Delegate a log call to the underlying logger, after prefixing its message with the rank of the process it's being logged from. If `'rank'` is

Extract the data from the csv files and store it in the DataFrames.

Initialize DataFrames for the metrics.

Convert a Tensor with dimensions [batch, series, time steps, ...] to one with dimensions [batch, series * time steps, ...]

Generate a LaTeX table for a metric with a fixed number of hypotheses across datasets and methods.

Parameters: ----------- input_encoded: torch.Tensor [batch, series, time steps, embedding dimension] An embedding

input : np.array of dim (num_simu,pred_length,target_dim) output : mean of total variation on the num_simu

Normalize the given values according to the historical data sent in the constructor. Parameters: ----------- value:

Train an ETS model on the entire dataset. Each column in the DataFrame is independently fitted with an ETS model. In this im

Format DataFrame to bold the best and underline the second best values per row.

Computes the weighted average of a given tensor across a given dim, masking values associated with weight zero, meaning instead of `nan *

Create and return the instance splitter needed for training, validation or testing. Parameters: ----------- mode: st

Generate a LaTeX table for a single dataset with varying numbers of hypotheses.

we overrided the distribution method! The goal was to retun a distribution objet. We return a DIFFUSION OBJECT

Transform the given value according to the given parameters, but does not compute the derivative of the transformation. param

Generates samples based on the specified stochastic process.

Compute accuracy metrics by comparing actual data to the forecasts. Parameters ---------- ts_iterator it

Create and return the instance splitter needed for training, validation or testing. Returns ------- Transformation

Unrolls the RNN encoder over past and, if present, future data. Returns outputs and state of the encoder, plus the scale of p

Create and return a predictor object. Returns ------- Predictor A predictor wrapping a `nn.Module` used

Create and return the transformation needed for training and inference. Returns ------- Transformation T

splitting between start_date and end_date in num_tests parts, and returning the test_dataset object you can directly use in you code, just lik

Returns the distribution of DeepVAR with respect to the RNN outputs. Parameters ---------- rnn_outputs O

Returns the distribution of DeepVAR with respect to the RNN outputs. Parameters ---------- rnn_outputs O

Returns the distribution of DeepVAR with respect to the RNN outputs. Parameters ---------- rnn_outputs O

Returns the distribution of DeepVAR with respect to the RNN outputs. Parameters ---------- rnn_outputs O

Returns the distribution of DeepVAR with respect to the RNN outputs. Parameters ---------- rnn_outputs O

Extract unique forecasts and their probabilities. Args: samples: Array of shape (n_hyp, forecast_length, target_dim) for TimeGrad

Extract unique forecasts and their probabilities. Args: samples: Array of shape (n_hyp, forecast_length, target_dim) for TimeGrad

Returns lagged subsequences of a given sequence. Parameters ---------- sequence the sequence from which l

Returns lagged subsequences of a given sequence. Parameters ---------- sequence the sequence from which l

Returns lagged subsequences of a given sequence. Parameters ---------- sequence the sequence from which l

Returns lagged subsequences of a given sequence. Parameters ---------- sequence the sequence from which l

Returns lagged subsequences of a given sequence. Parameters ---------- sequence the sequence from which l

Logarithm of the sigmoid function. Substract the epsilon to avoid 0 as a possible value for large x.

Computes the loss for training TimeMCL, all inputs tensors representing time series have NTC layout. Parameters ----

Compute the loss function of the model. Parameters: ----------- hist_time: Tensor [batch, series, time steps] or [ba

Computes the loss for training DeepVAR, all inputs tensors representing time series have NTC layout. Parameters ----

Computes the loss for training DeepVAR, all inputs tensors representing time series have NTC layout. Parameters ----

Computes the loss for training DeepVAR, all inputs tensors representing time series have NTC layout. Parameters ----

Computes the loss for training DeepVAR, all inputs tensors representing time series have NTC layout. Parameters ----

Args: ... (existing args) ... is_mcl: Whether the forecasts come from timeMCL model (with scores) or TimeGrad

"returns : - contexte_df : the real data, around of the point of interest - forecast_array : the predictions of the neural network - star

For each series in the context, generate num_samples future trajectories using the simulate method of the fitted ETS model.

Wrap a LaTeX table string with formatting, caption, and label.

Unrolls the RNN encoder over past and, if present, future data. Returns outputs and state of the encoder, plus the scale of p

Unrolls the RNN encoder over past and, if present, future data. Returns outputs and state of the encoder, plus the scale of p

Unrolls the RNN encoder over past and, if present, future data. Returns outputs and state of the encoder, plus the scale of p

Unrolls the RNN encoder over past and, if present, future data. Returns outputs and state of the encoder, plus the scale of p

Generate the base path for results CSV based on dataset and config.

Parameters: ----------- input_dim: int The dimension of the encoded representation (upstream data encoder).

A single layer of the Deep Sigmoid Flow network. Does not contains its parameters, they must be sent in the forward method.

Parameters: ----------- num_series: int Number of series of the data which will be sent to the model. Required fo

Parameters: ----------- attention_layers: int How many successive attention layers this encoder will use.

Only keep a small number of series for each of the input tensors. Which series will be kept is randomly selected for each batch. The

Check if there are gaps in the data indexed by days. Return the number of "gaps" and the total number of missing days.

just checking if there is any gap in the data return the number of "gaps", and the total number of hours missing/the total

Generate a MLP with the given parameters.

Generates batch_size AR(p) process samples, extracts sub-trajectories.