Open Interpreter 🖥️ ChatGPT Code Interpreter You Can Run LOCALLY!

you remember a code interpreter from chat EBT it said incredible feature that allows you to upload files to chat gbt and interact with them in natural language so for example you can upload PDFs and have chat GPD summarize it for you you can upload images and convert them into gifs you can upload data and run visualizations or analyze that data it's pretty incredible but now what if you could do all of that locally on your own computer a new project called open interpreter makes that possible open interpreter runs locally on your computer it supports both chat GPT and codelama and it doesn't have any of the limitations that code interpreter does so first I'm going to show you how to install it then I'm going to show you how to use it let's go so this is the GitHub page for open interpreter it is at the top of the trending repositories on GitHub right now and for good reason it is really incredible so here it says open interpreter lets llms run code python JavaScript shell and more locally you can chat with open interpreter through a chat gbt like interface in your terminal by running interpreter after installing so you can create and edit photos videos PDFs ETC control a Chrome browser to perform research plot clean and analyze large data sets and one thing it doesn't mention here is you can actually have it run scripts that change things on your computer so you can say change my computer to dark mode or change it to light mode or tell me how many files I have on my computer the possibilities are really endless so let me show you how to install it now I'm using a PC but I ran through this on a Mac and it works just as well and thank you so much to this video sponsor cap cut online with cap cut online video editor it is super simple to create YouTube videos just like this one I've used other video editors but when it comes to Simplicity and just getting the job done well cap cut online tends to be the best and they made things even easier with their magic tools Suite of features which is AI powered functionality that you can use to create incredible videos without needing to be a professional video editor one of the coolest features from Magic tools is a few feature called Script to video which allows you to just give it a prompt and it'll create a script for you one of the most overwhelming parts of creating video for me is staring at that blank page but with script to video you just input a few ideas and then cap cut will actually generate you that full script as a first draft and of course you can edit it and make it better and make it more akin to your own tone and from that script if you want to take it a step further you can actually convert the script directly into a video and that video will be copyright compliant and ready to export and then you can use the cap cut editor and edit the video to your needs so thank you again to capcut online video editor with all of the AI tools that they provide it truly is the easiest way to create professional looking videos for YouTube for tick tock or really any platform now back to the video so the first thing you're going to do is open up your terminal with Anaconda already installed we're going to be using conda today to isolate the environment now if you don't already have Anaconda installed go ahead and install it it's super easy easy now one thing to note and I generally always recommend using conda to isolate your environment but in this case if you want to be able to run an interpreter from any console then you wouldn't want to use conda and since I'm using conda every time I want to run interpreter I have to activate that conda environment so let's go ahead and do that now conda create dash n o i for open interpreter python equals 3.11.4 and yes I do want to install the packages next we're going to grab this Command right here to activate the environment so I'm going to highlight it copy it and paste it in conda activate oi and there we go we have the conda environment activated next and this really couldn't be easier this is one of the easiest installations that I've come across all you do is PIP install open dash interpreter hit enter and there it's done you now have interpreter ready to go so first I'm going to show you how to use it with gpt4 and that's the default but it does support codelama now admittedly I haven't gotten the codelama version to work but I know it's available and I know they're trying to make it a lot easier year for people to run it on various machines there's a bunch of different issues discussing it on their GitHub repo so check it out I hope you can get it installed I'll still show you how to install it but I haven't been able to do it myself but for now let's just type interpreter and that's it welcome to open interpreter the first thing it's going to ask you for is an open AI API key so if you don't already have an open AI account go ahead and sign up grab an API key come back here and paste it in alright there is my API key and of course I'm going to revoke this API key before publishing this video Don't Worry hit enter and if we don't want to have to enter our API key every time we run interpreter we would do either on a Mac we want to export it or we want to set X right here and it gives you both of those commands right when you're launching and that's it now we're in there so let's test it out now I'm going to say what files do I have on my desktop you hit enter and just like code interpreter it's going to tell you what it's going to do then it's actually going to write the code for that so here it is to find out what files are on your desktop we will use the shell command dir to list all the files in the desktop directory now when I I ran this on the Mac it knew I was on a Mac and now that I'm running it on Windows it knows I'm on Windows it's very impressive so there we go would you like to run this