learned_optimization reference documentation

learned_optimization is a research codebase for training learned optimizers. It implements hand designed and learned optimizers, tasks to meta-train and meta-test them on, and outer-training algorithms such as ES, gradients and PES.

Indices and tables