12. Financial Analyst App - GPT-4, Streamlit, and Llama-Index Tutorial

[@parttimelarry]

Like the video? Support my content by checking out Interactive Brokers using my link: www.interactivebrokers.com/mkt/?src=ptlPY1&url=%2Fen%2Findex.php%3Ff%3D1338 Source Code: github.com/hackingthemarkets/financial-news-llama-index Note: I have updated the source code to reflect recent changes to Llama Index 0.6 that happened after I recorded this video at the end of April. I have made notes on the changes in Github and how they differ from a few lines in the video. 0:00 What We’re Building 2:34 Recap of News Fetching and Indexing 5:56 Swapping out Models, GPT-4 10:31 Streamlit UI Tutorial 20:35 Wrapping Up, Other Financial Analyst Tasks

[@نشامي]

Really cool, I am saying this video after 9months of it's release, and I can say this world of llms has developed so much, and we don't know what there in the future, which makes me so excited to this world of programming and ai

[@onemanops]

Thanks Part Time Larry, I never heard of llama index. I wrote an app that generates embedding for pdfs and store them in pgvector, but was not getting the results I expected/wanted. Then you mentioned indexing and I remembered I forgot to create the pgvector index. I never heard of llama index before I will check it out. Your financial ai series is nothing short of brilliant

[@yellowboat8773]

Isn’t creating a vector indexing anyways?

[@ismailk7021]

This is fantastic. I was thinking how to build a stock analysis app and found your video.

[@cnaccio]

Love these hands on projects!

[@peralser]

Thanks Larry. Amazing video!!

[@ramelox]

Hey Larry, I wanted to thank you for teaching me Python. Thanks to your guidance, I've learned API consumption, Flask, Streamlit, pandas, data types, Google functions, databases, openai API and many more. You provide perfect starting points. I've absorbed everything you've taught here and built on top of it. Could you please create projects using Django, unit tests and more object oriented Python? If only there were another Larry for front-end development.

[@theawebster1505]

Very nice video, thanks for the tutorial and explanations!

[@unclecode]

Thank you for the excellent content, as always, Larry. The warning you have faces is important. I wanted to emphasize the importance of being cautious when using GPT 3.5 and 4. While they are conversational models, they have a different architecture than completion models (davinci-003). The way they works is different from completion models in many ways. Therefore, it's best to use something like "ChatOpenAI" or stick with "OpenAI" which intended to be used with completion models. Nonetheless, I appreciate your content and always enjoy watching your videos.

[@yellowboat8773]

Would love a video using the GitHub loader and building on repository adding code etc.

[@rajshailey92]

Thanks Larry, it helps a lot ❤

[@Sceptic850]

Brilliant Larry. Even if i load the index from disk, Query is using lots of tokens. I wonder why..

[@AP-hv5dh]

Pure fire Larry!!!!

[@klammer75]

Amazing! Such a cool application😎🦾

[@mahdishokri2370]

You are the best Larry

[@dservais1]

hello Larry, OpenAI is now 128K, how fast it is going, have you plan to demonstrate the new power ? in this video you were speaking about doing something cool with SQL ?

[@dservais1]

Hello Larry, long time with no news from you. Do you plan to produce soon?

[@parttimelarry]

I happened to post just after this :). I have been traveling a lot the past year so haven't made as many videos, greetings from Scotland!

[@90fulltime1]

Hello amazing tutoial, I am trying to test this our but I keep getting this error: ImportError: cannot import name 'GPTVectorStoreIndex' from 'llama_index' (unknown location) please any help with that?

[@jonconnor6697]

do you have any plans on doing a crypto research and report scraper that will find new and upcoming crypto tokens and will give brief reports about them also report on competing tokens and wrap it up with volume and twitter mention's?

[@SolidBuildersInc]

Hey Larry, Appreciate the video and the project you chose to build. I have been tailgating for about 1.5 years with you before AI came along and now it seems I am able to actually keep up with you ( Only Theoretically, lol ), which is proof positive of time saving on code production. With that being said, and the projects you are currently entertaining with OpenAI, what would be the best setup for building out an autonomous trading bot that doesn't execute the Trade with the broker, but makes the Trade in a Real Time simulation, similar to a Demo account. Have you done something like this in your archives that I might have missed? The idea is to build a ML that is getting trained on Real Time Streaming data and when the results are providing a strong positive or negative trend, the actual execution of Trading is turned on and if it is in question, it is turned off. Your thoughts are appreciated.

[@boba9610]

hey Larry, have you tried anything with assemblyAi yet?

[@jomohogames]

Formating looks like they use markdown to latex formatter, probably pandoc which interprets dollar sign as a start of mathmode

[@rverm1000]

Can you do more video's on vectorbt?

[@Alexrmacleod]

🔥

[@JohnCheninTW]

How to make it in agent with memory ? Please make it with another video

[@user-wr4yl7tx3w]

Is Llama Index better than LangChain? Is there a reason why you chose one over the other?

[@doords]

How much does the GPT tokens cost if you do this all day long

[@DavidGalan777]

I am looking for programmers to develop a web app project like this. Let me know if you are proficient in these related APIs.... Eleven Labs API, OpenAI, ChatGPT API, Whisper API

[@klammer75]

Hey Larry, couldn’t get the index_store.jsonand vector_store.json working…any thoughts? Please advise and love what you do🤔🤷🏼‍♂️😎

[@osagiezogie]

Hi, same problem here. Please did you find any work-around?

[@1ntrcnnctr608]

🦙

[@someetsingh1917]

getting docstore.not found error