hgq.layers package

Subpackages

Submodules

hgq.layers.activation module

class hgq.layers.activation.QUnaryFunctionLUT(*args, **kwargs)

Bases: Activation, QLayerBaseSingleInput

call(inputs, training=None)
get_config()

Returns the config of the object.

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

hgq.layers.batch_normalization module

class hgq.layers.batch_normalization.QBatchNormalization(*args, **kwargs)

Bases: QLayerBaseSingleInput, BatchNormalization

property beta
property bq
build(input_shape)
call(inputs, training=None, mask=None)
property gamma
get_config()

Returns the config of the object.

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

property kq
property qscaler_and_qoffset

hgq.layers.conv module

class hgq.layers.conv.QBaseConv(*args, **kwargs)

Bases: QLayerBaseSingleInput, BaseConv

property bq
build(input_shape)
call(inputs, training=None)
get_config()

Returns the config of the object.

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

property kq
property qbias
property qkernel
class hgq.layers.conv.QConv1D(*args, **kwargs)

Bases: QBaseConv

call(inputs, training=None)
get_config()

Returns the config of the object.

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

class hgq.layers.conv.QConv2D(*args, **kwargs)

Bases: QBaseConv

get_config()

Returns the config of the object.

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

class hgq.layers.conv.QConv3D(*args, **kwargs)

Bases: QBaseConv

get_config()

Returns the config of the object.

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

hgq.layers.einsum_dense_batchnorm module

class hgq.layers.einsum_dense_batchnorm.QEinsumDenseBatchnorm(*args, **kwargs)

Bases: QEinsumDense

property bq
build(input_shape)
call(inputs, training=None)
get_config()

Returns the config of the object.

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

get_fused_qkernel_and_qbias(training, mean, var)
property kq
property qbias
property qkernel

hgq.layers.linformer_attention module

class hgq.layers.linformer_attention.QLinformerAttention(*args, **kwargs)

Bases: QMultiHeadAttention

build(query_shape, value_shape, key_shape=None)

Builds layers and variables.

Parameters:

  • query_shape (tuple) – Shape of the query tensor.

  • value_shape (tuple) – Shape of the value tensor.

  • key_shape (tuple, optional) – Shape of the key tensor.

call(query, value, key=None, query_mask=None, value_mask=None, key_mask=None, attention_mask=None, return_attention_scores=False, training=None, use_causal_mask=False)
property ebops: int
get_config()

Returns the config of the object.

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

hgq.layers.multi_head_attention module

class hgq.layers.multi_head_attention.QMultiHeadAttention(*args, **kwargs)

Bases: MultiHeadAttention, QLayerBase

build(query_shape, value_shape, key_shape=None)

Builds layers and variables.

Parameters:

  • query_shape (tuple) – Shape of the query tensor.

  • value_shape (tuple) – Shape of the value tensor.

  • key_shape (tuple, optional) – Shape of the key tensor.

call(query, value, key=None, query_mask=None, value_mask=None, key_mask=None, attention_mask=None, return_attention_scores=False, training=None, use_causal_mask=False)
compute_output_shape(query_shape, value_shape, key_shape=None)
property ebops: int
get_config()

Returns the config of the object.

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

hgq.layers.pooling module

class hgq.layers.pooling.QAveragePooling1D(*args, **kwargs)

Bases: QBasePooling, AveragePooling1D

get_config()

Returns the config of the object.

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

class hgq.layers.pooling.QAveragePooling2D(*args, **kwargs)

Bases: QBasePooling, AveragePooling2D

get_config()

Returns the config of the object.

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

class hgq.layers.pooling.QAveragePooling3D(*args, **kwargs)

Bases: QBasePooling, AveragePooling3D

get_config()

Returns the config of the object.

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

class hgq.layers.pooling.QGlobalAveragePooling1D(*args, **kwargs)

Bases: QBasePooling, GlobalAveragePooling1D

get_config()

Returns the config of the object.

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

class hgq.layers.pooling.QGlobalAveragePooling2D(*args, **kwargs)

Bases: QBasePooling, GlobalAveragePooling2D

get_config()

Returns the config of the object.

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

class hgq.layers.pooling.QGlobalAveragePooling3D(*args, **kwargs)

Bases: QBasePooling, GlobalAveragePooling3D

get_config()

Returns the config of the object.

