Schedules and executes a set of tasks and their dependencies. Handles scheduling, execution, the movement of data between devices, and the l

r""" Dataset attributes.

r""" Computes and stores the average and current value.

r""" Arguments pertaining to which techniques we are going to fine-tuning with.

r""" Arguments pertaining to which model/config/tokenizer we are going to fine-tune.

r""" Wraps the tokenizer into metric functions, used in Seq2SeqPeftTrainer.

r""" Inherits PPOTrainer.

r""" Inherits PeftTrainer to compute generative metrics such as BLEU and ROUGE.

r""" Data collator for pairwise data.

Naive `torch.load` shard loading.

r""" Arguments pertaining to specify the decoding parameters.

Loader for pytorch files using a custom unpickler and vigorous monkeypatching.

r""" Logger handler used in Web UI.

A reference to a language model. Can be a hf hub path (username/repo), or local. Optionally includes a LoRA.

r""" Data collator for pairwise data.

r""" Inherits PeftTrainer to compute pairwise loss.

Class for lazily reading (sections of) files from a torch ZIP archive. Maintains a handle to the most recently opened file for faster access

Information about an individual weight tensor in a model. Attributes: name (str): The name of the tensor representing the wei

Custom YAML dumper to format lists of numbers in flow style.

r""" Arguments pertaining to what data we are going to input our model for training and evaluation.

r""" Arguments pertaining to specify the evaluation parameters.