Can someone suggest to me where to Learn Machine Learning PROPERLY?

[rare chasm]

Basically, I'm trying to learn from youtube, and maybe I didn't found better and proper tutorial that can makes me really understand.

I really want to train AI ( it's not always about llm, I also want to train bot for boxing in vr for example, AI for data analysis, and many more ) from start, but want to learn the basic so I can do it the right way ( as right as possible ).

Where can I learn this? I prefer the free one because of my money condition right now. But the paying one ( if really great ) also welcomed for the future :).

Thank youuuu!

[plain mauve]

+1 to this post. @mild pagoda, @rotund berry, @rancid hemlock, @honest zodiac may have some good leads for us learners

[honest zodiac]

Hi @rare chasm, what have you already been watching? I have quite a big list of favorites on YouTube ranging from liquid nets, training your own LLM, Andre stuff is pretty great

If you're looking to have foundational understanding of the mathematics, I could point you that way too

[rare chasm]

Hi! I've been watching many things on reinforcement learning, study foundational math on brilliant.org ( when I still subscribed ), a little on make transformers from scratch ( which is too fast for me lol! I can't understand much in this video tutorial ). Also learn about tokenizer, NLP, bert, etc.

But again... I feel I need to know from the basics.. So yeah, maybe starts from foundational math is great. Is there a great tutorial on this? If about my path, I really want to make AI for boxing VR game, AI for basketball VR game. Anything that needs natural movement and not repetitive. That's my preference path.

I also love going into LLM, but feels more hopeless since it's needs too many data and really high end to try making a good one later.

[honest zodiac]

https://youtube.com/playlist?list=PLZHQObOWTQDNU6R1_67000Dx_ZCJB-3pi&si=xC6oogj9d6dFC0H3

Along with

https://youtu.be/cAkMcPfY_Ns?si=SezifhBPGNIQ8LO-

https://youtu.be/w8yWXqWQYmU?si=oYvXemc5sU1fsy7s

If you're thinking virtual reality, watch out for when the AI starts to exploit your physics models

[mild pagoda]

Hi @rare chasm, Gpt from scratch.

https://youtu.be/kCc8FmEb1nY?feature=shared

https://github.com/karpathy/ng-video-lecture

@rare chasm You can find more about AI/ML here in these links:

https://xbe.at/index.php?uid=f878df618f9c19161881816988b9dae3

https://github.com/xbeat/Machine-Learning

[rotund berry]

@rare chasm just a learning program option in case you're interested. I'm currently in it but haven't advanced to ML stuff yet. https://qwasar.io/tech-talent-programs/ai-machine-learning-engineer-program/

[rancid hemlock]

Hi @rare chasm , I'm not sure which stage you're at right now, but I can suggest some resources that can benefit various stages of learning.
All the above mentioned resources and videos are great, make sure to check them out.

My advice is to focus on something you want to do and set it as a goal for learning. Get comfortable with any deep learning framework of your choosing: Tensorflow, Pytorch, Jax, it doesn't really matter which one.

And as for learning material, this one is great in my opinion: https://d2l.ai/
Also you can check the various courses for different domains on Hugging Face: https://huggingface.co/learn

And don't forget to have fun ✌️

[mild pagoda]

@rare chasm You can ask a lot of questions to Hugging Face discord like about learning and so on. This is the link to Hugging Face discord:
https://discord.gg/hugging-face-879548962464493619

[mild pagoda]

@rare chasm Machine Learning by Prof. Andrew NG, here you can find the assignments and you can do the assignments from scratch using numpy (and others), if the assignments it is not done already. You can access from this link:
https://github.com/greyhatguy007/Machine-Learning-Specialization-Coursera

[rare chasm]

This is the one that I also try to learn, and lost in the middle of the video. << Suddenly can't process what he taught. lol!!! Too advance.

Woah!! There's so much resources. Thank you @plain mauve @honest zodiac @mild pagoda @rotund berry @rancid hemlock !!! Really thank you so much. @rotund berry , right now I can't learn anything paid with my money condition, but I'll be sure to check it out later.

I really feel so excited. Just sneek peek all the resources you all gave. Really, thank you so much. I wonder why I can't find most of the resources that you all share ><..

[rare chasm]

Hmm.. I don't understand why and how it can be happening right now... Will take a note of it :). Thankss

[honest zodiac]

It just means, if you have the spare parts to do it, try building a robot or a simplified model when you are training for physics

[mild pagoda]

@rare chasm Feed Forward Neural Networks from scratch using numpy and others libraries, this is the link:

https://medium.com/mlearning-ai/building-a-neural-network-zoo-from-scratch-feed-forward-neural-networks-f754cc88eca2

[rare chasm]

Ah I see. Okay2. I'll keep in mind :D. Thanks

[rare chasm]

Thank you! Will see this after watching some of more fundamental ( so I didn't get lost ) while watching it.

[rare chasm]

Read d2l.ai , just still in preface, already in love when I read this:
"In this book, we teach most concepts just in time. In other words, you will learn concepts at the very moment that they are needed to accomplish some practical end. While we take some time at the outset to teach fundamental preliminaries, like linear algebra and probability, we want you to taste the satisfaction of training your first model before worrying about more esoteric concepts."

❤️ Really thank you to share this gem.