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

class hgq.layers.pooling.QGlobalMaxPooling1D(*args, **kwargs)

Bases: QBasePooling, GlobalMaxPooling1D

get_config()

Returns the config of the object.

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

class hgq.layers.pooling.QGlobalMaxPooling2D(*args, **kwargs)

Bases: QBasePooling, GlobalMaxPooling2D

get_config()

Returns the config of the object.

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

class hgq.layers.pooling.QGlobalMaxPooling3D(*args, **kwargs)

Bases: QBasePooling, GlobalMaxPooling3D

get_config()

Returns the config of the object.

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

class hgq.layers.pooling.QMaxPooling1D(*args, **kwargs)

Bases: QBasePooling, MaxPooling1D

get_config()

Returns the config of the object.

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

class hgq.layers.pooling.QMaxPooling2D(*args, **kwargs)

Bases: QBasePooling, MaxPooling2D

get_config()

Returns the config of the object.

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

class hgq.layers.pooling.QMaxPooling3D(*args, **kwargs)

Bases: QBasePooling, MaxPooling3D

get_config()

Returns the config of the object.

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

hgq.layers.softmax module

class hgq.layers.softmax.QSoftmax(*args, **kwargs)

Bases: QLayerBaseSingleInput

build(input_shape)
call(inputs, training=None, mask=None)
compute_output_shape(input_shape)
property ebops
get_config()

Returns the config of the object.

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

Module contents

class hgq.layers.QAdd(*args, **kwargs)

Bases: QMerge, Add

get_config()

Returns the config of the object.

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

class hgq.layers.QAveragePooling1D(*args, **kwargs)

Bases: QBasePooling, AveragePooling1D

get_config()

Returns the config of the object.

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

class hgq.layers.QAveragePooling2D(*args, **kwargs)

Bases: QBasePooling, AveragePooling2D

get_config()

Returns the config of the object.

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

class hgq.layers.QAveragePooling3D(*args, **kwargs)

Bases: QBasePooling, AveragePooling3D

get_config()

Returns the config of the object.

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

class hgq.layers.QAveragePow2(*args, **kwargs)

Bases: QAdd, Average

build(input_shape)
get_config()

Returns the config of the object.

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

hgq.layers.QAvgPool1D

alias of QAveragePooling1D

hgq.layers.QAvgPool2D

alias of QAveragePooling2D

hgq.layers.QAvgPool3D

alias of QAveragePooling3D

class hgq.layers.QBatchNormDense(*args, **kwargs)

Bases: QDense

build(input_shape)
call(inputs, training=None)
get_config()

Returns the config of the object.

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

get_fused_qkernel_and_qbias(mean, var)
property qbias
property qkernel
class hgq.layers.QBatchNormalization(*args, **kwargs)

Bases: QLayerBaseSingleInput, BatchNormalization

property beta
property bq
build(input_shape)
call(inputs, training=None, mask=None)
property gamma
get_config()

Returns the config of the object.

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

property kq
property qscaler_and_qoffset
class hgq.layers.QConv1D(*args, **kwargs)

Bases: QBaseConv

call(inputs, training=None)
get_config()

Returns the config of the object.

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

class hgq.layers.QConv2D(*args, **kwargs)

Bases: QBaseConv

get_config()

Returns the config of the object.

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

class hgq.layers.QConv3D(*args, **kwargs)

Bases: QBaseConv

get_config()

Returns the config of the object.

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

class hgq.layers.QDense(*args, **kwargs)

Bases: QLayerBaseSingleInput, Dense

property bq
build(input_shape)
call(inputs, training=None)
get_config()

Returns the config of the object.

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

property kq
property qbias
property qkernel
class hgq.layers.QDenseT(*args, **kwargs)

Bases: QLayerBaseSingleInput

build(input_shape)
call(x, training=None)
get_config()

Returns the config of the object.

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

property toq
class hgq.layers.QDot(*args, **kwargs)

Bases: QMerge, Dot

build(input_shape)
get_config()

Returns the config of the object.

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

class hgq.layers.QEinsum(*args, **kwargs)

Bases: QLayerBaseMultiInputs

build(input_shape)
call(inputs, training=None)
get_config()

Returns the config of the object.

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

class hgq.layers.QEinsumDense(*args, **kwargs)

Bases: QLayerBaseSingleInput, EinsumDense

property bq
build(input_shape)
call(inputs, training=None)
get_config()

Returns the config of the object.

