Green Computing Toolbox for Estimating Neural Network Model Performance
The Green Computing Toolbox is a python library that enables faster AI/ML model architecture and hyper-parameter searches via [Training Speed Estimation (TSE)](https://doi.org/10.48550/arXiv.2112.03364) and [Loss Curve Gradient Approximation (LCGA)](https://doi.org/10.1109/IPDPSW55747.2022.00123). The toolbox also includes the capability of of NVIDIA GPU power capping via a SLURM plugin and enables the comparison of energy consumed during training of different model architectures.
The Green Computing Toolbox depends on [Hydra-Zen](https://github.com/mit-ll-responsible-ai/hydra-zen) and the `submitit` plugin and has been tested to work with [SLURM](https://slurm.schedmd.com) managed clusters. Implementation of TSE is from the [LitMatter](https://github.com/ncfrey/litmatter) package.
The Green Computing Toolbox is a python library that enables faster AI/ML model architecture and hyper-parameter searches via Training Speed Estimation (TSE) and Loss Curve Gradient Approximation (LCGA). The toolbox also includes the capability of of NVIDIA GPU power capping via a SLURM plugin and enables the comparison of energy consumed during training of different model architectures.
The Green Computing Toolbox depends on Hydra-Zen and the submitit plugin and has been tested to work with SLURM managed clusters. Implementation of TSE is from the LitMatter package.