By MICHAEL CIRCLE
The Data Group News Service
Snowflake, the AI Data Cloud company, announced Wednesday it will host Meta's new Llama 3.1 large language models in its Snowflake Cortex AI platform, allowing enterprises to build AI applications at scale.
The partnership includes Meta's most advanced open-source model, Llama 3.1 405B. Snowflake has optimized the model for both inference and fine-tuning, supporting a 128K context window and enabling real-time inference with improved latency and throughput compared to existing open-source solutions.
Snowflake's AI Research Team developed a new inference and fine-tuning system stack, which they're open-sourcing in collaboration with DeepSpeed, Hugging Face, vLLM and the broader AI community. This system allows for fine-tuning of the massive model using a single GPU node, reducing costs and complexity for developers.
The company is also making its Cortex Guard feature generally available to protect against harmful content in AI applications built using Cortex AI. Cortex Guard leverages Meta's Llama Guard 2 to ensure model safety.
Vivek Raghunathan, VP of AI Engineering at Snowflake, said, "We're not just bringing Meta's cutting-edge models directly to our customers through Snowflake Cortex AI. We're arming enterprises and the AI community with new research and open source code that supports 128K context windows, multi-node inference, pipeline parallelism, 8-bit floating point quantization, and more to advance AI for the broader ecosystem."
The move aims to democratize access to powerful natural language processing and generation applications while prioritizing efficiency and safety.
To learn more about Snowflake, visit Newsroom | Snowflake - The AI Data Cloud.
***
If your business has a data problem, book your free consultation with The Data Group here!