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

property kq
property qbias
property qkernel
class hgq.layers.QEinsumDenseBatchnorm(*args, **kwargs)

Bases: QEinsumDense

property bq
build(input_shape)
call(inputs, training=None)
get_config()

Returns the config of the object.

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

get_fused_qkernel_and_qbias(training, mean, var)
property kq
property qbias
property qkernel
class hgq.layers.QGRU(*args, **kwargs)

Bases: QRNN, GRU

Gated Recurrent Unit - Cho et al. 2014.

The QGRU only allows the backend native implementation (no CuDNN kernel). When the jax backend is used, if any WRAP quantizers are used, unroll will be set to True to avoid the side effect issue in the jax.lax.scan loop.

Parameters:

  • units (int) – Positive integer, dimensionality of the output space.

  • activation (str, optional) – Activation function to use. Default: linear, effectively hard_tanh by the pre-activation quantizer.

  • recurrent_activation (str, optional) – Activation function to use for the recurrent step. Default: linear, effectively hard_sigmoid (slope=0.5) by the pre-activation quantizer.

  • use_bias (bool, optional) – Whether the layer should use a bias vector. Default: True.

  • kernel_initializer (str, optional) – Initializer for the kernel weights matrix, used for the linear transformation of the inputs. Default: “glorot_uniform”.

  • recurrent_initializer (str, optional) – Initializer for the recurrent_kernel weights matrix, used for the linear transformation of the recurrent state. Default: “orthogonal”.

  • bias_initializer (str, optional) – Initializer for the bias vector. Default: “zeros”.

  • kernel_regularizer (optional) – Regularizer function applied to the kernel weights matrix. Default: None.

  • recurrent_regularizer (optional) – Regularizer function applied to the recurrent_kernel weights matrix. Default: None.

  • bias_regularizer (optional) – Regularizer function applied to the bias vector. Default: None.

  • activity_regularizer (optional) – Regularizer function applied to the output of the layer (its “activation”). Default: None.

  • kernel_constraint (optional) – Constraint function applied to the kernel weights matrix. Default: None.

  • recurrent_constraint (optional) – Constraint function applied to the recurrent_kernel weights matrix. Default: None.

  • bias_constraint (optional) – Constraint function applied to the bias vector. Default: None.

  • dropout (float, optional) – Float between 0 and 1. Fraction of the units to drop for the linear transformation of the inputs. Default: 0.

  • recurrent_dropout (float, optional) – Float between 0 and 1. Fraction of the units to drop for the linear transformation of the recurrent state. Default: 0.

  • seed (int, optional) – Random seed for dropout.

  • return_sequences (bool, optional) – Whether to return the last output in the output sequence, or the full sequence. Default: False.

  • return_state (bool, optional) – Whether to return the last state in addition to the output. Default: False.

  • go_backwards (bool, optional) – If True, process the input sequence backwards and return the reversed sequence. Default: False.

  • stateful (bool, optional) – If True, the last state for each sample at index i in a batch will be used as initial state for the sample of index i in the following batch. Default: False.

  • unroll (bool or None, optional) – None is equivalent to False. However, for the JAX backend, if any WRAP quantizers are used, unroll will be set to True to avoid the side effect issue in the jax.lax.scan loop. If True, the network will be unrolled, else a symbolic loop will be used. Unrolling can speed-up a RNN, although it tends to be more memory-intensive. Unrolling is only suitable for short sequences. Default: None.

  • reset_after (bool, optional) – GRU convention (whether to apply reset gate after or before matrix multiplication). False is “before”, True is “after” (default and cuDNN compatible). Default: True.

  • iq_conf (QuantizerConfig or None, optional) – Quantizer configuration for input quantizer. Default: None (global default)

  • paq_conf (QuantizerConfig or None, optional) – Quantizer configuration for post-activation quantizer. Default: None (hard tanh like, w/ global default)

  • praq_conf (QuantizerConfig or None, optional) – Quantizer configuration for pre-recurrent activation quantizer. Default: None (hard sigmoid like, w/ global default)

  • sq_conf (QuantizerConfig or None, optional) – Quantizer configuration for state quantizer. Default: None (global default)

  • kq_conf (QuantizerConfig or None, optional) – Quantizer configuration for kernel quantizer. Default: None (global default)

  • rkq_conf (QuantizerConfig or None, optional) – Quantizer configuration for recurrent kernel quantizer. Default: None (global default)

  • bq_conf (QuantizerConfig or None, optional) – Quantizer configuration for bias quantizer. Default: None (global default)

  • oq_conf (QuantizerConfig or None, optional) – Quantizer configuration for output quantizer. Default: None (global default)

  • rhq_conf (QuantizerConfig or None, optional) – Quantizer configuration for recurrent hidden state quantizer. Default: None (global default)

  • parallelization_factor (int, optional) – Factor for parallelization. Default: 1.

