Skip to content

LLMs Related Playlist

🔥 Unleash the Power of Language Models! 💬✨ Dive into the world of cutting-edge AI with our Trending Concepts in LLMs playlist! 🚀 From mind-bending implementations to mind-blowing breakthroughs, we've curated the ultimate collection of videos that explore the latest and greatest in Language Model advancements. Whether you're an AI enthusiast, a tech guru, or just curious about the future, join us on a journey through the hottest trends shaping the world of AI-powered communication. Get ready to be amazed, inspired, and future-ready – press play now and ride the wave of linguistic innovation!

Videos available

Here are the videos available in the playlist

Video Name Link
Finetune Llama2 using QLoRA Link
Deploy Local LLM like a ChatGPT Chatbot Link
DPO Implementation Link
Idefics Image Captioner Link
CodeLlama Finetuning and CPU Inferencing Link
Evaluation of LLMs Is All You Need Link

1. Finetune Llama2 using QLoRA

🚀 LLM Model Finetuning in Minutes: Hands-On Guide with Llama2 - 7b!

📹 Unlock the secrets of custom dataset finetuning with the recently launched Llama2 - 7b parameters model from Meta AI. Join us in this video tutorial as we navigate through the process using just 50 lines of code, all within the convenience of Google Colab's free environment.

💡 Brace yourself for an exciting journey! Learn how to harness the power of QLoRA to quantize the model down to 4-bit precision. From there, we'll dive into the heart of the tutorial – finetuning the Llama2 - 7b model on your very own custom dataset.

⚙️ In a matter of minutes, you'll gain hands-on experience, enabling you to unlock the true potential of this cutting-edge model. Don't miss out on this incredible opportunity to elevate your AI skills. Hit play now and revolutionize your understanding of model customization in just a few simple steps. 🌈🤖

Resources: Notebook

2. Deploy Local LLM like a ChatGPT Chatbot

🚀 Create Your Own Local LLM Chatbots: Step-by-Step Deployment Guide!

📹 Get ready to build and run your very own chatbots using large language models (LLMs) on your own computer! In this video, I'll walk you through every step. We'll use the SantaCoder 1B model to help you write Python programs in a chat-like format.

💡 Watch and learn as I show you how to set up the SantaCoder 1B model as a chatbot. By the end, you'll have a chatbot similar to ChatGPT that can assist you with programming tasks.

⚙️ But that's not all! You can take it further by deploying your own models. Just by specifying the model's path in your local directory, you can make your very own custom chatbot.

🌟 Whether you're new to programming or a tech enthusiast, this video simplifies the process of creating interactive chatbots. With easy-to-follow steps, you'll be able to craft your own chatbot that suits your needs. Ready to get started? Hit play and begin your journey into chatbot creation! 🤖🚀

Resources: To do

3. DPO - Direct Preference Optimization Paper Implementation

🔥 Unleash DPO Power: Step-by-Step Implementation Guide using TRL Library on FREE Google Colab! 🚀

📚 Experience the magic of AI optimization hands-on in this electrifying second installment on DPO – Direct Preference Optimization! Witness the groundbreaking TRL (Transformer Reinforcement Learning) library by @HuggingFace in action, right here on Google Colab for FREE.

🌟 Brace yourself for an immersive journey as we demystify the implementation of DPO. Follow along as I guide you through the entire training pipeline, showcasing each intricate step on the SantaCoder1B model – all within the accessible realms of Google Colab's free version!

⚙️ Unlock the secret to DPO implementation with two pivotal steps. Step one: the training of an SFT (Supervised Fine-Tuning) model. Discover how we harness the prowess of the SFT Trainer to skillfully mold the model on the Dahoas/full-hh-rlhf dataset, setting the stage for DPO's transformative power.

✨ But wait, the journey doesn't stop there! Step two is where the true magic happens. Immerse yourself in the art of creating the DPO model using the prowess of the SFT model, all while enjoying the computational efficiency made possible by the ingenious QLoRA technique, perfectly tailored for Google Colab.

🚀 Don't miss this golden opportunity to witness AI optimization unfold, right at your fingertips. Whether you're an experienced coder or an enthusiastic learner, this video is your golden ticket to mastering DPO implementation, revolutionizing the way you enhance your models.

