tempor.models.clairvoyance2.treatment_effects.crn module

class tempor.models.clairvoyance2.treatment_effects.crn.RecurrentFFNet_ConcatTreatment(rnn_type: str, input_size: int, hidden_size: int, nonlinearity: str | None, num_layers: int, bias: bool, dropout: float, bidirectional: bool, proj_size: int | None, ff_out_size: int, ff_in_size_adjust: int = 0, ff_hidden_dims: Sequence[int] = (), ff_out_activation: str | None = 'ReLU', ff_hidden_activations: str | None = 'ReLU')[source]

Bases: RecurrentFFNet

rnn_out_postprocess(rnn_out: Tensor, **kwargs) → Tensor[source]
training : bool
x_used_in_autoregress : Any
class tempor.models.clairvoyance2.treatment_effects.crn.TreatBalancerNet(in_dim: int, out_dim: int, hidden_dims: Sequence[int] = (), out_activation: str | None = 'ReLU', hidden_activations: str | None = 'ReLU')[source]

Bases: FeedForwardNet

Initializes internal Module state, shared by both nn.Module and ScriptModule.

forward(x: Tensor) → Tensor[source]

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

training : bool
class tempor.models.clairvoyance2.treatment_effects.crn.CRNTreatmentEffectsModelBase(loss_fn: Module, params: dict[str, Any] | None = None)[source]

Bases: TreatmentEffectsModel, Seq2SeqCRNStylePredictorBase, OrganizedTreatmentEffectsModuleMixin

class tempor.models.clairvoyance2.treatment_effects.crn.CRNRegressor(params: dict[str, Any] | None = None)[source]

Bases: CRNTreatmentEffectsModelBase

requirements : Requirements = Requirements(dataset_requirements=DatasetRequirements(requires_static_covariates_present=False, requires_no_missing_data=True, static_covariates_value_type=<DataValueOpts.NUMERIC: 2>, temporal_covariates_value_type=<DataValueOpts.NUMERIC: 2>, temporal_targets_value_type=<DataValueOpts.NUMERIC_CATEGORICAL: 3>, temporal_treatments_value_type=<DataValueOpts.NUMERIC_BINARY: 4>, event_covariates_value_type=<DataValueOpts.ANY: 1>, event_targets_value_type=<DataValueOpts.ANY: 1>, event_treatments_value_type=<DataValueOpts.ANY: 1>, requires_all_temporal_data_samples_aligned=False, requires_all_temporal_data_regular=False, requires_all_temporal_data_index_numeric=False, requires_all_temporal_containers_shares_index=True), prediction_requirements=PredictionRequirements(target_data_structure=<DataStructureOpts.TIME_SERIES: 1>, horizon_type=<HorizonOpts.TIME_INDEX: 2>, min_timesteps_target_when_fit=6, min_timesteps_target_when_predict=3), treatment_effects_requirements=TreatmentEffectsRequirements(treatment_data_structure=<DataStructureOpts.TIME_SERIES: 1>, min_timesteps_treatment_when_fit=6, min_timesteps_treatment_when_predict=3, min_timesteps_treatment_when_predict_counterfactual=3))
DEFAULT_PARAMS : dict[str, Any] | NamedTuple = _DefaultParams(encoder_rnn_type='LSTM', encoder_hidden_size=100, encoder_num_layers=1, encoder_bias=True, encoder_dropout=0.0, encoder_bidirectional=False, encoder_nonlinearity=None, encoder_proj_size=None, decoder_rnn_type='LSTM', decoder_hidden_size=100, decoder_num_layers=1, decoder_bias=True, decoder_dropout=0.0, decoder_bidirectional=False, decoder_nonlinearity=None, decoder_proj_size=None, adapter_hidden_dims=[50], adapter_out_activation='Tanh', predictor_hidden_dims=[], predictor_out_activation=None, treat_net_hidden_dims=[], treat_net_out_activation=None, max_len=None, optimizer_str='Adam', optimizer_kwargs={'lr': 0.01, 'weight_decay': 1e-05}, batch_size=32, epochs=100, padding_indicator=-999.0)
process_output_for_loss(output: Tensor, **kwargs) → Tensor[source]
encoder_treat_net : FeedForwardNet | None
decoder_treat_net : FeedForwardNet | None
encoder : RecurrentFFNet | None
encoder_optim : torch.optim.Optimizer | None
decoder : RecurrentFFNet | None
decoder_optim : torch.optim.Optimizer | None
adapter : FeedForwardNet | None
encoder_output_and_h_dim : tuple[int, int, int]
decoder_output_and_h_dim : tuple[int, int, int]
loss_fn : nn.Module
params : DotMap
inferred_params : DotMap
training : bool
class tempor.models.clairvoyance2.treatment_effects.crn.CRNClassifier(params: dict[str, Any] | None = None)[source]

