Peter Paschalides

Meet Llama 3.2 on Databricks: faster and now multi-modal. The Llama 3.2 release pushes the frontier of enterprise #GenAI with new smaller models for latency + cost-sensitive use cases and larger multimodal models for visual understanding. Customers can easily get started tuning the Llama 3.2 models securely and efficiently on their data. Available today in Databricks Mosaic AI. 

3

So you've looked at DocAI and thought it had to be a very difficult and expensive way to tackle GenAI? Well, with Snowflake Arctic-TILT think again! #ArticTILT easily extracts content, like invoice amounts or contract terms, from documents via a built-in, multimodal large language model (LLM) we call Snowflake Arctic-TILT (Text Image Layout Transformer). Users now have a way of interacting with unstructured data that simplifies the setup and deployment, enabling users without any machine learning background to effortlessly package model builds for high-value, enterprise-scale extraction tasks. #LLM #GenAI #Snowflake #AI #multimodal

16

Join PyTorch at Snowflake's #DataCloudDevDay on June 6 in San Francisco to learn how to accelerate GenAI performance with Native PyTorch. Register: https://hubs.la/Q02xQStt0

18

"#AI/ML models [...] are “living organisms” that change with the underlying data. They require constant monitoring, retraining, and debiasing—a challenge with even a few #ML models but simply overwhelming with hundreds of them." (source see below). #MLOps builds on #DevOps / DevSecOps concepts and therefore it could make sense to use a platform like GitLab for it. Interested in learning how it can be done with GitLab? Let me know :) from https://lnkd.in/dPzXeXxQ

11

Comprehensive blog on the evolution of the Lakehouse Architecture from the time Databricks brought this to market in 2020 to today.

7

1 Comment

Meet Llama 3.2 on Databricks: faster and now multi-modal. The Llama 3.2 release pushes the frontier of enterprise #GenAI with new smaller models for latency + cost-sensitive use cases and larger multimodal models for visual understanding. Customers can easily get started tuning the Llama 3.2 models securely and efficiently on their data. Available today in Databricks Mosaic AI. 

7

This week marked a watershed moment for Snowflake: We launched #SnowflakeArctic, a state-of-the-art large language model uniquely designed to be the most open, enterprise-grade LLM on the market. Arctic not only outperforms many leading open models, but we brought it to market at one-eighth of the cost and in less than three months compared to other models. Arctic will significantly enhance our ability to deliver reliable, efficient AI to our customers. By delivering industry-leading intelligence and efficiency in a truly open way to the AI community, we are furthering the frontiers of what open-source AI can do. https://lnkd.in/eueT5Wcs

78

Snowflake is excited to partner with Anthropic to bring the latest Claude models to customers in #Snowflake Cortex AI. With Claude, enterprises can build cutting-edge generative AI apps, copilots, agents, and more — while benefiting from Snowflake’s industry-leading platform as their data foundation! Learn more: https://okt.to/O3ce5r

34

Introducing #DBRX: a general-purpose LLM that sets a new standard for efficient open source models. More organizations are leveraging open source models for efficiency and cost. The Databricks Mosaic Research team designed DBRX to be easily customizable so that enterprises can improve the quality of their GenAI applications. Leverage it for RAG or to build a custom LLM on your own private data.

2

Meet Llama 3.2 on Databricks: faster and now multi-modal. The Llama 3.2 release pushes the frontier of enterprise #GenAI with new smaller models for latency + cost-sensitive use cases and larger multimodal models for visual understanding. Customers can easily get started tuning the Llama 3.2 models securely and efficiently on their data. Available today in Databricks Mosaic AI. 

9

Stop by Speedata.io booth tomorrow at MinneAnalytics #datatech2024. We're demoing hardware accelerated Apache Spark: 🏎 20x-200x Faster Queries 💸 50x infrastructure savings vs. CPUs and GPUs ⚡ 91% less power than CPUs and GPUs Zero-code change to your DataFrames and SparkSQL

69

4 Comments

Here is a condensed version of my Braindate by e180 sessions at Databricks Data + AI Summit NVIDIA Spark RAPIDS in Amazon Web Services (AWS) EMR provides 2x-5x acceleration and 30% cost savings for the 25% most GPU friendly queries NVIDIA Spark RAPIDS in Microsoft Azure Databricks provides 34% cost savings Use NVIDIA Spark RAPIDS qualification tool to predict the speedup up your workload Use Speedata.io Spark workload analyzer to predict the speedup of your workload Speedata.io delivers 20x-200x acceleration and 50x better performance/$ than CPUs and GPUs

41

2 Comments

Meet Llama 3.2 on Databricks: faster and now multi-modal. The Llama 3.2 release pushes the frontier of enterprise #GenAI with new smaller models for latency + cost-sensitive use cases and larger multimodal models for visual understanding. Customers can easily get started tuning the Llama 3.2 models securely and efficiently on their data. Available today in Databricks Mosaic AI. 

7