Index

_

__init__() (base_attentive.backend.TorchDeviceManager method), [1] (base_attentive.components.MultiDecoder method)

A

activation (base_attentive.config.schema.BaseAttentiveSpec attribute) apply_dtw (base_attentive.config.schema.BaseAttentiveSpec property) architecture (base_attentive.config.schema.BaseAttentiveSpec attribute)

attention_heads (base_attentive.config.schema.BaseAttentiveSpec attribute) attention_levels (base_attentive.config.schema.BaseAttentiveSpec property) attention_units (base_attentive.config.schema.BaseAttentiveSpec attribute)

B

Backend (class in base_attentive.backend) backend_context (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) backend_name (base_attentive.config.schema.BaseAttentiveSpec attribute) base_attentive module base_attentive.api.property module base_attentive.backend module base_attentive.runtime module base_attentive.validation module BaseAttentive (class in base_attentive) (class in base_attentive.core.base_attentive)

BaseAttentiveComponentSpec (class in base_attentive.config.schema) BaseAttentiveSpec (class in base_attentive.config.schema) BaseAttentiveV2Assembly (class in base_attentive.resolver.assembly) BatchNormalization (base_attentive.backend.Backend attribute) blockers (base_attentive.backend.Backend attribute) (base_attentive.backend.JaxBackend attribute) (base_attentive.backend.TorchBackend attribute) build() (base_attentive.components.MemoryAugmentedAttention method) (base_attentive.components.VariableSelectionNetwork method) build_component() (in module base_attentive.resolver.component_resolver) built-in function check_torch_compatibility() get_torch_device() get_torch_version() torch_is_available()

C

call() (base_attentive.components.CrossAttention method) (base_attentive.components.HierarchicalAttention method) (base_attentive.components.MemoryAugmentedAttention method) (base_attentive.components.MultiDecoder method), [1] (base_attentive.components.MultiScaleLSTM method), [1] (base_attentive.components.TransformerDecoderLayer method) (base_attentive.components.TransformerEncoderLayer method) (base_attentive.components.VariableSelectionNetwork method)

check_tensorflow_compatibility() (in module base_attentive.backend) check_torch_compatibility() built-in function check_torch_compatibility() (in module base_attentive.backend) clear_gpu_cache() (TorchDeviceManager method) clone() (base_attentive.registry.ComponentRegistry method) ComponentRegistry (class in base_attentive.registry) components (base_attentive.config.schema.BaseAttentiveSpec attribute) CrossAttention (class in base_attentive.components)

D

decoder_cross_attention (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) decoder_cross_postprocess (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) decoder_fusion (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) decoder_hierarchical_attention (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) decoder_input_projection (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) decoder_memory_attention (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) decoder_residual_add (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) decoder_residual_norm (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) Dense (base_attentive.backend.Backend attribute) dependency_message() (in module base_attentive)

detect_available_backends() (in module base_attentive.backend) device (base_attentive.backend.TorchDeviceManager property), [1] (TorchDeviceManager attribute) Dropout (base_attentive.backend.Backend attribute) dropout (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) dropout_rate (base_attentive.config.schema.BaseAttentiveSpec attribute) dynamic_encoder (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) dynamic_input_dim (base_attentive.config.schema.BaseAttentiveSpec attribute) dynamic_processor (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) dynamic_projection (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) dynamic_window (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute)

E

embed_dim (base_attentive.config.schema.BaseAttentiveSpec attribute) encoder_positional_encoding (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) ensure_bh1() (in module base_attentive.validation) ensure_default_backend() (in module base_attentive.backend)

experimental (base_attentive.backend.Backend attribute) (base_attentive.backend.JaxBackend attribute) (base_attentive.backend.TorchBackend attribute) extras (base_attentive.config.schema.BaseAttentiveSpec attribute)