  • enable_oq (bool or None, optional) – Whether to enable output quantizer. Default: None (global default)

  • enable_iq (bool or None, optional) – Whether to enable input quantizer. Default: None (global default)

  • enable_ebops (bool or None, optional) – Whether to enable energy-efficient bit operations. Default: None (global default)

  • beta0 (float or None, optional) – Beta0 parameter for quantizer. Default: None (global default)

  • enable_ebops – Whether to enable EBOPs resource consumption estimation. Default: None (global default).

  • parallelization_factor – Number of cells to be computed in parallel. Default: 1.

Notes

inputsarray_like

A 3D tensor, with shape (batch, timesteps, feature).

maskarray_like, optional

Binary tensor of shape (samples, timesteps) indicating whether a given timestep should be masked (optional). An individual True entry indicates that the corresponding timestep should be utilized, while a False entry indicates that the corresponding timestep should be ignored. Defaults to None.

trainingbool, optional

Python boolean indicating whether the layer should behave in training mode or in inference mode. This argument is passed to the cell when calling it. This is only relevant if dropout or recurrent_dropout is used (optional). Defaults to None.

initial_statelist, optional

List of initial state tensors to be passed to the first call of the cell (optional, None causes creation of zero-filled initial state tensors). Defaults to None.

get_config()

Returns the config of the object.

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

class hgq.layers.QGlobalAveragePooling1D(*args, **kwargs)

Bases: QBasePooling, GlobalAveragePooling1D

get_config()

Returns the config of the object.

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

class hgq.layers.QGlobalAveragePooling2D(*args, **kwargs)

Bases: QBasePooling, GlobalAveragePooling2D

get_config()

Returns the config of the object.

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

class hgq.layers.QGlobalAveragePooling3D(*args, **kwargs)

Bases: QBasePooling, GlobalAveragePooling3D

get_config()

Returns the config of the object.

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

hgq.layers.QGlobalAvgPool1D

alias of QGlobalAveragePooling1D

hgq.layers.QGlobalAvgPool2D

alias of QGlobalAveragePooling2D

hgq.layers.QGlobalAvgPool3D

alias of QGlobalAveragePooling3D

hgq.layers.QGlobalMaxPool1D

alias of QGlobalMaxPooling1D

hgq.layers.QGlobalMaxPool2D

alias of QGlobalMaxPooling2D

hgq.layers.QGlobalMaxPool3D

alias of QGlobalMaxPooling3D

class hgq.layers.QGlobalMaxPooling1D(*args, **kwargs)

Bases: QBasePooling, GlobalMaxPooling1D

get_config()

Returns the config of the object.

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

class hgq.layers.QGlobalMaxPooling2D(*args, **kwargs)

Bases: QBasePooling, GlobalMaxPooling2D

get_config()

Returns the config of the object.

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

class hgq.layers.QGlobalMaxPooling3D(*args, **kwargs)

Bases: QBasePooling, GlobalMaxPooling3D

get_config()

Returns the config of the object.

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

class hgq.layers.QLinformerAttention(*args, **kwargs)

Bases: QMultiHeadAttention

build(query_shape, value_shape, key_shape=None)

Builds layers and variables.

Parameters:

  • query_shape (tuple) – Shape of the query tensor.

  • value_shape (tuple) – Shape of the value tensor.

  • key_shape (tuple, optional) – Shape of the key tensor.

call(query, value, key=None, query_mask=None, value_mask=None, key_mask=None, attention_mask=None, return_attention_scores=False, training=None, use_causal_mask=False)
property ebops: int
get_config()

Returns the config of the object.