Bases: CRNTreatmentEffectsModelBase

requirements : Requirements = Requirements(dataset_requirements=DatasetRequirements(requires_static_covariates_present=False, requires_no_missing_data=True, static_covariates_value_type=<DataValueOpts.NUMERIC: 2>, temporal_covariates_value_type=<DataValueOpts.NUMERIC: 2>, temporal_targets_value_type=<DataValueOpts.NUMERIC_CATEGORICAL: 3>, temporal_treatments_value_type=<DataValueOpts.NUMERIC_BINARY: 4>, event_covariates_value_type=<DataValueOpts.ANY: 1>, event_targets_value_type=<DataValueOpts.ANY: 1>, event_treatments_value_type=<DataValueOpts.ANY: 1>, requires_all_temporal_data_samples_aligned=False, requires_all_temporal_data_regular=False, requires_all_temporal_data_index_numeric=False, requires_all_temporal_containers_shares_index=True), prediction_requirements=PredictionRequirements(target_data_structure=<DataStructureOpts.TIME_SERIES: 1>, horizon_type=<HorizonOpts.TIME_INDEX: 2>, min_timesteps_target_when_fit=6, min_timesteps_target_when_predict=3), treatment_effects_requirements=TreatmentEffectsRequirements(treatment_data_structure=<DataStructureOpts.TIME_SERIES: 1>, min_timesteps_treatment_when_fit=6, min_timesteps_treatment_when_predict=3, min_timesteps_treatment_when_predict_counterfactual=3))
DEFAULT_PARAMS : dict[str, Any] | NamedTuple = _DefaultParams(encoder_rnn_type='LSTM', encoder_hidden_size=100, encoder_num_layers=1, encoder_bias=True, encoder_dropout=0.0, encoder_bidirectional=False, encoder_nonlinearity=None, encoder_proj_size=None, decoder_rnn_type='LSTM', decoder_hidden_size=100, decoder_num_layers=1, decoder_bias=True, decoder_dropout=0.0, decoder_bidirectional=False, decoder_nonlinearity=None, decoder_proj_size=None, adapter_hidden_dims=[50], adapter_out_activation='Tanh', predictor_hidden_dims=[], predictor_out_activation=None, treat_net_hidden_dims=[], treat_net_out_activation=None, max_len=None, optimizer_str='Adam', optimizer_kwargs={'lr': 0.01, 'weight_decay': 1e-05}, batch_size=32, epochs=100, padding_indicator=-999.0)
process_output_for_loss(output: Tensor, **kwargs) → Tensor[source]
encoder_treat_net : FeedForwardNet | None
decoder_treat_net : FeedForwardNet | None
encoder : RecurrentFFNet | None
encoder_optim : torch.optim.Optimizer | None
decoder : RecurrentFFNet | None
decoder_optim : torch.optim.Optimizer | None
adapter : FeedForwardNet | None
encoder_output_and_h_dim : tuple[int, int, int]
decoder_output_and_h_dim : tuple[int, int, int]
loss_fn : nn.Module
params : DotMap
inferred_params : DotMap
training : bool