r""" Newton-Schulz iteration to compute the zeroth power / orthogonalization of G. We opt to use a quintic iteration whose coefficient

Calculate the accuracy of predictions. Args: prediction (torch.Tensor): The predicted values as a tensor. targets (torch.Ten

Perform auto-attention mechanism on the input. Args: inp (torch.Tensor): Input data of shape [B, N, T], where B is the b

Check for NaN or Inf values in a tensor. Args: tensor (torch.Tensor): Input tensor to check. raise_ex (bool, optional): Whet

Decorator to measure the execution time of a function. This decorator prints the time taken for a function to execute.

Save an object to a pickle file. Args: obj (object): Object to save. file_path (str): Path to the file.

Fit the scaler to the training data. Args: data (torch.Tensor): Training data used to fit the scaler.

Fit the scaler to the training data. Args: data (torch.Tensor): Training data used to fit the scaler.

Forward pass of the moving average block. Args: inputs (torch.Tensor): Input tensor of shape [batch_size, seq_len, num_f

Forward pass of the moving average decomposition block. Args: inputs (torch.Tensor): Input tensor of shape [batch_size,

Forward pass of the series decomposition block. Args: inputs (torch.Tensor): Input tensor of shape [batch_size, seq_len,

Forward pass of the DFT decomposition layer. Args: inputs (torch.Tensor): Input tensor of shape [batch_size, seq_len, nu

Forward pass of the MLP layer. Args: inputs (torch.Tensor): Input tensor of shape [..., input_size] Returns:

Forward pass of the MLP layer. Args: inputs (torch.Tensor): Input tensor of shape [..., input_size] Returns:

Position embedding. Args: x (torch.Tensor): Input tensor of shape [batch_size, seq_len, hidden_size]. Returns:

Forward pass of the TimestampEmbedding layer. Args: timestamps (torch.Tensor): Input tensor of timestamps of shape [batc

Forward pass of the FeatureEmbedding layer. Args: inputs (torch.Tensor): Input tensor of shape [batch_size, seq_len, num

Forward pass of the SequenceEmbedding layer. Args: inputs (torch.Tensor): Input tensor of shape [batch_size, seq_len, nu

Forward pass of the PatchEmbedding layer. Args: x (torch.Tensor): Input tensor of shape [batch_size, seq_len, num_featur

:param x: input tensor of shape [batch_size, seq_len, num_features] :param mode: 'norm' for normalization, 'denorm' for denormalizati

Forward pass of STNorm. Args: inputs (Tensor): Input data with shape: [batch_size, input_len, num_features]

Forward pass of the decoder layer. Args: hidden_states (torch.Tensor): The input hidden states.

Forward pass of the auto-regressive decoder. Args: hidden_states (torch.Tensor): The input hidden states

Forward pass of the Seq2SeqDecoderLayer. Args: hidden_states (torch.Tensor): The input hidden states.

Forward pass of the Seq2SeqDecoderLayer. Args: hidden_states (torch.Tensor): The input hidden states.

Forward pass of the Seq2SeqDecoder. Args: hidden_states (torch.Tensor): The input hidden states.

Forward pass of the Seq2SeqDecoder. Args: hidden_states (torch.Tensor): The input hidden states.

Apply and cache RoPE. Args: q (torch.Tensor): The query tensor with shape [batch_size, n_heads, seq_len, head_size].

Feed-forward of BasicTS Transformer encoder layer. Args: hidden_states (torch.Tensor): Input hidden states.

Feed-forward of BasicTS Transformer encoder. Args: hidden_states (torch.Tensor): Input hidden states. attent

Forward pass of TimeXer. Args: inputs (torch.Tensor): Input tensor of shape [batch_size, input_len, num_features].

Feed forward of DLinear. Args: inputs (torch.Tensor): inputs data with shape [batch_size, input_len, num_features] Retur

Single-`inputs` forward to match runner API. Args: inputs (torch.Tensor): history input with shape [B, L, C] or [B, L, C

Feed forward of STID. Args: inputs (torch.Tensor): history data with shape [batch_size, input_len, num_features] inpu

Forward pass of the Leddam model. Args: inputs (torch.Tensor): Input tensor of shape [batch_size, input_len, num_feature

Forward pass of the Auto-Attention module. Args: P (int): The period for autoregressive behavior. inp (torch

Forward pass of the TimesNetBackbone. Args: inputs (torch.Tensor): Input tensor of shape [batch_size, input_len, num

Forward pass of the TimesNetForForecasting. Args: inputs (torch.Tensor): Input tensor of shape [batch_size, input_le

Forward pass of the Crossformer model. Args: inputs (torch.Tensor): Input tensor of shape [batch_size, input_len, num_fe

Args: x (torch.Tensor): Input tensor of shape [batch_size * num_features, num_patches, hidden_size] Returns:

Crossformer Encoder Layer Args: hidden_states (torch.Tensor): Input tensor of shape [batch_size * num_features, num_patc

Args: hidden_states (torch.Tensor): Input tensor of shape [batch_size * num_features, num_patches, hidden_size] key_v

Feed forward of Autoformer. Args: inputs: Input data with shape: [batch_size, input_len, num_features] targe

Feedforward function of StemGNN. Args: inputs (Tensor): Input data with shape: [batch_size, input_len, num_features] Ret

Args: inputs (Tensor): Input data with shape: [batch_size, input_len, num_features] Returns: torch.Tensor:

Forward pass of PatchTSTForForecasting. Args: inputs (Tensor): Input data with shape: [batch_size, input_len, num_featur

Forward pass of PatchTSTForForecasting. Args: inputs (Tensor): Input data with shape: [batch_size, input_len, num_featur

Forward pass of PatchTSTForReconstruction. Args: inputs (Tensor): Input data with shape: [batch_size, input_len, num_fea

Forward pass of LightTS. Args: inputs (torch.Tensor): Input tensor with shape [batch_size, input_len, num_features].

Forward pass of FITS. Args: inputs (torch.Tensor): Input tensor of shape [batch_size, input_len, num_features].

inputs : (length, ...) output : (length, ..., N) where N is the order of the HiPPO projection

Feed forward of FiLM. Args: inputs (torch.Tensor): inputs data with shape [batch_size, input_len, num_features] Returns:

Forward pass of Timer model. Args: inputs (torch.Tensor): Input tensor of shape [batch_size, input_len, num_features].