Fiddle is a Python-first configuration library used to set and manage values for functions and classes without deeply modifying application code. In machine learning systems, it supports reproducible configuration of experiments, models, and training infrastructure.
A Python-first configuration library that sets the values of functions and classes without invasive code or infrastructure. In the case of Pax---and other ML codebases---these functions and classes represent models and training hyperparameters. Fiddle assumes that machine learning codebases are typically divided into: - Library code, which defines the layers and optimizers. - Dataset "glue" code, which calls the libraries and wires everything together. Fiddle captures the call structure of the glue code in an unevaluated and mutable form.