AI controls the COMPUTER - Open-Interpreter FULL TUTORIAL!!! 

Thanks. Always good to hear a positive review from a critic 😊

Glad you like them! Thanks so much for your work on this. I wish this grows more and more !

Great!

You're so welcome!

Thanks 😁

You are so welcome

Thanks bro, I try every day, somedays I'm very tired and somedays code just doesn't work

Couldn't agree more!

Hey Kevin! dev of open interpreter here. this is coming by the end of the week :) internally working on the best way to support all these non-function calling models (the reason we started with GPT-4) like claude, cohere, a localhost, HF inference, together, etc. system prompt config is coming soon- it will just be a config.yaml you can easily edit (or edit from the command line with `interpreter --config`)

Hi@@killianlucas Yes, that 'function' without an API must be a challenge, but excellent news that you are working on it. I am already amazed and now flattered to be chatting with Mr. Open Interpreter.

@@killianlucas hi, i installed open interpreter and set my api keys. however i acccidentally deleted the api keys in openai account. i need help in changing my api keys in open interpreter . im not sure what argument to run to do this can you help?

The code is open source so there is some potential

Thanks for the info! I just thought if someone's on Windows or Linux and aren't familiar with this. I'll try to minimise or figure out a way to display as text

Don't assume everyone knows what you know. I'm new to python but not new to coding. Many people, like myself, need step by step with no detail omitted. Perhaps make a step by step of all commands and post it on a website or add it in the description.

@@jackflash6377 understood, but I'm saying it's pointless to tell people about pip3 as pip works in 99.9% of cases, and that's not impacted by how much or how little you know. It's 2023 we can move on. If he really wanted to go into micro detail, he could mention that strictly you should actually be usibg "python -m pip" but that's not worth side-tracking for either

@@jackflash6377agreed

Nope, it looks for the model in that folder and it won't download again

so, can i put my downloaded model into the downloads folder of the interpreter? where is the downloads folder of interpreter anyway?@@1littlecoder

I mentioned that in the video , I think you need use an additional argument fast or something like that.

it is true, always grateful@@1littlecoder

Sorry, I think the argument changed - interpreter --model gpt-3.5-turbo this is how you need to enable in the CLI

Is Buddy the tool name?

@@1littlecoder buddy is what I'm calling you😭

Love you for making video on open-interpreter :)

gpt-3 is not too bad too,using a gpu n larger models in my case.

My bad, I thought you were talking about some new tool that I didn't know :D

You can overwrite the api key by setting the new environment variable. It'd just simply overwrite it

@@1littlecoder ive tried, how do i overwrite it?

Honestly I don't think so but tomorrow we might get some idea if they are!

I couldn't find 7B WizardCoder GGUF and my Mac is Intel so I didn't try with the 13B GGUF, that's definitely worth a shot!

@@1littlecoder I've been experimenting with larger gguf models with a mix of GPU and CPU offload loading as many layers as possible into the GPU. Not sure how to setup open interpreter to load the model on both GPU and CPU though

@@justindressler5992 What's your GPU ?

@@1littlecoder I'm using a 3080 with 12gb this is good to load most 13b models in VRAM. But the 13b models are kind of simple, the 30b+ models behave alot more like gpt 3. I'm also using Runpod to test the biggest models. There has been some recent improvements with coherence in the smaller models especially the LLama 2 ones. When splitty the model over GPU and CPU you have to be vary careful not to overallocate VRAM,because modern systems will start to page VRAm this is far worse than just running on CPU I try to leave 1-2gb if VRAM free incase the system needs it

Yes interpreter.use_azure = True interpreter.api_key = "your_openai_api_key" interpreter.azure_api_base = "your_azure_api_base" interpreter.azure_api_version = "your_azure_api_version" interpreter.azure_deployment_name = "your_azure_deployment_name" interpreter.azure_api_type = "azure"

@@1littlecoder Thanks a lot 🙏

Do I need to add this piece of code after import interpreter...in the command prompt itself?

@@priyankak5589 Yes right

My bad I'm sorry. I missed that here. I had covered that intro in my previous AI news update so I kind of picked up from where I left there.

​@@1littlecoder​Oh ok!

Do you mean it's not working fine?

@@1littlecoder No this has too many moments of error for it to be fine

A side project with over 17k stars.

@@TheReferrer72 I had a side project with a million users. Stars don't indicate stability.

@@matbeedotcom Sure you did....