Large Language Models (LLMs) - Level 3

Meet Sanyam Bhutani, your guide for the following modules, who will assist you in crafting your own GPT with h2oGPT. Here's a glimpse of what's coming up: 💡 H2OGPT: Construct Your Offline ChatGPT: No internet connection required, giving you complete control. 🚀 Insider Tips for Your GPT Journey: Valuable insights to enhance your GPT creation experience. 💻 Fine-Tuning with H2O LLM Studio: A user-friendly, code-free tool to perfect your model. 📝 Put Your Skills to the Test: Challenge yourself with an engaging assignment and craft your own GPT. Stay tuned for the next modules and don't forget to subscribe! 📢 PS: The entire series is available in chronological order in the following playlists: https://www.youtube.com/playlist?list=PLNtMya54qvOHQHDpUDtZytwEV2Miali9l https://www.youtube.com/playlist?list=PLNtMya54qvOGK6vKlkU9_7hUvpJNKxFri https://www.youtube.com/playlist?list=PLNtMya54qvOHreXRc7JL0qMWGDYlLjz7u

🌟 Join us for a new session where we will learn more about the key considerations of making your own GPT with h2oGPT! 🤓 In this video, you will learn more about: Why grounding the model matters and how it enhances responses 🌐 The amazing capabilities of H2OGPT - our GPT open-source tool 🚀 A live demo to show you how it all works 🎥. You can watch the entire webinar of Rob Mulla here: https://youtu.be/2orLCp984HA?si=auSwE9We4QTN_Ld8 PS: This video is a part of a series published on our LLM Learning Path Playlist, that you can check out here: https://youtube.com/playlist?list=PLNtMya54qvOHQHDpUDtZytwEV2Miali9l&si=lIXmA0hqGZhftSZe

You can find our h2oGPT GitHub Repository here: https://github.com/h2oai/h2ogpt Welcome to this exciting demo video created by Andreea Turcu, our customer data scientist and instructor, where we will delve into the world of h2oGPT and explore the future of open-source AI! In this video, we'll cover a range of fascinating topics that will expand your knowledge and unleash the power of h2oGPT. First, we'll set the stage by providing a bit of context to help you understand the significance of h2oGPT in the AI landscape. Then, we'll dive into the concept of GPTs (Generative Pre-trained Transformers) and explain why it's crucial to familiarize yourself with Language Model Transformers (LLMs). To truly appreciate the capabilities of h2oGPT, we'll explore its origins and highlight the unique features that set it apart from other models. You'll gain a deep understanding of what makes h2oGPT an exceptional tool for various applications. One of the key aspects we'll discuss is how h2oGPT can significantly boost your productivity. You'll discover how this open-source AI model empowers you to accomplish complex tasks efficiently and effectively. No model is perfect, and we'll address the limitations of existing models on the market. By understanding these limitations, you'll be better equipped to harness the strengths of h2oGPT while mitigating potential challenges. The video will also highlight the numerous benefits of open-source models like h2oGPT. From flexibility and customization to collaborative development, open-source AI models offer a wealth of advantages that propel innovation in the field. Finally, Andreea Turcu will provide an engaging demo of h2oGPT in action. You'll witness firsthand the incredible capabilities of this open-source AI model, leaving you inspired and eager to explore its potential. Join us on this exhilarating journey as we unleash the power of h2oGPT and dive headfirst into the future of open-source AI. Don't miss out on this opportunity to expand your knowledge and discover the possibilities that h2oGPT holds. Hit the play button and let's get started! PS: This video is a part of a series published on our LLM Learning Path Playlist, that you can check out here: https://youtube.com/playlist?list=PLNtMya54qvOHQHDpUDtZytwEV2Miali9l&si=lIXmA0hqGZhftSZe

