Google Cloud’s Post

Google Cloud realize that strong and diverse LLM offering is a key to enterprise GenAI strategy and AI21 leads the way with #Jamba 1.5: - 2 models to fit your budget and needs - Mamba-Transformer MoE architecture designed for cost & efficiency - top quality: Jamba 1.5 Mini leads its size class with a 46.1 score on the Arena Hard benchmark, outperforming larger models like Mixtral 8x22B and Command-R+. Jamba 1.5 Large dominates with a 65.4 score, outpacing Llama 3.1 70B and 405B. - open weights - unmatched 256K effective context window - up to 2.5X faster including on long context than competition - enterprise features: function calling, JSON mode output, and more - available on cloud partners Vertex AI Model Garden, Microsoft Azure, and NVIDIA AI NIM and coming soon to Amazon Web Services (AWS) Bedrock, Databricks Marketplace, Snowflake Cortex, Together AI - available for private on-prem and VPC deployment

Are you at Google Cloud Next? Swing by the expo floor and check out the demo Jason De Lorme and I built to show off the Document Processing and Understanding solution. The demo showcases how you can combine the data and documents you already have with the power of AlloyDB, Vertex AI, and Cloud Run to build trustworthy Gen AI experiences for your customers. We also made the demo code available on GitHub so that you can deploy and customize it in your own environment. Check it out! https://lnkd.in/dyDSkiKg #googlecloudnext #genai #alloydb #vertexai #cloudrun

New #generativeAI news from Google Cloud! 1) Llama 3.1 models are now available on Vertex AI. This includes the new 405B model, the largest openly available foundation model to date, as well as 8B and 70B. 2) Google Cloud is the first hyperscaler to introduce Mistal AI’s Codestral, its first open-weight gen AI model explicitly designed for code generation tasks. We also added Mistral Large 2 and Mistral Nemo. 3) You can access the new models, and more, in just a few clicks using Model-as-a-Service, without any setup or infrastructure hassles. Learn more about what's exciting about these new models, how customers are using open models, and some strategies for success from Stephanie Wong, Head of Technical Marketing, and Warren Barkley, Senior Director of Product Management for Vertex AI ↓

Surely, not a single model services your needs whether it’s price, functionality, or whatever the future brings. #VertexAI continues to innovate and partner with that in mind. Llama 3.1, now available on our platform to service your #generativeAI requirements and powered by #GoogleCloud

Harnessing the Power of the Model Garden: The Benefits of Model as a Service Are you struggling to keep up with the ever-evolving landscape of AI models? 🌱 Model as a Service (MaaS) offers a solution by providing access to a curated collection of pre-trained models, often referred to as a model garden. Here's why MaaS is a game-changer: ➡ Efficiency: Quickly deploy and scale AI models without the need for extensive development or infrastructure. ➡ Accessibility: Access a wide range of models, from image recognition to natural language processing, tailored to your specific needs. ➡ Cost-effectiveness: Avoid the high costs associated with building and maintaining AI models in-house. ➡ Innovation: Stay ahead of the curve by leveraging the latest advancements in AI technology. These days, most companies are using a blend of open-source and proprietary models for a variety of use cases. Google Vertex AI allows users to conduct a model-by-model comparison to help you demonstrate side-by-side performance and cost factors across your models of choice. PRETTY helpful starting step to determine which model, deployment approach, cost, and architecture will work. Not to mention, this direction fits nicely with the AI agentic approach to help AI agents become increasingly capable of independent decision-making and goal-oriented behavior. Intelligent automation across multiple tasks is HERE. MaaS allows you to make the right model choice for your task given performance, cost, and accuracy and AI Agentic approach is the multiplier. #AI #MaaS #ModelGarden #DigitalTransformation

Exciting news from Google Cloud! 🚀 Check out the latest updates on #generativeAI: - Llama 3.1 models, including the groundbreaking 405B model, are now live on Vertex AI. With the largest openly available foundation model to date, also featuring 8B and 70B models. - Google Cloud leads as the first hyperscaler to unveil Mistal AI’s Codestral, a cutting-edge open-weight gen AI model crafted for code generation tasks. Introducing Mistral Large 2 and Mistral Nemo as well. - Access these innovative models effortlessly through Model-as-a-Service with just a few clicks, eliminating setup and infrastructure complexities. Learn more about the remarkable features of these new models, customer applications, and success strategies from industry experts Stephanie Wong, Head of Technical Marketing, and Warren Barkley, Senior Director of Product Management for Vertex AI. 🌟 #AI #GoogleCloud

Build a best-in-class RAG application with Qdrant! In this tutorial you'll see how to use: ✅ LlamaParse from LlamaIndex to parse documents ✅ Jina AI's latest embeddings to encode meaning ✅ Qdrant's hybrid cloud to store embeddings ✅ Mixtral 8x7b from Mistral AI to parse meaning It's all broken down step-by-step and even uses custom prompts! https://lnkd.in/gwKmNRHq

Great tutorial to get on track with AI and LLM APP deployment, specially if you are working with documents like PDFs using vector databases. #AI #LLM #RAG

Day 2 of Google cloud next conference in Las Vegas. Day 1 was very exciting, with Vids being the most interesting workspace release imo. For cloud it’s a whole different story: many new releases for making genAI production ready and scalable. Automated pipelines, model evaluations powered by AI and many more should take genAI app building to new heights. Very exciting times we live in! #cloudnext #newWayToCloud

"Proud to share the successful completion of Google's Cloud GenAI Jam Study, mastering the art of crafting advanced GenAI Apps using Gemini and Streamlit! 🚀 Verify it at ⬇️ https://lnkd.in/e8sbuFRK #GoogleCloud #GenAI #Gemini #Streamlit #AI"