As someone working in Infrastructure side, supporting Azure virtual desktops for more than 6+ years and trying to make a mark in Gen AI, I came across this course for LLM on Hugging face. However unlike other LLM basic courses that solely focused on LLM models and prompting, this one is so advanced (kind of). and It focusses on Transformers, NLP etc. Cramming stuff isn’t possible given my limited time owing to work schedule of 10 hours. How best can I prepare this? I have completed Data science, and have a fair idea about Gen AI, LLM, Prompting etc.
1 Like
The LLM course (probably a revamped version of the old NLP course that merged with LLM-related resources?) is more about how to build and fine-tune LLM models rather than how to use them…
Upon completing this course, you will have the necessary knowledge to fine-tune LM and LLM models as long as you have access to a GPU.
Well, if you have a foundation in programming and data science, and ideally some mathematical background, there shouldn’t be anything too unreasonable, so I think you should be able to handle it normally. Having a local GPU makes things easier, but the course is designed to be completed using Colab alone. (However, due to Colab version updates, dependencies may change, causing samples to stop working. In such cases, you can skip them or specify the library version to work around the issue… 
)
Additionally, even if you don’t run the code, just reading through it and mentally executing it in your head might give you a general idea of what’s happening… ?
For code where you can imagine the results, avoiding execution might save time.
If you’re looking to learn about actual use cases for LLMs, the Agent course might be more practical. That course treats LLMs as black-box components.