models.dilated_rnn.dilated_rnn

`Scaler`

chronax.dilated_rnn_scaler.Scaler · inherits Protocol

Per-window scaler. Implementations are pure.

`stats(self, x, axis=1, mask=None)`

Parameter	Type	Default	Description
`x`	`jnp.ndarray`	-	(undocumented)
`axis`	`int`	`1`	(undocumented)
`mask`	`jnp.ndarray \\| None`	`None`	(undocumented)

Returns: tuple[jnp.ndarray, jnp.ndarray] (undocumented)

`transform(self, x, shift, scale)`

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

Returns: jnp.ndarray (undocumented)

`inverse(self, z, shift, scale)`

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

Returns: jnp.ndarray (undocumented)

`IdentityScaler`

chronax.dilated_rnn_scaler.IdentityScaler

No-op scaler — the scaler_type="identity" path.

`stats(self, x, axis=1, mask=None)`

Parameter	Type	Default	Description
`x`	`jnp.ndarray`	-	(undocumented)
`axis`	`int`	`1`	(undocumented)
`mask`	`jnp.ndarray \\| None`	`None`	(undocumented)

Returns: tuple[jnp.ndarray, jnp.ndarray] (undocumented)

`transform(self, x, shift, scale)`

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

Returns: jnp.ndarray (undocumented)

`inverse(self, z, shift, scale)`

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

Returns: jnp.ndarray (undocumented)

`StandardScaler`

chronax.dilated_rnn_scaler.StandardScaler

Mean / population-std scaler — the scaler_type="standard" path.

`stats(self, x, axis=1, mask=None)`

Parameter	Type	Default	Description
`x`	`jnp.ndarray`	-	(undocumented)
`axis`	`int`	`1`	(undocumented)
`mask`	`jnp.ndarray \\| None`	`None`	(undocumented)

Returns: tuple[jnp.ndarray, jnp.ndarray] (undocumented)

`transform(self, x, shift, scale)`

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

Returns: jnp.ndarray (undocumented)

`inverse(self, z, shift, scale)`

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

Returns: jnp.ndarray (undocumented)

`RobustScaler`

chronax.dilated_rnn_scaler.RobustScaler

Median + MAD scaler with 0.6745*std fallback when MAD=0. Mirrors NF robust_statistics: the fallback uses the Gaussian relationship MAD ~= 0.6745*sigma to estimate MAD when the empirical MAD degenerates to zero (e.g. a near-constant window), then forces any remaining zero to 1.0 and adds eps.

`stats(self, x, axis=1, mask=None)`

Parameter	Type	Default	Description
`x`	`jnp.ndarray`	-	(undocumented)
`axis`	`int`	`1`	(undocumented)
`mask`	`jnp.ndarray \\| None`	`None`	(undocumented)

Returns: tuple[jnp.ndarray, jnp.ndarray] (undocumented)

`transform(self, x, shift, scale)`

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

Returns: jnp.ndarray (undocumented)

`inverse(self, z, shift, scale)`

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

Returns: jnp.ndarray (undocumented)

`resolve`

chronax.dilated_rnn_scaler.resolve

Return a Scaler instance from a registry name (or pass an instance through).

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

Returns: Scaler (undocumented) Raises: ValueError