code now every time it outputs code it asks for your approval and you can skip that step you can actually have it run the code automatically by typing interpreter space Dash y to Launch Open interpreter rather than just interpreter so we're going to say yes run the code and it actually ran into an issue and just like code interpreter it's going to try to figure out what the issue is and rewrite its code so there it did that let's try it again and that time it worked I have a PDF I have the cursor program I have a Microsoft Excel file I have Dropbox I have factorio one of the best games ever made OBS YouTube folder and so on so it told me everything on the desktop really impressive now let's interact with the Excel file I have what is the Excel file on the desktop about again I'm not giving it much to run on but it's going to figure out what I mean by that and run the command so to understand what the exact cell file is about we need to open it and read its contents so it's going to install some libraries then it's going to actually run the code open it up figure it out so there it's installing the libraries right from the window they've been successfully installed now it's writing the code to open up the file and see what it's about okay so it got another issue it figured out what the issue is and now it's going to rewrite the code and there we go there's the Excel file and so the file is about and it's actually giving me an example and it contains data about various fashion brands here are the first few rows and that's exactly right so now let's actually run some queries on it which company in that data set has the highest growth rate last year so there it's going to sort the data and then it's going to grab the one that's at top there we go Nike that's perfect that's exactly right but now let's take it a step further I actually want to see a graph of the growth rate of all the companies create a graph of the growth rate of all of those companies and it says it's going to use matplotlib to do so first it has to install great and again you can skip all the approvals simply by typing Dash y when and launching interpreter and there it is absolutely incredible there's the growth rate of the top 10 companies and all I had to do is describe what I want and it created this graph super cool next I don't like the name of that file I'm going to rename it to something much simpler rename that Excel file to Brand data so it ran into an issue so it's trying something different just like code interpreter would let's try that and there we go it worked perfect and let's verify yep right there brandata.xlsx perfect now I have a PDF on my desktop and I'm going to ask it to summarize that PDF summarize the first page of the PDF on my desktop okay so it ran into an issue try it again yep there it goes and again it is self-correcting code which is so so cool okay the packages have been installed now it's actually writing the code to extract the text from that file okay it's having trouble again there is an issue opening the PDF let's try again now it says PDF file reader is deprecated which I found on my Mac as well now let's try something different and there we go it worked now we're going to summarize using the Gen Sim package so let's install it okay so it ran into another issue let's try it again and it's having an issue with Gen Sim and this isn't a problem with open interpreter this is a problem with python and gen Sim so we're going to try a different way same thing now it's saying let's use the Bert based model and I'm going to say no let's use the sum y package okay so got it installed now let's use some y or is it pronounced Sumi I'm not sure there it wrote some code let's do it and it has an issue and it says the nltk data needs to be downloaded so let's download it downloaded it unzipped it great successful now let's try it again and here's the summary it worked perfect so it talks about the attribution of course it's only the first page of the attention is all you need document and I tried on my Mac now the authors of open interpreter are well aware of this they're working on it and I can't wait until code interpreter is available so that this is truly fully open source completely local now if you did want to run it with cold Lama here's how you would do it and I'm going to show you and then I'm going to show you where I get stuck so first you type interpreter and then dash dash local and the first thing you're met with is asking what parameter count model you want to download so we're going to download the smallest one because I know it's not going to work so let's save some time then I'm going to ask for the lowest size and then it's asking if I want to use a GPU which I do because I have one on this computer so the cold Lama interface is not found let's install it and this is where I always get stuck failed to install code llama we have likely not built the proper code llama support in your system and they even say here running large language models locally is difficult and if you have any insight hop in their Discord and help them out so that's it I can't wait for this to have code llama but in the meantime this is absolutely incredible this is a fully open source version of code interpreter that you can run on your machine and I want to know what you do with it tell me the use cases you think of in the comments below and if you want me to create another video going over all the use cases that you all mentioned let me know in the comments I'll do it if you like this video please consider giving a like And subscribe and I'll see you in the next one

Channel: Matthew Berman

Views: 175,936

Keywords: chatgpt code interpreter, chatgpt code interpreter plugin, code interpreter, open interpreter, llama 2, llama, meta ai, chatgpt, openai, open source, tutorial, ai, artificial intelligence

Id: SqnXUHwIa3c

Length: 9min 59sec (599 seconds)

Published: Thu Sep 07 2023