This content describes a CS50 final project: a Resource-Augmented Generation (RAG) chatbot deployed on a personal website, designed to answer questions specifically about the creator, Taylor Stersz.
Key Points
Mind Map
Clicca per espandere
Clicca per esplorare la mappa mentale interattiva completa
Hello, my name is Taylor Stersz and this
video will be a very quick walkthrough
of my CS50 final project which I've
called the portfolio chatbot. Uh here's
my GitHub username, edex username, my
name, where I'm from, and the date the
video is recorded. Uh I've restarted the
server because I updated my database. So
I want to make sure that's fresh.
Okay, so uh let me jump over here and
say thanks real quick to the entire
staff. I've learned a lot despite having
a good deal of experience. Uh, one of
the things I learned was Python and I
wanted to put that to work in an AI
engineering project and this final
project was a great ex uh, excuse that.
Um, so this is a my final project is a
re resource augmented generation chatbot
uh that I put on my personal website. So
we've got my Python app here um, which
is a flask app that exposes a prompt
endpoint. Uh, I've got a lang chain
agent that uses some tools to augment
its responses. I implemented the chatbot
in my uh personal website. Um the Python
app is hosted in hugging face uh spaces
which is free. It's a ephemeral server.
Um I've got some vector uh embeddings
stored in Chrome cloud which is nearly
free. And then Google's cheapest model.
Uh so that's also nearly free. I'm
talking about a couple of cents a month.
So, I come over here and let's refresh
my personal site and I'll say hello
and it'll give me some information.
Then I'll say tell me about monkeys and
it won't be able to tell me about
monkeys because it can only tell me
about Taylor St. So I'll say tell me
about Taylor's Python experience of
which I have a tiny bit but I I learned
a lot in this class. So um yeah. Oh, I
think I added those recently. So, I jump
over here. This is the hugging face
space. Um, I expected to see some logs
there. Um, but I'll see some logs over
here. This is the Netlefi function
wrapper that I use to offiscate the
hugging face API token. Um, and Netlefi
is where I host the website itself. Uh,
this is the Chrome database. Uh I have
one collection stored with the uh
basically the information from these
docs which is my resume, some explicit
facts about me, some video game credits
since I work at Blizzard and my uh
GitHub profile. Uh this is my personal
website uh repo which last 10 commits or
something are me updating it. Uh and
this is the brand new repo I created for
the hugging face uh chatbot. Uh so let
me go ahead and uh just note like let's
say 240 is the first ID here. And I'll
just note that because uh I'll show how
I make the uh database.
This is a manual process to uh just save
me money basically and I only need to do
it sometimes. So created and populated
after it would have deleted everything
that was in there and we see that the
IDs change. So that's a quick walk
Clicca su qualsiasi testo o timestamp per andare direttamente a quel momento del video
Condividi:
La maggior parte delle trascrizioni è pronta in meno di 5 secondi
Copia in un clicOltre 125 lingueCerca nel contenutoVai ai timestamp
Incolla l'URL di YouTube
Inserisci il link di qualsiasi video YouTube per ottenere la trascrizione completa
Modulo di estrazione trascrizione
La maggior parte delle trascrizioni è pronta in meno di 5 secondi
Installa la nostra estensione per Chrome
Ottieni le trascrizioni all'istante senza uscire da YouTube. Installa la nostra estensione per Chrome e accedi con un clic alla trascrizione di qualsiasi video direttamente dalla pagina di riproduzione.
