A custom sampler for PyTorch DataLoader that creates batches with sequences of similar lengths.. Args: lengths (list): A lis

A PyTorch module implementing a ChordMixer Encoder as a stack of ChordMixer layers. The number of layers in the stack is determined by the m

A PyTorch module implementing 1D Batch Normalization for token embeddings. Args: embedding_size (int): The size of the token embeddi

A PyTorch module implementing the ChordMixerBlock. This module combines two main steps in the ChordMixer layer: Rotate and Mix. The drop

A PyTorch module implementing Group Normalization for token embeddings. Args: embedding_size (int): The size of the token embeddings

Linear decoder for fixed-length input sequences. This module computes the mean of the input tensor along the sequence length dimension and

Linear decoder for variable-length input sequences. This module computes the mean of the input tensor along the sequence length dimension fo

A simple Multi-Layer Perceptron (MLP) module in PyTorch. Args: in_features (int): Number of input features. hidden_features (int, op

Path Finder dataset.

Path Finder dataset.

A PyTorch module that performs a parameter-free rotation of tracks within token embeddings. This module can be used to augment or modify the

A PyTorch module that performs a parameter-free rotation of tracks within variable-length token embeddings. This module can be used to augme

The ChordMixer model. Encoder is a stack of ChordMixer blocks. Decoder a global average pooling, followed by a linear layer. Args: i

The ChordMixer model for the Retrieval task. Has no Decoder

A PyTorch module that performs a parameter-free rotation of tracks within token embeddings with circular padding. The rotation is performed

A custom dataset class for variable lengths sequences masking. Args: df (DataFrame): A DataFrame containing sequences, labels, and l