Software libraries

MLoggers

A package of loggers well-suited for machine learning experiments. It integrates with Weights and Biases and provides structured file and console logging.

LData

A library for managing datasets and benchmarks in NLP experiments. It provides a simple API for using datasets to train and evaluate models across different pre-defined or user-defined language tasks.

LModels

A library defining a common API to work with state-of-the-art LLMs. It provides wrappers for the Hugging Face and OpenAI APIs, amongst others.

LMethods

A library containing methods to enhance the capabilities of LLMs, such as Chain-of-Thought, Tree of Thoughts, Decomposed Prompting, etc. A common interface can be used to implement new methods.