I took this course because I like the ML Specification of Andrew Ng very much. For me, that was really the gold standard for MOOC. This LLM course was, unfortunately, a bit disappointing. The labs were underwhelming. To me, not really grading the labs looks like the lack of engagement on the part of the authors. I also had technical problems with the third lab -- I think, some issues with hard-coded versions of the packages -- and did not manage to run it at all! I also missed a forum to interact with fellow students and the instructors. I am still very thankful for this course, it is really hard to make one at this stage. I especially liked that authors went into technical details and provided links to the original papers. I see this course as a germ of a future cool GenAI specification. I am confident that the authors are capable of developing it further.

"Intermediate" level in the sense that you perhaps need some basic understanding of machine learning, but this is definitely not a course that challenges you. You get a very high level conceptual explanation of basic concepts (including things like LoRA, RLHF), but definitely no specifics on the implementation level. The assignment "Labs" consists of executing pre-written code in notebooks, and seeing the result output. No coding of your own, and typically just making function calls to Huggingface libraries, but not actually seeing how the algorithms are implemented.

Lots of theory with very little practice. You will not walk away from this course feeling confident that you know how to code any of it. The labs that are offered do not teach you much either. The instructors just tell you to not worry about how it works, and that it just works. The instructors need to add a whole lot more practice code to get you practicing the theory they teach. At this point, I am looking for where I can find that information because theory without practice is pointless.

The course material was all pretty superficial; the lectures never really delve into the nitty gritty details. The labs require no coding, which is disappointing. That being said, it's a good overview of the current landscape. If you want to learn implementation and the way things work (the how rather than the what), you will probably be disappointed.

While the course content and organization was great, I had issues in accessing the AWS labs (Week 2 and Week 3) where I couldn't execute the Python notebook steps after few steps and got stuck. When I tried to resume by restarting the terminal it said invalid authentication and could not complete the labs and had to do forceful submission.

The videos in the course were helpful, with the exception of the lab videos.

I found those simply regurgitated what was already in the lab notes.

The labs themselves were only partially helpful due to the high-level code being used.

I would have actually preferred to go a bit lower-level in implementing a few pieces.

I think the course needs to have a better balance between the contents. For example, it spends tons of effort talking about different parameters in model inference, which is as simple as 1+1, but it touches very lightly regarding how transformer works.

The lab parts do not make the student work nor present a challenge, they only make the student run through someone else's code.

(+) Excellent info, representation and organization

(-) The practical part is not good (some hands-on need to be added)

Week 1 and 2 are great but 3 isn't that exciting. I was expecting the course to be more technical.

Helpful introduction to LLMs but I wish we got the chance to go in-depth on implementation

A lot of issues with the labs. Contacted supported and waited for long but no resolution

Not explained as clearly as you would expect from an Andrew Ng course.

Not very technical course, but gives quite an overview

Not too practical Just a broad view on the subjecgt

Theory was good, but labs should be more practical

its LLM 101 explaining intuitions behind it

need more practic labs and videos.

more hands on

Most of the course content was quite shallow, only skimming the surface of each topic. I felt that it was a good primer on LLMs and I'm glad I took the course, but if I could go back in time I'd simply audit it. The labs and quizzes add very little value and are in no way worth $50.

Poor explanations. Can't hold the attention of the student. Start with problems and solve them using Generative AI and gradually explain different ways and concepts.

I cant acess the lab 3!!! Show this. "AWS account deactivated at 2023-12-29T02:15:01-08:00" This is not fair!!! I paid for this course. Can't finish it.

The Amazon SageMaker lab is hit or miss. It wasn't functional for me 80% of time.

Lab2 doesn't work. The kernel crashes at one place all the time.

i was not able to finish at 99 percent