F

final_agg (base_attentive.config.schema.BaseAttentiveSpec property) final_pool (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) final_pool_flatten (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) final_pool_last (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) final_pool_mean (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) final_residual_add (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) final_residual_norm (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) forecast_horizon (base_attentive.config.schema.BaseAttentiveSpec attribute) framework (base_attentive.backend.Backend attribute) (base_attentive.backend.JaxBackend attribute) (base_attentive.backend.TensorFlowBackend attribute) (base_attentive.backend.TorchBackend attribute) from_config() (base_attentive.BaseAttentive class method) (base_attentive.components.CrossAttention class method) (base_attentive.components.HierarchicalAttention class method) (base_attentive.components.MemoryAugmentedAttention class method) (base_attentive.components.MultiDecoder class method) (base_attentive.components.MultiDecoder method) (base_attentive.components.MultiScaleLSTM class method) (base_attentive.components.MultiScaleLSTM method) (base_attentive.components.VariableSelectionNetwork class method) (base_attentive.core.base_attentive.BaseAttentive class method)

fusion (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) future_encoder (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) future_input_dim (base_attentive.config.schema.BaseAttentiveSpec attribute) future_positional_encoding (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) future_processor (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) future_projection (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute)

G

get_available_backends() (in module base_attentive) (in module base_attentive.backend) get_available_devices() (base_attentive.backend.TorchDeviceManager method), [1] (TorchDeviceManager method) get_backend() (in module base_attentive) (in module base_attentive.backend) get_backend_capabilities() (in module base_attentive) (in module base_attentive.backend) get_backend_version() (in module base_attentive.backend) get_capabilities() (base_attentive.backend.Backend method) get_config() (base_attentive.BaseAttentive method) (base_attentive.components.CrossAttention method) (base_attentive.components.HierarchicalAttention method) (base_attentive.components.MemoryAugmentedAttention method) (base_attentive.components.MultiDecoder method), [1] (base_attentive.components.MultiScaleLSTM method), [1] (base_attentive.components.TransformerDecoderLayer method) (base_attentive.components.TransformerEncoderLayer method) (base_attentive.components.VariableSelectionNetwork method) (base_attentive.core.base_attentive.BaseAttentive method)

get_device_info() (base_attentive.backend.TorchDeviceManager method), [1] (TorchDeviceManager method) get_layer_class() (in module base_attentive) get_model_class() (in module base_attentive) get_params() (base_attentive.api.property.NNLearner method) get_torch_device() built-in function get_torch_device() (in module base_attentive.backend), [1] get_torch_version() built-in function get_torch_version() (in module base_attentive.backend), [1]

H

has() (base_attentive.registry.ComponentRegistry method) (base_attentive.registry.ModelRegistry method) head_type (base_attentive.config.schema.BaseAttentiveSpec attribute) hidden_projection (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute)

hidden_units (base_attentive.config.schema.BaseAttentiveSpec attribute) HierarchicalAttention (class in base_attentive.components) hybrid_encoder (base_attentive.config.schema.BaseAttentiveComponentSpec attribute)

I

is_available() (base_attentive.backend.Backend method)

J

JaxBackend (class in base_attentive.backend)

L

Layer (base_attentive.backend.Backend attribute) layer_norm_epsilon (base_attentive.config.schema.BaseAttentiveSpec attribute) LayerNormalization (base_attentive.backend.Backend attribute)

list_keys() (base_attentive.registry.ComponentRegistry method) LSTM (base_attentive.backend.Backend attribute) lstm_units (base_attentive.config.schema.BaseAttentiveSpec attribute)

M

make_fast_predict_fn() (in module base_attentive) (in module base_attentive.runtime) max_window_size (base_attentive.config.schema.BaseAttentiveSpec property) maybe_reduce_quantiles_bh() (in module base_attentive.validation) memory_size (base_attentive.config.schema.BaseAttentiveSpec property) MemoryAugmentedAttention (class in base_attentive.components) mode (base_attentive.config.schema.BaseAttentiveSpec property) Model (base_attentive.backend.Backend attribute) ModelRegistry (class in base_attentive.registry) module base_attentive base_attentive.api.property base_attentive.backend base_attentive.runtime base_attentive.validation

multi_horizon_head (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) multi_scale_agg (base_attentive.config.schema.BaseAttentiveSpec property) MultiDecoder (class in base_attentive.components) MultiHeadAttention (base_attentive.backend.Backend attribute) MultiScaleLSTM (class in base_attentive.components)

