Snowflake Polar models are now accessible in Amazon SageMaker JumpStart | AWS Machine Learning Blog

Snowflake Arctic Instruct Model Available Through Amazon SageMaker JumpStart

Today, we are excited to announce that the Snowflake Arctic Instruct model is available through Amazon SageMaker JumpStart to deploy and run inference. The Snowflake Arctic is an enterprise-grade large language model (LLM) catering to the needs of enterprise users, exhibiting exceptional capabilities in SQL querying, coding, and following instructions accurately.

Features of Snowflake Arctic Instruct

Snowflake Arctic is an enterprise-focused LLM that delivers top-tier enterprise intelligence and cost-efficiency. It is built with a Dense Mixture of Experts (MoE) hybrid transformer architecture and efficient training techniques, enabling it to have a large capacity for enterprise intelligence while being resource-efficient for training and inference.

Deployment and Inference Efficiency

Snowflake Arctic comes with innovations and optimizations for efficient inference. It has fewer memory reads and compute compared to other models, leading to faster performance. The models are available under an Apache 2.0 license, providing access to weights and code for customers.

Discovering and Deploying the Model

With SageMaker JumpStart, users can deploy the Arctic Instruct model with a few clicks in Amazon SageMaker Studio or programmatically through the SageMaker Python SDK. The model is deployed in an AWS secure environment under the user’s virtual private cloud controls, ensuring data security.

Using Snowflake Arctic Instruct Model in SageMaker Studio

SageMaker Studio is an integrated development environment that provides access to purpose-built tools for ML development. Users can access SageMaker JumpStart, find the Snowflake Arctic model in the Hugging Face hub, and deploy it through the example notebook for inference tasks.

Example Prompts for Enterprise Use Cases

Snowflake Arctic Instruct can be used for tasks such as text summarization, code generation, sentiment analysis, and more. Example prompts and outputs demonstrate the model’s capabilities in generating text based on user instructions, including math reasoning and SQL queries.

Cleaning Up Resources After Deployment

After running inference with the Snowflake Arctic Instruct model, users should delete all resources created to stop billing. Resources can be deleted using provided code or through the SageMaker Studio console.

This rewrite includes the main content of the article, structured into paragraphs with appropriate headings for each section.