UNCENSORED GPT4 x Alpaca & Vicuna [Local Install + Tutorial]

Links:

------------------------------------------------------------------------------------------
:start
pause
goto start

----------------------------------------------------------------------------------------------

#llama #gpt4xalpaca #vacuna

[oliverli]

holy moly! this is much faster than the previous CPU chat bots! awesome! keep going, open source community!

[mondainthewizard]

Thank you so much for creating such a fantastic video! Your attention to detail and clear explanations made it so easy to follow along and learn. I can tell you put a lot of effort and thought into making this, and it definitely paid off. Keep up the amazing work!

[nartrab]

Super easy and straight foward tutorial, thank you!

[juanmacias]

Hey thanks for this video, super cool to see that we can utilize open source models. :D

[rip-emix]

A video on an easy way to connect the GPT4-x-Alpaca to a chatbot assistant would be great as well!

[jeffwads]

Wow. This thing actually remembers context of your previous prompts. That is a huge benefit. Also, after using this 13b vicuña for bit, its comprehension is pretty incredible. It is bad at math, but it passed the GPT-4 stack items test with flying colors. The 30b alpaca fails that test. Interesting.

[stevebruno]

man the files in those repositories change every hour tracking down the ones you used was a challenge but once i did it worked fine. THANK YOU!

[ninocrudele]

Great video, I was looking for this since long time, I subscribed immediately, please keep going!

[Desaster]

Thats a pretty good model. From my opinion the strongest you can run locally at the moment

[DrWne-WasteLand]

that's the best tutorial on that topic, thanks you very much.

[Karvan]

For those who care, the 7B version is uncensored, the 13B one is censored

[francescabazzi]

A very interesting and well-explained video! thank you!!! To speed up my computer, I set it to 8 threads, I asked for an explanation of the "Pythagorean theorem" and after 22 minutes of processing, the answer I got was the "Theorem of the bisector of the isosceles triangle"... I'm a bit disappointed about this... I was hoping for something better.. in the sense of performance but also for the type of response obtained

[CuteAssTiger]

best feeling when you dowload the thing and the contents of the zip look completely different .

why cant anything related to github ever be easy ?

[xlrusa]

Is there a way to run the alpaca models on multiple gpus? Is there a Linux based OS that can be setup to run models on multiple gpus. Basically something like HiveOS but for running machine models. This could open up a new service/platform that would greatly accelerate machine learning and it should be open source and open to all.

[frankebert]

can you review some prompts to make it more on the jailbroken side? seems to be more on the censored side than I thought it would be. Thank you for your vids!!

[dimasavin]

When I try to run the model I get an error.
Does anyone know how to fix this?

error loading model: failed to open ggml-model-q4_1.bin: No such file or directory
llama_init_from_file: failed to load model
main: error: failed to load model 'ggml-model-q4_1.bin'

[xzendordigitalartcreations]

This was good and easy to follow video presentation. I tried the three versions of the Vicuna model based on AVX, AVX2 and AVX512 and only the AVX512 version failed to work. I may have one of the newer CPUs in which Intel removed support for this extension.

When I asked a question the response was far more detailed than what I received from ChatGPT and it even sited sources for the research it did. Quite an impressive application.

[xmartyxcorex]

Awesome video, I'd love to use these unrestricted models for writing film scripts and stories. We need a Mac version though!

[cleverestx]

Thanks for the video. Is it possible to leverage a RTX 4090 as per this model instead of a CPU? If so, HOW?

[LoGoBDA]

How can I force these models to use my GPU instead of the CPU? Also how do I set the parameters when using the CPU that it uses all of CPU and cores?