🌈 Ready to embark on a journey that merges theory with hands-on practice? Hit that play button and join me in unraveling the marvels of DPO implementation – all done on the Google Colab platform. Your voyage to becoming an AI implementation virtuoso starts NOW! 🎓🤖

Resources: SFT NB Train NB Inference NB

4. IDEFICS - Image Captioning

🔥 Brace yourselves for a mind-bending journey into the future of AI! Join us in this captivating video as we dive deep into the groundbreaking State-of-The-Art multimodal model - IDEFICS. 🌌

What if I told you that there's an open-access gem that rivals Google Deepmind's closely guarded Flamingo model? Enter IDEFICS, a cutting-edge creation from HuggingFace that's set to redefine the AI landscape. 🚀

From Image Question Answering to Image Captioning and beyond, the possibilities of IDEFICS are limitless. Curious about how to fine-tune this incredible model for image captioning on the MSCOCO dataset? We've got you covered, all within the free Google Colab platform. 📸✨

But wait, there's more! We're not stopping at image captioning. Witness firsthand how the versatile IDEFICS can seamlessly adapt to diverse use cases involving both text and images. 📚🖼️

Here's the million-dollar question: Is multimodality truly the path forward in AI evolution? Get ready to ponder as we not only unravel the enigma of IDEFICS but also explore the very future of AI itself. 🤖🔮

🔔 Don't miss this chance to broaden your horizons and witness the next big leap in AI technology. Click now to embark on a journey that might just reshape your perception of what's possible. Your future in AI begins now. 🌐🚀

Resources: Notebook

5. CodeLlama Finetuning and CPU Inferencing

🚀 Ready to take your coding game to new heights? 🔥 Join me in this thrilling tutorial where we unravel the secrets of finetuning the state-of-the-art coding model, Codellama, using Google Colab's free version! 🖥️💡

🔍 Dive deep into the art of refining Codellama with the power of QLoRA in this step-by-step guide. 🎨✨ Learn how to masterfully fine-tune Codellama using QLoRA, while making the most of the free resources in Google Colab. 💻📈

And the excitement doesn't stop there! 🌟 As a special treat, I'll walk you through the exhilarating process of performing CPU inference for Llama 7b using GGUF. 🦙💨💼 Watch the magic unfold as we showcase the potential of your local CPU through a Gradio Chat Interface. 🎉🤖

💯 Elevate your coding skills without breaking a sweat! Whether you're a novice or a coding maestro, this video is your golden opportunity. 🎟️🔑 Let's conquer Codellama together, revolutionizing your coding experience one step at a time! 🌐📝

🔔 Don't miss out! Remember to like, subscribe, and ring the notification bell to stay updated on the latest coding hacks and tech tutorials! 🛎️🔗 Get ready to embark on this coding journey with us. See you in the video! 🚀👨‍💻👩‍💻

Resources: Finetune NB CPU Inference NB

6. Evaluation of LLMs Is All You Need

🔍 Unlocking the Unseen: Delve into the world of LLM evaluation, a facet often underestimated yet holding the power to transform your model's prowess! 💎

🔢 Crucial Queries Explored:

  • Why Evaluate LLMs? Discover the critical need for evaluation in maximizing model potential. Uncover why this step is a cornerstone of success, and how it can take your model from good to extraordinary.

  • What to Evaluate? Explore the multifaceted approach to evaluation. Understand the dimensions that demand attention – from coherence to context handling – and how they collectively shape model effectiveness.

  • Where to Evaluate From? Learn about the datasets that fuel evaluation. Explore the diverse sources required to gauge the model's performance across various aspects addressed in the 'What' section.

  • How to Execute Evaluation? Unveil the strategies and techniques to conduct effective evaluation. From methodologies to metrics, this section provides the roadmap to assessing and enhancing model powers

🚀 Elevate Your Model: Elevating your model's performance is just a click away! Through comprehensive evaluation, you can turn setbacks into triumphs, failures into stepping stones towards success. 📈 This video is your complete A-to-Z guide, providing insights that can reshape your approach to LLMs.

🎥 Embrace Transformation: Ready to unlock the full potential of LLM evaluation? Click the link below to watch the video and gain insights that can reshape your approach to LLMs.

Resources: Eval Page