I learned so much through the material and topics presented during this course. The topics are explained in detail and easy-to-understand way, and the labs and quizzes solidify the learning.

It is very informative and practical. It can really help machine learning engineers to understand and fine tune their own LLMs to adapt to various application scenarios.

Great overview of how to build, fine-tune and enhance the LLM model and how it can connect to the applications layer.

Wanted some more technical depth

This course had the potential to excel, but faced several shortcomings. Access to the labs was challenging, and it was unclear where to seek assistance, leading to a decline in my motivation. The course language and technical terminology presupposed prior knowledge, and the presentation lacked techniques, like metaphors, that could aid in retention. The labs, rather than being interactive and reinforcing the video content, were merely a series of clicks, lacking the engagement and educational depth found in Courseau's lab experiences. Additionally, the graded assignments posed their own set of issues. On several occasions, the system either altered my responses or failed to register them, necessitating multiple attempts to complete a task. This recurring problem suggested that the course might be more focused on learning from the users’ interactions rather than facilitating effective learning for the participants themselves.

Enlightening, Inspirational, and Riveting. Three words I wouldn't typically associate with online learning courses. Throughout the weeks of immersive lessons, the once mysterious structure of LLMs became clear and comprehensible to me. I was astonished to discover that the phrase "Attention Is All You Need" originates from a 2017 paper authored by individuals from Google and the University of Toronto. This transformative paper introduces a transformer architecture that employs an attention mechanism to assess the interconnections between words within a sentence.

While the course doesn't delve excessively into complex mathematics, a basic grasp of matrix arithmetic and calculus can be beneficial in understanding the foundational mechanisms utilized for modeling and training Artificial Neural Networks. Yet, it is important to note that this course is designed to cater to learners of all levels, whether they are novices or well-acquainted experts in the STEM field. While no coding is required during the labs, a working knowledge of the Python programming language can prove advantageous, as it serves as the primary tool for developing and analyzing these LLMs.

However, the course goes beyond the surface and delves into the full Generative A.I. lifecycle, which seeks to manage and direct the desired outcomes for specific applications. Concepts like PEFT (Parameter Efficient Fine Tuning) are thoroughly explored, enabling a significant reduction in the number of tuned parameters in the model, thereby leading to a smaller memory footprint and less computational strain.

Considering the emergence of this transformative and generation-defining technology, I firmly believe that it is crucial for individuals from various fields to grasp the intricacies of generative A.I. This collective understanding will enable us to wield its potential responsibly and ensure that it ultimately benefits humanity.

Furthermore, the course emphasizes the paramount importance of governance, regulation, and policy guidance in the training of these models. In its concluding segments, the course takes a serious approach to exploring strategies for mitigating unwanted A.I. behaviors, such as toxicity and misinformation, fostering a responsible and ethical approach to A.I. development.

A great introductory course to LLM's and generative AI. If you're somewhat technical (i.e. know python and have some basic knowledge of data science, for example NLP) and you want to get up to speed on the hot new topic of generative ai, this is a great course for you. It has a a lot of information packed into 3 weeks. The focus of the course is on improving an existing LLM to fit your specific needs (as opposed to creating LLM's from scratch). The lab assignments are done in AWS Sagemaker Jupyter notebooks. They're very straight forward follow-along style assignments which go over the mechanics of how to do various things with the LLM. The instructors are articulate and make the subject matter very approachable. Overall I found this course to be very helpful to get from nearly no knowledge, to a point of where I have enough to start doing lots of experiments.

This course has been an incredible journey, guiding me through the fascinating world of LLMs, from the fundamentals of the Transformers architecture to tackling complex computational challenges through fine-tuning. I also gained valuable insights into reinforcement learning from human feedback, which was both intriguing and intellectually rewarding. What truly impressed me was the course's comprehensive approach, featuring thoughtfully crafted quizzes and hands-on labs. The structured nature of the materials allowed me to dive deep into each week's content, and the example workbooks provided a practical bridge between theory and application. It's evident that the creators of this course put significant effort into ensuring a meaningful and engaging learning experience.

