vilcek

Training and Inference of LLMs with PyTorch Fully Sharded Data Parallel and Better Transformer

In this blog we show how to perform efficient and optimized distributed training and inference of large language models using PyTorch’s Fully Sharded Data Parallel and Better Transformer implementations, on the Spark platform. In this implementation, we combine Microsoft Fabric for data preparation and model inference, and Azure Databricks for model training, having all our […]

Training and Inference of LLMs with PyTorch Fully Sharded Data Parallel and Better Transformer Continue Reading

A Solution Template for Soft Sensor Modeling on Azure – Part 2

In this two-part blog series we explore a solution template for creating models for soft sensors, taking advantage of the scalability and automation provided by the Microsoft Azure platform.   In the first part, we explored what a soft sensor is, the use case, dataset, common approaches and major steps usually needed to model soft

A Solution Template for Soft Sensor Modeling on Azure – Part 2 Continue Reading