New on NGC Catalog: Samsung SDS Brightics, an AI Accelerator for Automating and Accelerating Deep Learning

The Kubernetes-based, containerized application, is now available on the NVIDIA NGC Catalog – a GPU-optimized hub for AI and HPC containers.

NVIDIA logo
NVIDIA
•
Brad Nemire
•2 min read•advanced•
--
•View Original
AutoMLDeep LearningHelmKerasKubernetesPyTorchResNetTensorFlow

Overview

The article discusses the Samsung SDS Brightics AI Accelerator, a Kubernetes-based application designed to automate and accelerate deep learning model training. It highlights its availability on the NVIDIA NGC catalog and its key features that enhance model training efficiency across various industries.

What You'll Learn

1

How to automate machine learning processes using the Brightics AI Accelerator

2

Why using distributed clusters can significantly reduce model training time

3

When to apply automated feature engineering and hyper-parameter tuning in AI projects

Key Questions Answered

What is the Samsung SDS Brightics AI Accelerator?
The Samsung SDS Brightics AI Accelerator is a Kubernetes-based, containerized application that automates machine learning and accelerates model training. It is designed to improve model accuracy and is available on the NVIDIA NGC catalog, facilitating easier AI development and deployment.
How does the Brightics AI Accelerator improve model training efficiency?
The Brightics AI Accelerator enhances model training efficiency by automating processes such as feature engineering, model selection, and hyper-parameter tuning. It can utilize up to 512 NVIDIA GPUs per training job, reducing training time from 3 weeks to just 1 hour.
What industries can benefit from the Brightics AI Accelerator?
The Brightics AI Accelerator can be applied across various industries, including healthcare, manufacturing, retail, and automotive. It supports diverse use cases such as computer vision and natural language processing, making it versatile for different applications.

Key Statistics & Figures

Training time reduction
From 3 weeks to 1 hour
This statistic highlights the efficiency gained by using the Brightics AI Accelerator compared to traditional training methods.
Speed improvement factor
126 times
Using the Brightics AI Accelerator, the training time for a ResNet 50 image classification model is reduced significantly, showcasing its effectiveness.

Technologies & Tools

Some links below are affiliate links. We may earn a commission if you make a purchase.

Orchestration
Kubernetes
Used as the basis for the Brightics AI Accelerator application.
Framework
Tensorflow
One of the frameworks supported for training deep learning models.
Framework
Keras
Another framework supported for deep learning model training.
Framework
Pytorch
Supported framework for training deep learning models.
Library
Scikit-learn
Used for automated machine learning processes.

Key Actionable Insights

1
Utilize the Brightics AI Accelerator to streamline your AI model training process.
By automating key aspects of model training, such as feature engineering and hyper-parameter tuning, teams can focus on higher-level tasks and reduce the time spent on model development.
2
Leverage the distributed training capabilities of Brightics to enhance performance.
With the ability to use up to 512 NVIDIA GPUs, organizations can achieve significant reductions in training time, making it feasible to iterate on models more rapidly.

Common Pitfalls

1
Failing to leverage the full capabilities of distributed training can lead to unnecessarily long training times.
Many teams may not utilize the available GPU resources effectively, resulting in missed opportunities for performance improvements.