Marqo

Meta AI just released Llama3.1, one of the most capable, openly available LLM’s to date 🔥 The models come as a collection of 8B, 70B and 405B parameter size models with 8B best suited for limited computational resources (like local deployment) and 405B setting new standards for AI. To test out its capabilities, I created a simple, local RAG demo with Llama 3.1 8B and Marqo. The project allows you to locally run a RAG Knowledge Question and Answering System 🚀 The demo also works with a variety of other models. Give it a try! Full walkthrough article: https://lnkd.in/e3MBc3PH Code: https://lnkd.in/etui9CxE #llama #llama3_1 #meta #ai #ml #model

Meta AI just released Llama 3.1, one of the most capable openly available LLM to date, with claims that it outperforms GPT-4o and Anthropic Claude 3.5 Sonnet on various benchmarks. Here's a tutorial for how to build a simple, local RAG application using Llama 3.1 8B off HuggingFace, complete with a GitHub repo of python code, written by Ellie Sleightholm. Let us know what you think in the comments! https://lnkd.in/gJDSTGsh #Meta #AI #Llama #LLM #Anthropic #HuggingFace #GitHub #Python #RAG

A full house for our talk on "Driving Revenue Through Hyper-Personalised Search and Recommendations" at Online Retailer Conference & Expo today! Our CEO, Tom Hamer 🦛, discussed: - Our success stories with customers such as Redbubble and Temple & Webster - How Marqo drives user engagement - Why personalised discovery is so important in ecommerce - How Marqo adapts based on user behaviour to adjust personalisation - Plus more! To discover how you can use Marqo to help drive more sales on your website, book a demo with us today: https://marqo.ai/book-demo More on our success story with Redbubble: https://lnkd.in/emKc2sG8 #OnlineRetailer #Ecommerce #Search

Don't miss our talks today at Online Retailer Conference & Expo! Tom Hamer 🦛 will be giving a presentation on "Driving Revenue Through Hyper-Personalised Search and Recommendations" at 14:00-14:30 ACT. Jesse N. Clark will be taking part in a Fireside Chat at 14:35-15:05 ACT with Briony Kennedy on "The Power of Authentic Personalisation - Lessons Learned on the Importance of Deeply Understanding Your Customers' Needs". The Marqo team are also at Stand G104. Come and chat with us about all things Ecommerce Search! More information: https://lnkd.in/e5gGrNPA #OnlineRetailer #EcommerceSearch #Ecommerce #VectorSearch #Search

Marqo have made fine-tuning embedding models so much easier 🤗🔥 You can select from a wide variety of CLIP and Sentence Transformer models to fine-tune. Marqtune also supports Generalized Contrastive Learning (GCL), Marqo's internally developed embedding model training regime for user interaction data. With this user interaction data, Marqtune optimizes the model for the queries that your business handles and the results your customers care about. These fine-tuned models bring rank and query awareness directly into the embedding model. Pretty cool! With the recent announcement of Meta's Llama 3.1, it's more prevalent than ever that fine-tuning such models is going to be the key to business success. Full Marqtune code walkthrough: https://lnkd.in/eY9F6nUH #ML #AI #MachineLearning #FineTune #Embeddings #FineTuning #Tuning

We released an embedding model fine tuning API today. Using Marqtune, your user interaction data optimizes the model for the queries that your business handles and the results your customers care about. These fine-tuned models bring rank and query awareness directly into the embedding model. We do this by using Generalized Contrastive Learning, an approach we've pioneered ( paper here: arxiv.org/abs/2404.08535 ). In our internal evals, our GCL trained model has greater than 200% improvements in nDCG vs. pretrained -- even when comparing ViT-L-14 to ViT-B-32. In the real world, we've seen GCL models take embedding retrieval from failed pretrained model experiments to revenue generating. Getting started is simple. You can specify your dataset, which business objective or metadata is important for ranking, choose from dozens of base models and checkpoints, and get out a usable model. Request Access to Marqtune: https://lnkd.in/ejJk6yYG Getting Started with Marqtune: https://lnkd.in/eEMabkcv GCL results: https://lnkd.in/en-c2HuX

We recently announced the launch of our embedding model training platform, Marqtune. With Marqtune, we’re making it easier to improve search relevance based on your specific business needs. To help you get set up with Marqtune, we created a 'Getting Started with Marqtune' guide. It walks you through the process of fine-tuning a pre-trained model using a multi-modal dataset. It covers: ⭐️ Setting up datasets in Marqtune ⭐️ Fine-tuning a pre-trained model with a training dataset ⭐️ Evaluating models with an evaluation dataset ⭐️ Downloading fine-tuned models Get started with Marqtune today: https://buff.ly/3A0lgAe Links: Request Access to Marqtune: https://buff.ly/3YeXk6q Marqtune Announcement: https://buff.ly/3LAFcMO Join our Community: https://buff.ly/4cIbKjZ #VectorSearch #Search #InformationRetrieval #Embeddings #Models #AI #MachineLearning

We are delighted to introduce Marqtune, the embedding model training platform that will change the game for search systems. With Marqtune, we’re making it easier to improve search relevance based on your specific business needs. We’ve built Marqtune on the foundation of our innovative training framework, Generalized Contrastive Learning (GCL). With GCL, you can fine-tune embedding models to rank search results not only by semantic relevance but also by a ranking system defined by your search team. This means better, more relevant search results that cater to your business. “Marqtune has been developed in response to feedback from our customers. Every vector search system in production needs to have its models continuously retrained and updated. Doing this manually is simply not feasible,” said Tom Hamer 🦛, our CEO and co-founder. “Marqtune introduces a user-friendly process for fine-tuning embedding models and allows users to achieve significant improvements in search relevance with minimal engineering effort.” For more information and to get started with Marqtune, read our full article: https://lnkd.in/esX8-EEE Request Access to Marqtune: https://lnkd.in/ejJk6yYG Getting Started with Marqtune: https://lnkd.in/eEMabkcv Join our Community: https://lnkd.in/e2hxAnJR #VectorSearch #Search #InformationRetrieval #Embeddings #Models #AI #MachineLearning

We walked around the booths getting set up today for #OnlineRetailer, and although we're admittedly biased, we probably have the coolest giveaway of any vendor 🔥 Book a Marqo demo through the QR code below for your chance to win a fine dining experience at a restaurant of your choice! The Marqo team will be at Stand G104. Come and chat with us about all things Ecommerce Search! #Ecommerce #Search #VectorSearch Online Retailer Conference & Expo

We're speaking at #OnlineRetailer! Our CTO & Co-Founder, Jesse N. Clark, will be taking part in a Fireside Chat at Online Retailer Conference & Expo with Briony Kennedy on Thursday 25th July, 14:35-15:05 ACT. The discussion is on 'The Power of Authentic Personalisation - Lessons Learned on the Importance of Deeply Understanding Your Customers' Needs'! The Marqo team will be at Stand G104 during the conference. Come and chat with us about all things Ecommerce Search! For more information: https://buff.ly/3Y3MW1j Join our community: https://buff.ly/4cIbKjZ #OnlineRetailer #EcommerceSearch #Ecommerce