Attention

class tfts.layers.attention_layer.Attention(*args, **kwargs)

Bases: Layer

Multi-head attention layer

Initialize the Attention layer.

Parameters:

hidden_sizeint

The number of hidden units, hidden_size = attention_dim_each_head x num_attention_heads.

num_attention_headsint

The number of attention heads.

attention_probs_dropout_probfloat, optional

Dropout rate for the attention weights. Defaults to 0.0.

inherited-members:

Methods

add_loss(loss)	Can be called inside of the call() method to add a scalar loss.
add_metric(*args, **kwargs)
add_variable(shape, initializer[, dtype, ...])	Add a weight variable to the layer.
add_weight([shape, initializer, dtype, ...])	Add a weight variable to the layer.
build(input_shape)
build_from_config(config)	Builds the layer's states with the supplied config dict.
call(q, k, v[, mask, past_key_value, ...])	use query and key generating an attention multiplier for value, multi_heads to repeat it
compute_mask(inputs, previous_mask)
compute_output_shape(input_shape)
compute_output_spec(*args, **kwargs)
count_params()	Count the total number of scalars composing the weights.
from_config(config)	Creates an operation from its config.
get_build_config()	Returns a dictionary with the layer's input shape.
get_config()	Returns the config of the object.
get_weights()	Return the values of layer.weights as a list of NumPy arrays.
load_own_variables(store)	Loads the state of the layer.
quantize(mode[, type_check])
quantized_build(input_shape, mode)
quantized_call(*args, **kwargs)
rematerialized_call(layer_call, *args, **kwargs)	Enable rematerialization dynamically for layer's call method.
save_own_variables(store)	Saves the state of the layer.
set_weights(weights)	Sets the values of layer.weights from a list of NumPy arrays.
stateless_call(trainable_variables, ...[, ...])	Call the layer without any side effects.
symbolic_call(*args, **kwargs)

Attributes

compute_dtype	The dtype of the computations performed by the layer.
dtype	Alias of layer.variable_dtype.
dtype_policy
input	Retrieves the input tensor(s) of a symbolic operation.
input_dtype	The dtype layer inputs should be converted to.
input_spec
losses	List of scalar losses from add_loss, regularizers and sublayers.
metrics	List of all metrics.
metrics_variables	List of all metric variables.
non_trainable_variables	List of all non-trainable layer state.
non_trainable_weights	List of all non-trainable weight variables of the layer.
output	Retrieves the output tensor(s) of a layer.
path	The path of the layer.
quantization_mode	The quantization mode of this layer, None if not quantized.
supports_masking	Whether this layer supports computing a mask using compute_mask.
trainable	Settable boolean, whether this layer should be trainable or not.
trainable_variables	List of all trainable layer state.
trainable_weights	List of all trainable weight variables of the layer.
variable_dtype	The dtype of the state (weights) of the layer.
variables	List of all layer state, including random seeds.
weights	List of all weight variables of the layer.

call(q: Tensor, k: Tensor, v: Tensor, mask: Tensor | None = None, past_key_value=None, training: bool | None = None, return_attention_scores: bool = False, use_causal_mask: bool = False, **kwargs)

use query and key generating an attention multiplier for value, multi_heads to repeat it

Parameters:

• q (tf.Tenor) – Query with shape batch * seq_q * fea

• k (tf.Tensor) – Key with shape batch * seq_k * fea

• v (tf.Tensor) – Value with shape batch * seq_v * fea

• mask (tf.Tensor, optional) – important to avoid the leaks, by default None

Returns:

Tensor with shape batch * seq_q * (units * num_attention_heads)

Return type:

tf.Tensor

get_config()

Returns the config of the object.

An object config is a Python dictionary (serializable) containing the information needed to re-instantiate it.