In this video I show I was able to install an open source Large Language Model (LLM) called h2oGPT on my local computer for 100% private, 100% local chat with a GPT. Links * h2o website: https://h2o.ai/ * h2oGPT UI: https://falcon.h2o.ai/ * h2oGPT GM-UI : https://gpt-gm.h2o.ai/ * h2oGPT github repo: https://github.com/h2oai/h2ogpt * h2o Discord: https://discord.gg/WKhYMWcVbq Timeline: 00:00 100% Local Private GPT 01:01 Try h2oGPT Now 02:03 h2oGPT Github and Paper 03:11 Model Parameters 04:18 Falcon Foundational Models 06:34 Cloning the h2oGPT Repo 07:30 Installing Requirements 09:48 Running CLI 11:13 Running h2oGPT UI 12:20 Linking to Local Files 14:14 Why Open Source LLMs? Links to my stuff: * Youtube: https://youtube.com/@robmulla?sub_confirmation=1 * Discord: https://discord.gg/HZszek7DQc * Twitch: https://www.twitch.tv/medallionstallion_ * Twitter: https://twitter.com/Rob_Mulla * Kaggle: https://www.kaggle.com/robikscube

This video serves as a concise refresher for fine-tuning in H2O LLM Studio, a condensed version of our earlier webinar led by our Head of Global Training, Andreea Turcu. For a more comprehensive review, please check out the detailed version here: https://youtu.be/u48QaIAIFw4 💡 Want to try H2O LLM Studio hands-on without setup? Watch our Aquarium walkthrough here: https://youtu.be/FSBlJeSadgw PS: This video is a part of a series published on our LLM Learning Path Playlist, that you can check out here: https://youtube.com/playlist?list=PLNtMya54qvOHQHDpUDtZytwEV2Miali9l&si=lIXmA0hqGZhftSZe

Take a look at the latest updates in H2O.ai Aquarium, featuring a refreshed user interface and seamless integration with H2O.ai University. These enhancements make it easier than ever to access hands-on labs and learning resources, all in one place. ▶ Start exploring Aquarium here: https://aquarium.h2o.ai/ ▶ Learn more at H2O.ai University: https://h2o.ai/university/

This is a short video describing what your practical assignment is: to create your own GPT model. Here's how: Just before starting it, you will will need to: 1. Install LLM Studio and load datasets from Kaggle and Hugging Face. 2. Train your model on your chosen backbone. That's it! Good luck in experimenting and creating your first GPT model! If interested, please check out the solution here: https://youtu.be/HFMSqOZHSsg PS: This video is a part of a series published on our LLM Learning Path Playlist, that you can check out here: https://youtube.com/playlist?list=PLNtMya54qvOHQHDpUDtZytwEV2Miali9l&si=lIXmA0hqGZhftSZe

Happy that you made it! Here's the solution to the earlier task, which is available right here: https://youtu.be/bDBpdfkrl-U 🔍 Pick the Perfect Dataset: Head over to Hugging Face and choose a conversation dataset that excites you. Look for conversation-themed datasets like mental health chats, function calling, or clinical conversations. 🚀 Get Creative, Ignore Licensing: For this assignment, don't worry too much about licensing. Just focus on getting your hands dirty with data. 📥 Download & Customize: Download your chosen dataset and trim it down to around 10,000 examples. This will save you from overwhelming your resources. 🛠️ Start the Experiment: Begin by randomly selecting a portion of the dataset and train your first GPT model, following the steps from the previous videos. ❓ Need Help? Join Us on Discord: If you hit any bumps in the road, leave a comment or reach out on Discord. We're here to help you out. 🙌 PS: This video is a part of a series published on our LLM Learning Path Playlist, that you can check out here: https://youtube.com/playlist?list=PLNtMya54qvOHQHDpUDtZytwEV2Miali9l&si=lIXmA0hqGZhftSZe

Join us on a new video based on Large Language Models (LLMs), where we will be discussing about the fascinating process of evaluating and benchmarking these LLM models. In this module, we'll explore: - How we measure LLM performance with various metrics - The challenges we face with common datasets and evaluation techniques We'll also showcase the user-friendly LLM EvalGPT platform, where you can explore rankings, evaluation methods, and select the best model for your needs.