geoprior.params.LearnableQ

class geoprior.params.LearnableQ(initial_value=0.0, name=None, trainable=True, **kws)

Bases: BaseLearnable

Learnable Source/Sink term (Q).

Indicates that the PINN’s source/sink term \(Q\) should be learned (trainable) if TensorFlow is available; otherwise acts as a fixed NumPy‐based parameter. Unlike K and Ss, Q may be positive or negative, so we learn it directly (no log‐transform). The user supplies an initial_value, and the object initializes:

(1)\[Q \;=\; ext{initial\_value}.\]

Unconstrained: may be positive or negative.

Examples

>>> q = LearnableQ(0.0)
>>> q.get_value()
0.0

Parameters:

• initial_value (float)

• name (str | None)

• trainable (bool)

__init__(initial_value=0.0, name=None, trainable=True, **kws)

Parameters:

• initial_value (float)

• name (str | None)

• trainable (bool)

Methods

__init__([initial_value, name, trainable])
from_config(config)	Re-instantiate from get_config().
get_config()	Return a JSON-serialisable dict for tf.keras.
get_value()	Return raw \(Q\) value.

__init__(initial_value=0.0, name=None, trainable=True, **kws)

Parameters:

• initial_value (float)

• name (str | None)

• trainable (bool)

get_value()

Return raw \(Q\) value.

Returns:

Source/sink strength.

Return type:

Union[Tensor, float]