Hello, when will the new course be available?

Thank you very much!!!!

I'll give it s try after watching your videos. I even subscribe to your channel I had my doubts about these AI agents but will give openHand(formerly called opendevin) and maybe SWE-Agent from Princeton

Thanks for this. I've been looking for a video it coding and solving GitHub issues

Hi Graham, when will the new video of the course come out?

board is not visible

Could you kindly upload the recitations as well? It would be greatly beneficial.

Thanks a lot for making these great materials to be open to public.

Thanks Prof .

Thanks for sharing. It helps me a lot!

Hi Professor, Thank you for the insightful lecture. I have one quick question. In the lecture, you mentioned that Llama2 is safer than MoE-based models such as Mixture of Experts (Mixtral). I wonder, with heavy context distillation or other tuning methods that Meta conducted, if these methods were applied to MoE structures, would the MoE-based model, such as Mixtral, gain comparable scores in safety? Or are there fundamental differences resulting from different structures that make dense models inherently better than MoE models in terms of safety?

Soooo quiet

I love the lectures by PhD students. This and the Prompting lecture (iirc) were fantastic

Prompt Tuning explanation @ 58:50 isn't very clear imo

This was an awesome lecture, thanks! Certainly the best in the course so far.

Thank you for sharing this knowledge publicly :)

Hello~ How to install opendevin? (which one macbook or windows?) I can't install windows and MacBook. Please explain the installation method in detail.

A very comprehensive overview of RAG related technologies, thanks!

which llm model and codeagent？

thanks for the very useful vids!! i'm wondering where lect 19 is

lol. "OpenAI has been doing sketchy things for a long time and look where they are". ya so true. thanks for the vids, they are amazing content!

Thanks professor. Tuition too expensive, lol

Thanks

Thanks

This is great channel for know about NLP and all concepts that have relation with NLP like transformer and etc ....

Thank you for your introducing interesting papers!

Thanks professor！This is the latest and most complete series of large-scale model courses that the whole RU-vid can see at present😃😃

Thank you so much professor! However, could you also release the rest videos and the 16th video? Thanks in advance!

great lecture！

I've really enjoyed the 14 episodes of this series so far - each one has been incredibly insightful and helpful. I noticed that there haven't been any new uploads recently, and I'm eagerly looking forward to the next episodes. Can you please let us know if you plan to continue the series? Many of us are excited to learn more and can't wait to see what's next. Thanks for all your hard work!

Thanks for sharing！

There is no sound?

Its hard to see the board, what is the version of bilinear attention that is commonly used mentioned at 54:25?

Hi Professor Neubig, really like your advanced nlp series, would you upload the latest lecture videos? Thank you.

Thank you for sharing!

theres a mistake on the slide at 3:58 in the interpolation smoothing. the index should be i-n+2 not 1-n+2.

thank you for the fantastic video! great slides and up to date explanation

waiting for next sessions ....

Hi, just a quick question. I noticed that the new playlists e.g. 2024 don't cover old approaches in NLP. For someone who don't have any background in NLP would you advice me to begin with the 2017 playlist or should I begin from 2024 playlist?

Nice Lecture Prof . Thanks for making it available for us.

It's great content. Will you continue uploading subsequent lectures? Thanks!

興味深い講義ありがとうございました。 16:15 In Attention is All You Need, I was under the impressions that the number of vectors in K and Q had to be the same in order to do the dot product. Or is permissible for K and Q to be different because we are talking about decoders?

thank u for give us these lectures u help me alot

Can we have the link for Facebook OPT lessons? (Here 1:02:15 )

Thanks a lot for sharing these! Will you consider making the camera screen a bit larger? It's really hard to see what's being written on the blackboard.

Thank you. It's really useful

What a legend

Thank you so much

is these assessments are free to write and get the certficate

a lot of helpful information out there, thanks for sharing Graham, I really couldn't resists watching all of the video. If you could consider making another lecture about Jailbreaking/Injections it would be nice.