Meta’s MultiRay, a ML Platform for Running Large Foundational Models

2 years ago thenewstack.io

Summary: This is a summary of an article originally published by The New Stack. Read the full original article here →

Training an AI model is no small feat. With the costs associated with training each model, it’s easy to see how this can spiral out of control.

Multiple specialized, smaller models can now run off of the input (also known as embedding) from the universal model.

PostRay models are complex to train, deploy, and maintain because they incorporate advanced research in multiple fields but only need training once.

This is a moving target that undergoing continuous innovation in new model architectures, heavy investment in model refresh, and training flows.

DevOps Articles