Brightics AccelAI

Brightics AccelAI is a service offering for Machine Learning and Distributed Deep Learning running on a cluster of GPU servers in a Data Center. AccelAI can be hosted on AWS, Azure, Google Cloud and on premises.

Here are the main benefits:

• Reduce Model Training from week(s) to hour(s) using Server GPU Clusters
• Distributed Training in TensorFlow and Keras using Google Parameter Server Technology
• Distributed Training in PyTorch and Tensorflow using Uber Horovod Technology
• Distributed and Scalable Inference API and Jobs for Tensorflow, Keras and PyTorch
• Distributed Data Transformation using Python & OpenCV (without Spark Knowledge)
• Automated Experiments with Hyper-Parameter Search using Grids
• Collaboration & Asset Sharing amongst Data Scientist & AI Engineering Team Members
• Jupyter based Multi-User and Multi-Tenant Environment using Kubernetes
• Scalability as shown below:

AccelAI API

in html:

in pdf:

AccelAI 1.6 Notebooks

AccelAI 2.0 Notebooks