A Generalized Framework For Population Based Training

# A Generalized Framework for Population Based Training

**arXiv ID:** 1902.01894
**Authors:** Ang Li, Aleksandra Spyra, Sagi Perel, Valentin Dalibard, Max Jaderberg, Chenjie Gu, David Budden, Tim Harley, Pramod Gupta
**Published:** 2019-02-05T20:11:17Z
**Abstract:**
Population Based Training (PBT) is a recent approach that jointly optimizes neural network weights and hyperparameters which periodically copies weights of the best performers and mutates hyperparameters during training. Previous PBT implementations have been synchronized glass-box systems. We propose a general, black-box PBT framework that distributes many asynchronous "trials" (a small number of training steps with warm-starting) across a cluster, coordinated by the PBT controller. The black-box design does not make assumptions on model architectures, loss functions or training procedures. Our system supports dynamic hyperparameter schedules to optimize both differentiable and non-differentiable metrics. We apply our system to train a state-of-the-art WaveNet generative model for human voice synthesis. We show that our PBT system achieves better accuracy, less sensitivity and faster convergence compared to existing methods, given the same computational resource.

## Skill Description

This skill is generated from the arXiv paper: A Generalized Framework for Population Based Training (1902.01894).

## How to Use

[To be filled in by the user or by future automation]

## References

- [arXiv:1902.01894](http://arxiv.org/abs/1902.01894v1)