is it possible

#1

Fluxcicide

tryna make a small scale ai ( not really an ai, idk what u can call) which can read books (max 2 ) and give me questions based on it ( i want hands on thats why im making it from scratch)

#2

morgue

1

Frags

+

–

i believe in you

#6

Fluxcicide

1

Frags

+

–

thank u morgue

#9

Aquemini

0

Frags

+

–

In python you can do that with like two work days. First half with pdfplumber, then train your model with nlp and some scikit learn libraries. Also i would recomend to give some sort of guideline with a regresion rather than a clustering straight away.

#3

1nshh

0

Frags

+

–

I mean can't u just ask chatgpt to give u questions

#4

Fluxcicide

0

Frags

+

–

yes but i wanna make smt, i wanna make ts fun ( cs student)

#5

seven_glazer

0

Frags

+

–

some indian/chinese dude probably made it already

#7

Fluxcicide

0

Frags

+

–

ehhh ,i mean sure

#8

Ldrago

0

Frags

+

–

i have made exactly what u have in mind the method is called Retrieval-Augmented Generation (RAG) . to keep it short, basically what ever books u have u convert them to vectors, the input query will also we converted to vectors. Then u do a simple similarity search between the input query and the vectors of ur book , take the top most matching query ,parse it and there is ur o/p