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

hgq.layers.QMaxPool1D

alias of QMaxPooling1D

hgq.layers.QMaxPool2D

alias of QMaxPooling2D

hgq.layers.QMaxPool3D

alias of QMaxPooling3D

class hgq.layers.QMaxPooling1D(*args, **kwargs)

Bases: QBasePooling, MaxPooling1D

get_config()

Returns the config of the object.

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

class hgq.layers.QMaxPooling2D(*args, **kwargs)

Bases: QBasePooling, MaxPooling2D

get_config()

Returns the config of the object.

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

class hgq.layers.QMaxPooling3D(*args, **kwargs)

Bases: QBasePooling, MaxPooling3D

get_config()

Returns the config of the object.

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

class hgq.layers.QMaximum(*args, **kwargs)

Bases: QMerge, Maximum

get_config()

Returns the config of the object.

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

class hgq.layers.QMeanPow2(*args, **kwargs)

Bases: QSum

build(input_shape)
get_config()

Returns the config of the object.

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

class hgq.layers.QMinimum(*args, **kwargs)

Bases: QMerge, Minimum

get_config()

Returns the config of the object.

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

class hgq.layers.QMultiHeadAttention(*args, **kwargs)

Bases: MultiHeadAttention, QLayerBase

build(query_shape, value_shape, key_shape=None)

Builds layers and variables.

Parameters:

  • query_shape (tuple) – Shape of the query tensor.

  • value_shape (tuple) – Shape of the value tensor.

  • key_shape (tuple, optional) – Shape of the key tensor.

call(query, value, key=None, query_mask=None, value_mask=None, key_mask=None, attention_mask=None, return_attention_scores=False, training=None, use_causal_mask=False)
compute_output_shape(query_shape, value_shape, key_shape=None)
property ebops: int
get_config()

Returns the config of the object.

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

class hgq.layers.QMultiply(*args, **kwargs)

Bases: QMerge, Multiply

get_config()

Returns the config of the object.

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

class hgq.layers.QSimpleRNN(*args, **kwargs)

Bases: QRNN, SimpleRNN

Quantized Fully-connected RNN where the output is to be fed back as the new input.

When the jax backend is used, if any WRAP quantizers are used, unroll will be set to True to avoid the side effect issue in the jax.lax.scan loop.

Parameters:

  • units (int) – Positive integer, dimensionality of the output space.

  • activation (str, optional) – Activation function to use. Default: linear, effectively hard_tanh by the pre-activation quantizer.

  • use_bias (bool, optional) – Whether the layer uses a bias vector. Default: True.

  • kernel_initializer (str, optional) – Initializer for the kernel weights matrix, used for the linear transformation of the inputs. Default: “glorot_uniform”.

  • recurrent_initializer (str, optional) – Initializer for the recurrent_kernel weights matrix, used for the linear transformation of the recurrent state. Default: “orthogonal”.

  • bias_initializer (str, optional) – Initializer for the bias vector. Default: “zeros”.

  • kernel_regularizer (optional) – Regularizer function applied to the kernel weights matrix. Default: None.

  • recurrent_regularizer (optional) – Regularizer function applied to the recurrent_kernel weights matrix. Default: None.

  • bias_regularizer (optional) – Regularizer function applied to the bias vector. Default: None.

  • activity_regularizer (optional) – Regularizer function applied to the output of the layer (its “activation”). Default: None.

  • kernel_constraint (optional) – Constraint function applied to the kernel weights matrix. Default: None.

  • recurrent_constraint (optional) – Constraint function applied to the recurrent_kernel weights matrix. Default: None.

  • bias_constraint (optional) – Constraint function applied to the bias vector. Default: None.

  • dropout (float, optional) – Float between 0 and 1. Fraction of the units to drop for the linear transformation of the inputs. Default: 0.

  • recurrent_dropout (float, optional) – Float between 0 and 1. Fraction of the units to drop for the linear transformation of the recurrent state. Default: 0.

  • return_sequences (bool, optional) – Whether to return the last output in the output sequence, or the full sequence. Default: False.

  • return_state (bool, optional) – Whether to return the last state in addition to the output. Default: False.

  • go_backwards (bool, optional) – If True, process the input sequence backwards and return the reversed sequence. Default: False.

  • stateful (bool, optional) – If True, the last state for each sample at index i in a batch will be used as initial state for the sample of index i in the following batch. Default: False.

