Dataset that pulls from UniRef/Uniclust downloads. The data folder should contain the following: - 'consensus.fasta': consensus sequences, no

Raised when length isn't acceptable.

Protein structure representation.

r""" Subset of a dataset at specified indices. Args: dataset (Dataset): The whole Dataset indices (sequence): Indices in the

Dataset that pulls from UniRef/Uniclust downloads. The data folder should contain the following: - 'consensus.fasta': consensus sequences, no

Implementation of dropout with the ability to share the dropout mask along a particular dimension.

Raised when resolution isn't acceptable.

Dataset that pulls from UniRef/Uniclust downloads. The data folder should contain the following: - 'consensus.fasta': consensus sequences, no

Aggregation of the various losses described in the supplement.

Implements Algorithm 20, lines 11-14.

Parameters: ----------- sampler : Pytorch Sampler Choose base sampler class to use for bucketing max_tokens : int

Parameters: ----------- sampler : Pytorch Sampler Choose base sampler class to use for bucketing max_tokens : int

Parameters: ----------- sampler : Pytorch Sampler Choose base sampler class to use for bucketing max_tokens : int

Implements part of Algorithm 23.

A column containing text.

Pytorch pickle loading workaround. https://github.com/pytorch/pytorch/issues/16797

Wrapped for OA Collater to operate on ESM w/ ESM alphabet and batch converter/tokens.

Data exception.

Implements Algorithm 22.

Raised when mmcif parsing fails.

Rotary position embeddings based on those in. [RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer). Query and keys are transfo

Rotary position embeddings based on those in. [RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer). Query and keys are transfo

Returns indices such that inputs with similar lengths are close together.

Returns indices such that inputs with similar lengths are close together.

Returns indices such that inputs with similar lengths are close together.