rustformers / llm

[Unmaintained, see README] An ecosystem of Rust libraries for working with large language models

Home Page:https://docs.rs/llm/latest/llm/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Video Tutorials

okpatil4u opened this issue · comments

It is possible that this will be the first framework for a lot of people who are entering into the field. Is it possible to create a video tutorial series such as Andrej's for the newcomers ? This will improve the adoptability by a huge margin.

I could be persuaded into taking this on. Do you have any thoughts on the types of topics you'd like to see covered?

Haha, thanks @danforbes

Tutorials such as how would anyone build GPT-2 inference engine using llm ? Then increasing the difficulty level bit by bit. Introducing maybe T5, Bloom then llama. What kind of architecture has been mentioned in the actual publication ? How would you implement them in the llm.rs ? How does rust's borrowing / traits help / hinder one in this process ? How would one validate / test the results ? What kind of tests one needs to build in order to avoid common pitfalls / failures ?

The thought process that goes on in the expert's head while looking the problem at bird's eye view and then deep dive into specifics of the implementation.

Apologies for scattered ramblings. I was typing it as it was coming to my mind.

I actually came across this crate from this tutorial though it's mostly focused on leptos than the llm crate. So, I'd also appreciate it.