N

name (base_attentive.backend.Backend attribute) (base_attentive.backend.JaxBackend attribute) (base_attentive.backend.PyTorchBackend attribute) (base_attentive.backend.TensorFlowBackend attribute) (base_attentive.backend.TorchBackend attribute)

NNLearner (class in base_attentive.api.property) normalize_backend_name() (in module base_attentive.backend) num_encoder_layers (base_attentive.config.schema.BaseAttentiveSpec property) num_heads (base_attentive.config.schema.BaseAttentiveSpec property)

O

objective (base_attentive.config.schema.BaseAttentiveSpec property)

output_dim (base_attentive.config.schema.BaseAttentiveSpec attribute) output_head (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute)

P

parse_version() (in module base_attentive.backend) point_head (base_attentive.config.schema.BaseAttentiveComponentSpec attribute)

positional_encoder (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) PyTorchBackend (class in base_attentive.backend)

Q

quantile_distribution_head (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute)

quantile_head (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) quantiles (base_attentive.config.schema.BaseAttentiveSpec attribute)

R

register() (base_attentive.registry.ComponentRegistry method) (base_attentive.registry.ModelRegistry method) register_keras_serializable() (in module base_attentive) required_modules (base_attentive.backend.Backend attribute) (base_attentive.backend.JaxBackend attribute) (base_attentive.backend.TensorFlowBackend attribute) (base_attentive.backend.TorchBackend attribute)

reset_cache() (base_attentive.backend.TorchDeviceManager method), [1] residual_projection (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) resolve() (base_attentive.registry.ComponentRegistry method) (base_attentive.registry.ModelRegistry method) resolve_keras_dep() (in module base_attentive) runtime (base_attentive.config.schema.BaseAttentiveSpec attribute)

S

scales (base_attentive.config.schema.BaseAttentiveSpec property) select_best_backend() (in module base_attentive.backend) sequence_pool (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) sequence_pooling (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) Sequential (base_attentive.backend.Backend attribute) set_backend() (in module base_attentive) (in module base_attentive.backend) set_device() (base_attentive.backend.TorchDeviceManager method), [1] (TorchDeviceManager method) set_params() (base_attentive.api.property.NNLearner method) static_input_dim (base_attentive.config.schema.BaseAttentiveSpec attribute)

static_processor (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) static_projection (base_attentive.config.schema.BaseAttentiveComponentSpec attribute) (base_attentive.resolver.assembly.BaseAttentiveV2Assembly attribute) supports_base_attentive (base_attentive.backend.Backend attribute) (base_attentive.backend.JaxBackend attribute) (base_attentive.backend.TensorFlowBackend attribute) (base_attentive.backend.TorchBackend attribute) supports_base_attentive_v2 (base_attentive.backend.Backend attribute) (base_attentive.backend.JaxBackend attribute) (base_attentive.backend.TensorFlowBackend attribute) (base_attentive.backend.TorchBackend attribute)

T

Tensor (base_attentive.backend.Backend attribute) TensorFlowBackend (class in base_attentive.backend) torch_is_available() built-in function torch_is_available() (in module base_attentive.backend), [1]

TorchBackend (class in base_attentive.backend) TorchDeviceManager (built-in class) (class in base_attentive.backend), [1] TransformerDecoderLayer (class in base_attentive.components) TransformerEncoderLayer (class in base_attentive.components)

U

use_batch_norm (base_attentive.config.schema.BaseAttentiveSpec property) use_residuals (base_attentive.config.schema.BaseAttentiveSpec property) use_vsn (base_attentive.config.schema.BaseAttentiveSpec property)

uses_keras_runtime (base_attentive.backend.Backend attribute) (base_attentive.backend.JaxBackend attribute) (base_attentive.backend.TensorFlowBackend attribute) (base_attentive.backend.TorchBackend attribute)

V

v2_blockers (base_attentive.backend.Backend attribute) (base_attentive.backend.JaxBackend attribute) (base_attentive.backend.TorchBackend attribute) validate_model_inputs() (in module base_attentive.validation)

VariableSelectionNetwork (class in base_attentive.components) verbose (base_attentive.config.schema.BaseAttentiveSpec property) version_at_least() (in module base_attentive.backend) vsn_units (base_attentive.config.schema.BaseAttentiveSpec attribute)