  • unroll (bool or None, optional) – None is equivalent to False. However, for the JAX backend, if any WRAP quantizers are used, unroll will be set to True to avoid the side effect issue in the jax.lax.scan loop. If True, the network will be unrolled, else a symbolic loop will be used. Unrolling can speed-up a RNN, although it tends to be more memory-intensive. Unrolling is only suitable for short sequences. Default: None.

  • seed (int, optional) – Random seed for dropout.

  • iq_conf (QuantizerConfig or None, optional) – Quantizer configuration for input quantizer. Default: None.

  • sq_conf (QuantizerConfig or None, optional) – Quantizer configuration for state quantizer. Default: None.

  • kq_conf (QuantizerConfig or None, optional) – Quantizer configuration for kernel quantizer. Default: None.

  • rkq_conf (QuantizerConfig or None, optional) – Quantizer configuration for recurrent kernel quantizer. Default: None.

  • bq_conf (QuantizerConfig or None, optional) – Quantizer configuration for bias quantizer. Default: None.

  • oq_conf (QuantizerConfig or None, optional) – Quantizer configuration for output quantizer. Default: None.

  • paq_conf (QuantizerConfig or None, optional) – Quantizer configuration for pre-activation quantizer. Default: None.

  • parallelization_factor (int, optional) – Factor for parallelization. Default: -1 (automatic).

  • enable_sq (bool or None, optional) – Whether to enable state quantizer. When the output is already quantized and the state is 0-inited, the state quantizer should be disabled to avoid double quantization. Default: False.

  • enable_oq (bool or None, optional) – Whether to enable output quantizer. Default: None.

  • enable_iq (bool or None, optional) – Whether to enable input quantizer. Default: None.

  • enable_ebops (bool or None, optional) – Whether to enable energy-efficient bit operations. Default: None.

  • beta0 (float or None, optional) – Beta0 parameter for quantizer. Default: None.

Notes

sequencearray_like

A 3D tensor, with shape [batch, timesteps, feature].

maskarray_like, optional

Binary tensor of shape [batch, timesteps] indicating whether a given timestep should be masked. An individual True entry indicates that the corresponding timestep should be utilized, while a False entry indicates that the corresponding timestep should be ignored.

trainingbool, optional

Python boolean indicating whether the layer should behave in training mode or in inference mode. This argument is passed to the cell when calling it. This is only relevant if dropout or recurrent_dropout is used.

initial_statelist, optional

List of initial state tensors to be passed to the first call of the cell.

get_config()

Returns the config of the object.

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

class hgq.layers.QSoftmax(*args, **kwargs)

Bases: QLayerBaseSingleInput

build(input_shape)
call(inputs, training=None, mask=None)
compute_output_shape(input_shape)
property ebops
get_config()

Returns the config of the object.

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

class hgq.layers.QSubtract(*args, **kwargs)

Bases: QMerge, Subtract

get_config()

Returns the config of the object.

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

class hgq.layers.QSum(*args, **kwargs)

Bases: QLayerBaseSingleInput

build(input_shape)
call(inputs, training=None)
get_config()

Returns the config of the object.

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

property keepdims
property scale
class hgq.layers.QUnaryFunctionLUT(*args, **kwargs)

Bases: Activation, QLayerBaseSingleInput

call(inputs, training=None)
get_config()

Returns the config of the object.

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

class hgq.layers.Quantizer(*args, **kwargs)

Bases: Layer

The generic quantizer layer, wraps internal quantizers to provide a universal interface. Supports float, fixed-point (KBI, KIF) quantization. Can be initialized with a QuantizerConfig object or with the quantizer type and its parameters.

property bits
bits_(shape)
build(input_shape)
call(inputs, training=None)
epsilon_(shape)
property fbits
fbits_(shape)
classmethod from_config(config)

Creates an operation from its config.

This method is the reverse of get_config, capable of instantiating the same operation from the config dictionary.

Note: If you override this method, you might receive a serialized dtype config, which is a dict. You can deserialize it as follows:

```python if “dtype” in config and isinstance(config[“dtype”], dict):

policy = dtype_policies.deserialize(config[“dtype”])

```

Parameters:

config – A Python dictionary, typically the output of get_config.

Returns:

An operation instance.

get_config()

Returns the config of the object.

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

get_quantizer_config_kwargs(*args, **kwargs)
max_(shape)
min_(shape)
property q_type