Together AI

🚀 Announcing the launch of Llama 3.2 and Llama Stack on Together AI, in partnership with AI at Meta. 🎉 We are excited to offer free access to the Llama 3.2 vision model for developers to build and innovate with open source AI. Start building with the Llama-Vision-Free model today: 👉 https://lnkd.in/gWxwdaVd ▶ What we are launching: - Free Llama 3.2 Vision Model (11B): Develop and experiment with our high-quality Llama-Vision-Free endpoint for multimodal tasks. - Together Turbo Inference Endpoints (11B, 90B): High performance and accuracy for tasks like image captioning, visual question answering, and image-text retrieval. - New Llama Stack APIs: Standardized APIs to simplify building agentic and retrieval-augmented generation (RAG) conversational apps. ▶ Unlock powerful use cases: - Interactive Agents: Build AI agents that process both image and text inputs Image Captioning: Create high-quality image descriptions for e-commerce and digital accessibility - Visual Search: Enable users to search via images, enhancing search efficiency for retail and e-commerce ▶ Industry applications: - Healthcare: Accelerate medical image analysis for faster diagnostics Retail & E-commerce: Revolutionize shopping with image-based search and personalized recommendations - Finance & Legal: Streamline workflows by analyzing visual and textual content to optimize contract reviews and audits 💡 Check out napkins.dev: Our open-source demo app uses Llama 3.2 vision to transform sketches and wireframes into React code! Try it out at https://napkins.dev 🔧 Get started today: Experiment with the Llama-Vision-Free endpoint, or build for production with Llama 3.2 Together Turbo endpoints. 🌟 Read more in the blog https://lnkd.in/gDfmXxzu

From Nov 8-10, the Llama Impact Hackathon brings together developers and AI enthusiasts to address critical challenges in healthcare, education, and the environment with Meta AI’s Llama 3.1 & 3.2. We’re proud to be a partner alongside AutogenAI, MindsDB, LlamaIndex, and others committed to advancing AI for real-world impact. 📍 Register here: SF Onsite: lu.ma/1mkqg22a Online: lnkd.in/drzm2d-P

New Cookbook: An open source implementation of NotebookLM from Google. In short it involves: 1. Structured decoding with Llama 3.1 70B on Together AI to extract a JSON script. 2. Cartesia Text to Speech to bring the script to life! Check it out below! https://lnkd.in/gcGYjGWE

Thrilled to announce Together AI’s sponsorship of The New Club’s Hack & Relax, a women-only retreat for women in engineering, taking place November 8-10 in beautiful Monterey! 🌊🚀 This retreat is an amazing opportunity to support women in tech at every career stage, whether you're looking to get started in AI or expand your skills. Together AI’s own Virginia Adams will be leading a hands-on workshop to show how to build and deploy AI applications using open-source models on our platform. It’s all about empowering women in engineering to learn, connect, and grow in an inspiring environment. Along with Virginia’s workshop, enjoy tech and career panels, peer networking, and a fantastic location. Secure your spot today! ➡️ https://lnkd.in/dgNv2F3E

🚨 New work on LLM quantization! Want even better, even faster quantized LLMs? Check out QTIP, a new work led by Albert Tseng in collaboration with Qingyao Sun, David Hou and Chris De Sa. QTIP stands for Quantization with Trellises and Incoherence Processing and builds upon the incoherence processing (IP) framework introduced in QuIP and QuIP#. QTIP adopts the “bitshift trellis” and introduces a series of compute-based pseudorandom Gaussian codes that use as few as 3 ALU instructions. Together, these enable QTIP to achieve high quality models and fast decoding. QTIP’s results show that trellis coding significantly improves over vector quantization, even after recovery-based fine-tuning. Even on Llama 3.1, QTIP is able to achieve strong results below 4 bits! Finally, QTIP’s decoding speed shows that these gains are essentially free on modern hardware. QTIP achieves the same fast inference speeds as QuIP# with significantly better quality. Check out the paper, code, and prequantized models for more! QTIP will appear at NeurIPS 2024 as a Spotlight 🔦 as well! Blog: https://lnkd.in/dz-VjpV7

📢 Excited to announce our upcoming webinar on November 6th, at 9AM PT: on"How We Built LlamaCoder (400k Users) – a Full-Stack AI App with Next.js"! Join Hassan El Mghari from our Dev Rel team as he dives into the story behind LlamaCoder, an open-source code generation tool with 400k users and 3k stars on GitHub. He’ll share exactly how he built it a single weekend, scaled it, and tips on creating your own full-stack AI apps with Next.js. Don’t miss out! Click the link to RSVP 👉https://lu.ma/euvemy0q

Join Hassan El Mghari from the Together AI Dev Rel team as he goes over how he built LlamaCoder – an open-source code generation tool where you can build small apps with a single prompt. It currently has over 400k users and 3k stars on GitHub. He’ll go over exactly how he built it in a single weekend, how he scaled it to 400k users, and advice for how to build your own full-stack AI apps with Next.js

We’re excited to share that our CEO, Vipul Ved Prakash, represented Together AI on an incredible panel titled "AI Disruption: The Bigger Picture" at #TecShift2024, the aramco Entrepreneurship Summit at Ithra in Dhahran, Saudi Arabia. Joined by fellow AI visionaries Aidan Gomez (CEO of Cohere), Alexander Ratner (CEO of Snorkel AI), and Yuan (Alan) Qi (CEO of INF TECH), Vipul explored AI’s transformative impact on healthcare—from diagnostic advancements to personalized treatments and the evolving role of healthcare practitioners in this technology-driven landscape. A big thank you to Aysar Tayeb of Prosperity7 Ventures for moderating this forward-looking discussion. #TecShift #FII8

New Cookbook: An open source implementation of Contextual RAG from Anthropic! In short, it involves: 1. Using Llama 3.2 3B to efficiently generate context for each chunk 2. Building two data indices: vector and BM25 3. Performing hybrid search using reciprocal rank fusion 4. Reranking hybrid search results 5. Generate with Llama 3.1 405B Check it out below! https://lnkd.in/gR9MniE5

👨🍳 New Cookbook: A step-by-step walkthrough of the PDF to Podcast workflow! In short it involves: 1. Structured decoding with Llama-3.1-70B on Together AI to extract a JSON script. 2. Cartesia TTS to bring the script to life! Notebook: https://lnkd.in/g7vS_4nV