Another course from DeepLearning.AI that doesn't disappoint. 🔥 Course kicks things off with some classics:🏗️ LLM Architecture🧐 Attention Mechanism🏋️ LLMS Training🚀 Applications . But it doesn't stop there! It delves into what I believe is even more crucial for those of us in applied data science:💡 Fine-tuning strategies and their associated costs 💰💻 Viewing LLMS as pieces of software (which they really are)🔗 Integrating them into the rest of a tech company's stack with the LangChain libraryPlus, the talented instructors share some cool tricks for solving complex tasks with LLMs using in-context learning 🤓 and Chain-of-thought prompting. 🧠Highly recommended! 👌😄

This course was really worth it. It covers all current aspects of LLMs starting from prompt engineering to PEFT techniques like LoRA to advanced concepts such as RLHF with PPO, RLAIF and how RAG, ReAct and so on can be implemented with orchestrator models such as LangChain. This course also covers aspects which should be maintained in any responsible AI system and offers glimpses in current fields of research. Specially helpful were the actual hands-on implementations of these concepts in AWS sage maker studio. I recommend it to anyone who has some concepts of deep learning and transformers and have implemented some ML projects using python.

I went through few courses to know about a high level overview about Generative AI lifecycle. But this course has delivered 200% more than my expectations. Very well explained the most important concepts like transformers, fine tuning, Reinforcement learning with human feedback, the best part is handson even if I am not able to understand few concepts during lecture it has cleared while doing handson. I dont have prior knowledge of the concepts or machine learning before but I am able to understand very well. I can train llm model and delpoy it and also integrate with external applications. Thanks to the whole team who designed this course.

Maybe the best course in the world 2023: Most effective matter and structure: The best blend of simpleness and deep knowledge. It is scientific, it contains labs, it is not too complicated. Can be started with little pre-knowledge, but a high motivation! This is not a course scratching only at the surface. It presents some of the most important scientific papers in the LLM area, but in the most explainable way with very little pre-knowledge. Every aspect of Generative AI is mentioned. After this course you can decide where you want to deep dive further (programming and books; no courses on coursera will take you much deeper I fear)

The course "Generative AI with Large Language Models," made together by AWS and DeepLearning.AI, is a great way to learn about Generative AI. It covers both the ideas and how to use them in real life. With AWS, I get to do practical exercises, which makes learning more hands-on. The course also talks about the right ways to use AI, which is important. The faculty explains things clearly and knows a lot, making it easier to understand hard topics. This course is really useful for anyone who wants to learn more about advanced AI and the LLM development lifecycle, as it teaches both the theory and how to apply it in the real world.

This course is an absolute gem for anyone looking to dive into the world of Generative AI and LLMs. The content is presented in a remarkably clear and concise manner, making complex concepts accessible to learners of all backgrounds. The course structure is well-paced, ensuring a smooth learning journey. I was particularly impressed by the depth of coverage while maintaining a focus on practical applications. The instructor's expertise shines through in every module, making the learning experience both engaging and informative. Without a doubt, this is the best course I've encountered on the subject. Highly recommended!

I highly recommend this (Intermediate level) course to start with LLMs. Having a bit of knowledge in RL (Reinforcement Learning) will help you grasp some concepts covered in the course quickly (and correctly). I have no foundational knowledge of RL and therefore referred to this [https://karpathy.github.io/2016/05/31/rl/] beautiful article written by Andrej Karpathy. He nailed it as usual. Whether you are a researcher or an application developer, you have a lot to learn from this course. Thanks to DeepLearning.AI and the instructors of the course: Mike Chambers Antje Barth, Shelbee Eigenbrode and Chris Fregly

The program offers a comprehensive insight into the GenAI project lifecycle through a structured approach over 3 weeks. Whether you're an AI novice or an expert, this course is tailored to all experience levels. Each week covers a different project phase, featuring bite-sized videos, labs, quizzes, and additional reading materials for a deep dive into research papers. 📚🔍 This format allowed me to learn in sync with the project lifecycle and explore source research papers that piqued my interest. 👩💻📝 If you're eager to elevate your knowledge of GenAI, I highly recommend this course!

This is a very complete and insightful course on Generative AI and LLMs. It starts with the foundations of LLMs, and proceeds throughout the various phases of the Generative AI project lifecycle. The course is filled with comprehensive explanations, challenges, techniques to address them, and resources to explore hands-on and understand the various techniques. If you want to understand how Generative AI and Large Language Models work and how you can use them in your projects, this is a great course to get you up to speed in understanding, planning and implementation details.

Recommending this very solid foundational course I just completed on the current state GenAI with a focus on practical knowledge around fine-tuning, optimization, and using LLM in applications using frameworks like LangChain. It is supplemented with a number of hands-on exercises on Amazon SageMaker Studio, which is an amazing place to accelerate prototyping and app development with GenAI in a secure environment.

Would have liked to have additional hands-on lab for the last part (LangChain) or at least pointing to resources on GitHub with sample Jupyter notebooks / guidance.

The "Generative AI with Large Language Models" course proved to be a game-changer for my understanding of Generative AI. It offered a structured approach, blending theory with hands-on exercises on Large Language Models (LLMs). Key highlights included in-depth discussions on GPT models, practical use of tools like OpenAI and Hugging Face, and seamless integration with PyTorch and TensorFlow. The course fostered collaboration and provided a strong foundation for real-world applications. Highly recommended for anyone interested in mastering Generative AI with LLMs.

Some who is an AI engineer, as an user, who is trying to apply technology in AI/ML use cases, this course was quite detailed and very helpful. Really thankful to Professor Andrew NG, his team, whole AWS team and all other contributors.

AI/ML deeper development concepts were difficult to understand because of poor knowledge at that level, but the course gave a good go ahead information with what it takes to train and use the generative AI, some good general knowledge of the field, and use for applications (with Langchain and all the concepts)

Thank you again.

This is an amazing course, team taught by DeepLearning.ai and AWS, an intermediate level course that gives full explanations about architecture of LLMs, prompting, connecting applications, and labs to help give you a "hands on view" of how to string together the various features. Some Python is highly recommended, but no coding is required. Quizzes are fair, not easy but not hard, and reinforce key concepts. Highly recommended for "novice AI enthusiasts" as well as DIY application developers. Videos are well done, with slides and transcripts.

The course is great. Unfortunately, I don't think it gives me an education that allows me to think for myself. Too much cookbook. I'll have to invent my own projects, but by the time I am done, DPO, DiffEQ ML, 1.58 but LLM, and SSM/Mamba will have changed the whole game. It is less a matter of keeping up with tech and a more a matter of broader control, education and empowerment. Too many players, too much money... recall the experience of a new CS student with Unix and C/C++ and the empowerment there. How to bring it back for AI?

Amazing Course, covering the whole LLM Development Life Cycle , very good explanation for details, wished just some more focus on parameters for functions but that wasn't a problem since the main function was explained during the walkthrough for the lab, but the concepts are well covered, highly recommend but sure would need good ML and Generative AI concepts that would help in that course, though there was a quick pass over that part as well ( Attention is what you need paper , but quick pass not the whole details ,, abstract view )

Excellent! This class is great for many learning styles and includes lots of visual information and extremely well-planned lectures. The labs top it all off with experiential learning; the labs are well-explained and designed. They are understandable for many levels of learners and bring all the previously learned lecture information together. The lecturers are all fantastic- with information that builds making it accessible. This is an excellent course. Thank you to all those at AWS and Deep Learning AI who designed this class!

A great foundational course.

This course teaches about things like: Instruction fine-tuning, FLAN, PEFT, LoRA, RLHF/RLAIF, PPO, PAL, RAG, ReAct, optimization and lifecycle of LLMs.

The 3 labs use Hugging Face transformers and cover prompt engineering, full fine-tuning, LoRA fine-tuning, and tuning with RLHF to reduce "harmfulness" of responses.

Studying of this course should be accompanied with study/overview of research papers, since the LLMs area is rather advanced; the course includes quite a few such references